Metformin versus dietary treatment in nonalcoholic hepatic steatosis: a randomized study

Article metrics



We aimed at evaluating whether the addition of low-dose metformin to dietary treatment could be an effective approach in nondiabetic patients with nonalcoholic fatty liver disease (NAFLD).


We carried out a 6-month prospective study in a series of overweight or obese patients with ultrasonographic diagnosis of hepatic steatosis. In total, 50 patients were enrolled and randomized into two groups: the first group (n=25) was given metformin (1 g per day) plus dietary treatment and the second group (n=25) was given dietary treatment alone.


At the end of the study, the proportion of patients with echographic evidence of fatty liver was reduced in both the metformin (P<0.0001) and the diet group (P=0.029). Moreover, patient body mass index and waist circumference significantly decreased in both groups (P<0.001). Fasting glucose, insulin resistance (evaluated as homeostasis model assessment of insulin resistance (HOMA-IR)) and serum adiponectin decreased in both groups, although these changes reached statistical significance only in the metformin group. In this group, HOMA-IR decreased from 3.3±1.6 to 2.4±1.2 (P=0.003), whereas it decreased from 3.2±1.6 to 2.8±1.1 (not significant, NS) in the diet group. Similarly, the proportion of patients with impaired fasting glucose declined from 35 to 5% (P=0.04) in the metformin and from 32 to 12% (NS) in the diet group. At baseline, 40% of patients in both groups met the diagnostic criteria of metabolic syndrome. This proportion decreased to 20% in the metformin group (P=0.008) and to 32% in the diet group (NS).


In our 6-month prospective study, both low-dose metformin and dietary treatment alone ameliorated liver steatosis and metabolic derangements in patients with NAFLD. However, metformin was more effective than dietary treatment alone in normalizing several metabolic parameters in these patients.


Nonalcoholic fatty liver disease (NAFLD) is the most common cause of elevated liver enzymes and also one of the most common forms of liver disease in the world. It is now estimated to affect about 20–30% of people in Italy and other Western countries. In contrast, the prevalence of chronic hepatitis C virus infection is estimated at 3% of the world's population.1 NAFLD includes a broad spectrum of liver alterations, ranging from pure steatosis to cirrhosis, and nonalcoholic steatohepatitis (NASH). NAFLD is characterized by liver damage and functional impairment similar to those observed in alcoholic liver disease, although occurring in patients who do not drink or drink only a moderate amount of alcohol.2 NAFLD occurs in 60–95% patients with obesity, in 28–55% patients with type 2 diabetes mellitus (T2DM) and in 27–92% patients with dyslipidemia.3

Insulin resistance with compensatory hyperinsulinemia is the common denominator of obesity, T2DM and dyslipidemia, and is believed to have a pathogenetic role in NAFLD.

Accordingly, it has been reported that insulin resistance is the single laboratory finding most closely associated with NAFLD in a large series of patients, irrespective of body mass index (BMI), fat distribution or glucose tolerance.4 Indeed, NAFLD has been recently proposed as an additional feature of the metabolic syndrome (MS).5

Moreover, cytokines have a pivotal role in pathogenesis and progression of NAFLD. For this reason, in the present study we evaluated changes in circulating levels of adiponectin that seems to be a protective factor in liver inflammation and fibrosis.6

The therapeutical approach of NAFLD is currently based on lifestyle intervention, whereas there is no consensus on an effective pharmacological treatment. Several drugs (insulin sensitizers, hepatoprotective agents and lipid-lowering drugs)5, 7 have been evaluated as potential therapeutical agents, but results are inconclusive. In view of the fact that insulin resistance is the key metabolic abnormality in NAFLD, insulin-sensitizing agents such as peroxisome proliferator-activated receptor-γ agonists belonging to the thiazolidinediones class and metformin, have been recently proposed as a rational therapeutic strategy in NAFLD.

Three small trials and one large randomized placebo-controlled trial have evaluated thiazolidinedione pioglitazone in the treatment of NAFLD.8, 9, 10, 11 The small trials reported an improvement in insulin sensitivity and in alanine transaminase (ALT) levels. The larger randomized study showed that treatment with pioglitazone was associated with a significant improvement in ALT, hepatic fat content and insulin sensitivity as compared with placebo.

Three additional trials, two open-label and one placebo-controlled, have evaluated the effect of a second thiazolidinedione, rosiglitazone.12, 13, 14 All these studies have reported a significant decrease in ALT and a histological improvement of liver disease. However, thiazolidinediones may have significant side effects, and long-term safety is yet to be established.12, 15 Some studies have provided evidence supporting beneficial effects of metformin in patients with NAFLD. However, most of them have involved small groups of patients and have been designed as single-arm trials.15, 16, 17, 18, 19 The use of metformin in NAFLD patients was associated with a reduction in serum aminotransferase levels in three single-arm trials and with a statistically significant amelioration in liver histology in one of these studies. Only two small randomized studies compared the efficacy of metformin versus diet in NAFLD, overweight or obese, diabetic patients.15, 17 These studies have reported amelioration of liver function tests in the metformin and diet group, but the reduction was not as significant in the diet group as in the metformin group, and they have not shown the superiority of metformin versus dietary treatment on liver histology. In fact, only one of two studies evaluated hepatic histology, and did not find significant changes in histology.17 A recent randomized controlled trial of metformin versus vitamin E versus dietary treatment, reported a greater reduction of liver aminotransferases and a greater overall improvement of metabolic parameters in patients treated with metformin.18 Although encouraging, these results are far from being conclusive. In fact, most of these studies have involved small series of patients, had short treatment period and were not well controlled. Moreover long-term benefits and histological benefit are not shown.

In the present prospective, randomized study, we compared the efficacy of low-dose metformin plus dietary treatment versus dietary treatment alone in 50 NAFLD patients during a period of 6 months. We found that both treatments ameliorated liver steatosis and metabolic derangements. However, metformin was more effective than dietary treatment alone in normalizing several metabolic parameters in NAFLD patients.

Materials and methods


The study was carried out in 50 patients with BMI>25 kg m−2 and diagnosis of liver steatosis at ultrasonography (US). Patients were recruited, from July 2006 to January 2008, among those referred to the Endocrine Unit of University ‘Magna Graecia’ of Catanzaro for treatment of obesity or overweight.

We excluded patients with evidence of heart disease, renal failure and diabetes, even if newly discovered. Exclusion criteria also included smoking habits, alcohol intake of more 20 g per day, signs of hepatic virus infection (patients were tested for hepatitis B antigen or hepatitis C antibodies), the presence of clinical or biochemical evidence of autoimmune, metabolic or genetic liver diseases and use of drugs known to induce liver steatosis (for example, valproato, prednisone, amiodarone, calcium-channel blockers, cocaine and tamoxifen). Notably, circulating values of aminotransferases were in the normal range. All subjects gave a written informed consent to take part to the study.

Patients were randomized into two groups. The first group (25 patients) was given metformin (1000 mg per day) plus hypocaloric (1300 kcal) diet, the second group (25 patients) was given dietary treatment alone. Metformin was started at a dose of 250 mg twice a day and it was increased after a week to the final dose of 500 mg twice a day, to reduce gastrointestinal side effects. Both groups were treated for 6 months. Primary end points of this study were liver steatosis amelioration or disappearance, as evaluated by US. Secondary end points included amelioration of the metabolic derangements associated with liver steatosis evaluated as BMI, waist circumference, homeostasis model assessment (HOMA)-Index, fasting glucose, lipid profile, aspartate aminotransferase, ALT and circulating adiponectin.

All patients had a complete clinical, anthropometric and laboratory investigation at baseline. Patients were carefully interviewed to obtain a social and family history, and anthropometric data were obtained (height, waist, hip circumference and weight). BMI was calculated as body weight in kg divided by the square of the patient' s height in m. Waist-to-hip ratio was calculated as an index of splanchnic fat.

Laboratory investigation and US evaluation

Laboratory investigations included a routine biochemistry and a complete evaluation of iron status (serum iron, transferrin, ferritin concentration). C-reactive protein was measured in all patients as an inflammation marker. A 2-h glucose tolerance test was also performed in all patients. Fasting glucose and insulin levels were used to calculate insulin resistance, according to the homeostasis model assessment technique (HOMA-insulin resistance, HOMA-IR). HOMA-IR was calculated as previously described (HOMA-IR=insulin (μU ml−1) × glucose (mmol l−1)/22.5),20 and a HOMA-IR value 1.5 was considered indicative of normal insulin sensitivity.21

A blood sample for measurement of adiponectin was collected. Plasma samples were immediately separated and stored frozen at −80 °C until measurement. Plasma levels of adiponectin were assayed by an enzyme-linked immunosorbent assay (Mediagnost, Reutlinger, Germany) with intra- and interassay coefficients of variations of 6.7 and 4.7%, respectively. The assay sensitivity is 0.0006 μg ml−1.

All patients underwent US liver evaluation by a single experienced operator (M.D.S.), blinded to the clinical data, using a 3.75 MHz convex transducer (Toshiba Power vision 6000, Toshiba, Rome, Italy). The diagnosis of liver steatosis was made on the basis of echographic liver features22 as follows: (1) liver hyperechogenicity relative to adjacent right kidney or spleen, (2) unclear display of intra-hepatic lacuna structures, (3) mild to moderate hepatomegaly with a round and blunt border and (4) unclear display of envelop of right liver lobe and diaphragm. Liver steatosis was graded as mild, moderate and severe. Mild steatosis was characterized by a mild increase in liver echogenicity. Moderate steatosis appeared as increased liver echogenicity that obscured hepatic and portal vein walls. Severe steatosis was characterized by posterior attenuation of the deep liver parenchyma.23 All patients then underwent clinical, laboratory and liver US evaluation once in every 8 weeks till the end of the study.

Definition of metabolic variables and diagnosis of MS

Patients were classified into overweight (BMI>25 to <30 kg m−2) and obese (BMI 30 kg m−2) as previously described.24 Men and women with waist circumference values of at least 94 and 80 cm, respectively, were considered to be affected by central obesity.25

Patients were considered hypertensive if the systolic blood pressure was 130 mm Hg or greater and/or a diastolic blood pressure of 85 mm Hg or greater, or if they were using antihypertensive drugs, according to the guidelines of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure.26

Impaired fasting glucose (IFG) level was defined as a fasting plasma glucose level of 100 mg per 100 ml or greater (5.6 mmol l−1), but less than 126 mg per 100 ml (<7.0 mmol l−1).27 Hypertriglyceridemia and low high-density lipoprotein cholesterol level were defined according to the Adult Treatment Panel III (ATP III) of the National Cholesterol Education Program guidelines (triglyceride level >150 mg per 100 ml; high-density lipoprotein cholesterol level, <40 mg per 100 ml men or <50 mg per 100 ml (<1.3 mmol l−1) in women).28 A diagnosis of MS required fulfilment of at least three criteria.25

Statistical analysis

All study variables were treated as continuous variables. Summary outcome measures were reported as mean±s.d. Statistical differences of laboratory and clinical variables between the two treatment groups at baseline were evaluated by the Student's t-test. The χ2-test or Fisher test was used to compare the frequency data. One-way analysis of variance (ANOVA) and post hoc Bonferroni's test were used to evaluate intragroup differences over the treatment period (baseline, 8 weeks, 16 weeks and 24 weeks). Logistic regression analysis was used to examine independent factors associated with NAFLD amelioration. A P-value <0.05 was considered statistically significant.

Data analysis was performed using the statistical package SPSS version 11.0 for Windows (SPSS, Chicago, IL, USA).

Review of previous studies using metformin in NAFLD/NASH patients

To identify all the trials evaluating the efficacy of metformin in NAFLD, we searched MEDLINE for studies that focused on the use of metformin in adult NAFLD/NASH patients. We also checked the reference list of any original article and review found in MEDLINE regarding therapy for NAFLD/NASH patients. The search was limited to English-language articles. In total, we found nine studies complying with our searching criteria (Table 4).


In total, 20 patients (2 males and 18 females) randomized to metformin treatment (metformin group) and 25 patients (5 males and 20 females) randomized to dietary intervention alone (diet group) completed 6 months of treatment. Five patients (all in the metformin group) dropped out because of noncompliance to the treatment (three patients after 2 months and one patient after 4 months because of change of residence, and one patient after 5 months because of pregnancy).

At baseline, the two treatment groups showed similar clinical and biochemical variables (Table 1). In particular, all patients were either overweight (10 of 45=22.2%) or obese (35 of 45=77.8%). BMI was 36.5±4.9 kg m−2 in the metformin group and 34.7±3.5 kg m−2 in the diet group (Table 1). All patients were affected by central obesity, that is, the mean value of waist circumference was 108.4±12.5 cm in the metformin group and 102.8±8.8 cm in the diet group.

Table 1 Clinical and laboratory features in NAFLD patients of the metformin and the diet group at baseline

Mean serum concentrations of fasting glucose and insulin were at the upper limit of normal values in both groups. IFG was present in 15 patients (7 in the metformin and 8 in the diet group) and impaired glucose tolerance was present in 6 patients (3 in the metformin and 3 in the diet group). The great majority of patients in both groups were clearly insulin resistant. Mean post-oral glucose tolerance test insulin concentration was equally increased in both groups (76.5±53.7 and 85.1±90.5 mU ml−1, respectively, in the metformin and in the diet group) and 27 patients (12 in the metformin and 15 in the diet group) showed a concentration of >50 mU ml−1. Mean HOMA-IR values were also increased in both groups (Table 1). A HOMA-IR value >1.5 was observed in 37 patients (15 in the metformin and 22 in the diet group). However, only six patients (four in the metformin and two in the diet group) showed basal hyperinsulinemia (insulin >25 mU ml−1).

Serum aspartate aminotransferase and ALT were within the normal range. Only two patients presented a mild elevation in serum concentration of aminotransferases, one in the metformin group and one in the diet group.

Low-density lipoprotein cholesterol and PCR values were either at the upper limit of the normal range or slightly increased. MS, as defined by the ATP III criteria, was present in 45% and in 36% of patients in the metformin and in the diet group, respectively (Table 1).

At the end of the study period, BMI significantly decreased in both treatment groups. In the metformin group, it decreased from 36.5±4.9 to 33.3±5.4 kg m−2 (P=0.0001), whereas in the diet group, it decreased from 34.7±3.5 to 32.9±3.6 kg m−2 (P=0.0001). The mean weight loss in patients on metformin was 8 kg (P=0.0001), whereas in patients on diet alone it was 5 kg (P=0.0001). In both groups, weight loss was maximal during the first 2–3 months and then progressively declined during the last 3 months of treatment (Figure 1).

Figure 1

BMI variations in patients of the metformin and the diet group during the study. Squares represent values (m±s.d.) in the metformin group, circles represents values (m±s.d.) in the diet group. In both groups, weight loss was greater in the first 4 months, with a smaller additional reduction during the last 2 months.

Waist circumference also progressively and significantly decreased both in the metformin group (108.4±12.7 versus 101.4±12.7 cm, basal versus final, P=0.0009) and in diet group (102.8±8.8 versus 95.6±8.5 cm, basal versus final, P=0.0001) (Tables 2 and 3).

Table 2 Clinical and biochemical features (m±s.d.) in the NAFLD patients (n=20) of the metformin group at baseline and during the study
Table 3 Clinical and biochemical features (m±s.d.) in the NAFLD patients of the diet group (n=25) at baseline and during the study

Fasting glucose, basal serum insulin and HOMA-IR index values decreased in both groups. However, differences both in fasting glucose (from 92.4±9.9 to 89.1±9.3 mg per 100 ml, P=0.04) and HOMA-IR index (from 3.3±1.6 to 2.4±1.2, P=0.003) reached statistical significance only in the metformin group (Table 2 and Figure 2). HOMA-IR index improved markedly after the first 8 weeks of metformin treatment (Table 2 and Figure 2).

Figure 2

Changes in fasting glucose (a) and HOMA-IR index (b) in NAFLD patients of the metformin and the diet group during the study. In both patient groups, fasting glucose and HOMA-IR index decreased reaching statistical significance only in the metformin group. °NS, *P<0.05 (ANOVA test).

At baseline, IFG was present in 7 of 20 (35%) patients of the metformin group and in 8 of 25 (32%) of the diet group. At final evaluation, IFG persisted in 1 of 20 (5%) patients of the metformin group (P=0.04, final versus baseline) and in 3 of 25 (12%) patients of the diet group (P=0.17, final versus baseline) (Figure 3).

Figure 3

Patients with IFG (a) or metabolic syndrome (b) in the metformin and the diet group at baseline and at the end of the study. The proportion of patients with IFG or metabolic syndrome decreased in both groups, although this reduction reached statistical significance only in the metformin group. °NS, *P<0.05 (χ2-test).

At baseline, impaired glucose tolerance was present in 3 of 20 (15%) patients of the metformin group and in 3 of 25 (12%) of the diet group. At final evaluation, impaired glucose tolerance persisted in two patients of the metformin group (not significant) and in all three patients of the diet group.

High-density lipoprotein cholesterol and triglyceride serum concentrations decreased in both treatment groups, but differences did not reach statistical significance.

At baseline, 8 of 20 (40%) patients of the metformin group and 9 of 25 (36%) of the diet group met the diagnostic criteria for MS. At the end of study, patients with this diagnosis were reduced to 4 of 20 (20%) in the metformin group (P=0.008, final versus baseline) and to 8 of 25 (32%) in the diet group (P=0.9, final versus baseline) (Figure 3).

Liver steatosis, as evaluated by US, improved/disappeared in 5 of 20 (25%) patients of the metformin group (P<0.0001, final versus baseline) and in 6 of 25 (24%) patients of the diet group (P=0.029, final versus baseline). In particular, in the metformin group, liver steatosis changed from moderate to mild grade in two patients and disappeared in three patients, whereas in the diet group, it changed from moderate to mild grade in five patients and disappeared in one.

Plasma adiponectin measurement was available in 12 patients of the metformin group and in 19 patients of the diet group. Adiponectin increased in both groups, although this increase reached statistical significance only in the metformin group (from 5.8±2.7 to 7.0±3.3 μg ml−1, P=0.005, in the metformin group and from 7.9±4.4 to 8.5±4.6 μg ml−1, P=0.17 in the diet group) (Figure 4). In contrast, BMI significantly decreased in all these patients (from 35.2±5.6 to 32.19±6.5 kg m−2, P=0.006, in the metformin group, and from 31.4±3.7 to 29.2±3.8 kg m−2, P=0.0002, in the diet group) (Figure 4).

Figure 4

Changes in BMI (a) and adiponectin levels (b) in 12 patients of metformin group and in 19 patients of the diet group during the study. In both patient groups, BMI significantly decreased. Adiponectin increased in both groups, but reached statistical significance only in the metformin group. °NS, *P<0.05 (ANOVA test).

According to logistic regression analysis, no clinical/biochemical feature at baseline or a specific treatment group was independently associated with improvement/disappearance of liver steatosis. Covariates included gender, age (>50 years), BMI (30 kg m−2), waist circumference (94 cm in men and 80 cm in women), diastolic blood pressure (85 mm Hg), systolic blood pressure (130 mm Hg), fasting glucose (100 mg per 100 ml), triglycerides (150 mg per 100 ml), high-density lipoprotein (<40 mg per 100 ml in men or <50 mg per 100 ml in women), HOMA-IR index (1.5) and group of treatment (data not shown).

Review of previous studies using metformin in NAFLD/NASH patients

From a review of the English-language articles, we found nine studies using metformin in adult patients with NAFLD or NASH. Five of them were randomized and only one placebo controlled. One study was multicentric. Most studies included diabetic patients and/or patients with NASH and/or elevated aminotransferases. Only one small open-label, single-arm study used 1 g per day metformin, whereas higher and variable doses of metformin were generally used. In two of these studies, metformin-treated patients were apparently not prescribed a diet. Monitoring of liver function and metabolic parameters was very variable. None measured adiponectin. The differential characteristics of these studies are reported in Table 4.

Table 4 Schematic summary of existing studies reporting the use of metformin in adult patients with NAFLD/NASH as compared with our present study


In the present study, we evaluated the efficacy of a pharmacological treatment with low-dose metformin as compared with dietary measures alone in obese, nondiabetic patients with NAFLD in a prospective, randomized study. We found that metformin therapy was associated with an improvement or disappearance of hepatic steatosis, although it was not significantly more effective than diet alone. However, metformin treatment was also associated with a significantly greater amelioration of a number of metabolic parameters, as compared with diet alone. In fact, metformin increased insulin sensitivity and reduced fasting glucose more effectively than diet alone. At baseline, IFG was found in 35% of NAFLD patients in both groups. At the end of the study, it disappeared in 86% of patients of the metformin group and only in 62% of patients of the diet group. Metformin treatment was also significantly more effective than diet alone in reducing the proportion of patients who met the diagnostic criteria of MS. The proportion of patients with this diagnosis decreased from 40 to 20% in the metformin group (P=0.0008) and only from 36 to 32% in the diet group (P=0.9). This result is important given the high proportion of patients diagnosed with MS in our series of NAFLD patients and the potential of MS to increase the risk for T2DM and cardiovascular diseases.29, 30 Accordingly, NAFLD itself is strongly and independently associated with increased risk of T2DM and cardiovascular disease.29

The first-line approach to NAFLD is currently based on diet and lifestyle modification. Several studies have suggested that weight loss may have beneficial effects in NAFLD and NASH patients. The majority of these studies have found an improvement in liver biochemistry profile after weight loss.28, 31, 32, 33, 34, 35 However, most of these studies are nonrandomized and short term. Therefore, the paucity of data makes it difficult to produce evidence-based recommendations about dietary modifications and exercise in NAFLD patients. Moreover, dietary treatment is limited by the lack of compliance and the frequent regain of weight at follow-up.

A pharmacological treatment in patients with NAFLD is not recommended yet. Because insulin resistance is nearly universal in patients with NASH, it is not surprising that several studies have evaluated insulin sensitizers as a possible treatment. However, a large number of them are proof-of-concept studies with small number of patients and with no rigorous study design, making it difficult to make definite recommendations. Biguanides (metformin) and thiazolidinediones (pioglitazone and rosiglitazone) are the two classes of insulin sensitizers studied in humans.5 Literature supports a potential therapeutic role of metformin in NAFLD. Metformin decreases hepatic glucose output, increases glucose utilization in peripheral tissues and has an antilipolytic effect, reducing serum-free fatty acid concentrations.36 Although the exact mechanism of action of metformin is not yet fully elucidated, it is known that several metabolic effects of metformin are mediated through the activation of LKB1 kinase in liver, leading to upregulation of adenosine monophosphate-activated protein kinase (AMPK), a master regulator of glucose metabolism.37 Activation of AMPK by metformin, not only inhibits gluconeogenesis by inducing the dissociation of the CREB-CBP-TORC2 trascriptional complex and transcriptional inhibition of key enzymes38 but also inhibits the sterol regulatory element-binding protein-1c, which is a major regulator of fatty acid synthesis and is inappropriately increased in liver of NAFLD patients.39 Unlike other insulin sensitizers, such as thiazolidinediones, whose side effects may raise concern, the long-term safety of metformin is well established.12, 29, 40

Our present findings are in agreement with results of Bugianesi et al.,18 who reported that metformin was more efficacious than dietary treatment on metabolic parameters in NAFLD patients and of Haukeland et al.41, who found that the use of metformin was associated with weight loss and amelioration in the metabolic profile in NAFLD patients. Other studies have also found a beneficial effect of metformin on metabolic parameters,15, 16, 17, 41 and on liver enzymes and histology in NAFLD patients.16 Bugianesi et al.18 reported a greater improvement in liver enzymes and in histological features with metformin, as compared with diet alone or vitamin E. However, a recent 24-month observational study in pediatric patients by the same group was unable to confirm this greater efficacy of metformin as compared with lifestyle intervention.19

Our prospective and randomized study differs from previous studies in many features (Table 4) and provides novel information. At variance with previous studies that included both nondiabetic and also diabetic patients,16, 18, 37 we enrolled exclusively nondiabetic overweight/obese adult patients. This is a very important point because diabetes is the only independent variable associated with advanced-stage NAFLD.42 Moreover, all our patients were asymptomatic and with normal values of liver enzymes. In contrast, most previous studies have dealt with patients with clinical symptoms and elevated ALT (Table 4). Therefore, we focused on an early stage of the disease. It is, however, well known that serum transaminases may be normal in patients with histologically proven NAFLD and are not useful for making this diagnosis.42, 43

We also clearly showed that a low metformin dose (1 g per day) is sufficient to induce clinically significant beneficial effects on metabolic parameters. These effects are greater than those observed with diet alone, in spite of similar weight loss in the two groups. This metformin dose is lower than that used in other studies16, 17, 18 and it was extremely well tolerated by all our patients. Although the possibility exists that a higher metformin dose could be more effective, this should be balanced against an increased rate of side effects and drop out. Moreover, a recent meta-analysis showed that a metformin dose of 500–750 mg per day was apparently as effective as 1700 mg per day in preventing the progression of prediabetes to overt T2DM.44

Finally, we observed that metformin led to a significant increase of plasma adiponectin levels, whereas diet alone led to a nonsignificant adiponectin increase. These results are in line with the greater amelioration of insulin sensitivity we observed in the metformin group as compared with the diet group and suggest a possible role of this cytokine as a future noninvasive marker of NAFLD treatment. Clearly, larger studies are needed to confirm this data.

Recent studies have reported that adiponectin activates AMPK,45 has anti-inflammatory effects in liver and prevents lipid accumulation by increasing β-oxidation of free fatty acids and/or by decreasing de novo free fatty acids within hepatocytes, through the downregulation of sterol regulatory element-binding protein-1.6 Our findings, therefore, may support the hypothesis that metformin activates AMPK both directly and indirectly, through adiponectin. In turn, AMPK downregulates sterol regulatory element-binding protein-1 and decreases liver steatosis.

Although liver biopsy is still considered the gold standard for distinguishing between the broad range of chronic liver diseases,46 the fact that we did not use liver biopsy to monitor NAFLD should not be regarded as a serious limitation. Several reasons, including ethical reasons, accounted for this choice. First, our patients were not eligible to liver biopsy. In fact, liver biopsy is not recommended in NAFLD patients, because of its cost, the potential risk of bleeding and the absence of consensus regarding the histopatological criteria that firmly define NASH and differentiate between NAFLD entities.47, 48 Accordingly, a repeated liver biopsy after a short-term therapeutic trial is considered too invasive, especially because valid surrogate markers are available.7 In a recent study, US showed a high sensitivity (91.7%) and specificity (100%) in detecting fatty liver49, and a recent review confirmed that US can accurately identify steatosis with a sensitivity and a specificity of 80–100%.46 Moreover, the accuracy of US in detecting steatosis is not affected by obesity.43 One limitation of US is to be somewhat operator-dependent when used for evaluating small changes in liver fat with time.50 To circumvent this limitation, all US examinations were carried out by a single experienced operator (M.D.S.) blinded to the clinical data. Another potential US limitation is its relative insensitivity in detecting hepatic steatosis when liver fibrosis is present. However, our patients were asymptomatic, nondiabetic and showed normal liver enzymes. They had, therefore, no features predicting liver fibrosis.

Finally, a recent prospective study51 evaluated US accuracy in NAFLD patients undergoing liver biopsy. US had a high sensitivity (100%) and specificity (90%) in these patients, especially when the total area of hepatocytes with steatosis exceeded 20%. At lower levels of fat content, although the sensitivity was significantly reduced, the specificity of US remained high, showing that serial US is an accurate tool for noninvasive monitoring of efficacy of interventions in NAFLD patients.

In conclusion, our study suggests that a low dose of metformin might be proposed to NAFLD patients, especially if they meet the diagnostic criteria of MS. However, long-term studies on a larger group of patients are needed to definitively determine the efficacy of metformin in ameliorating liver characteristics in NAFLD patients.


  1. 1

    Bellentani S, Marino M . Epidemiology and natural history of non-alcoholic fatty liver disease (NAFLD). Ann Hepatol 2009; 8 (Suppl 1): S4–S8.

  2. 2

    Festi D, Colecchia A, Sacco T, Bondi M, Roda E, Marchesini G . Hepatic steatosis in obese patients: clinical aspects and prognostic significance. Obes Rev 2004; 5: 27–42.

  3. 3

    Bugianesi E, Zannoni C, Vanni E, Marzocchi R, Marchesini G . Non-alcoholic fatty liver and insulin resistance: a cause-effect relationship? Dig Liver Dis 2004; 36: 165–173.

  4. 4

    Marchesini G, Brizi M, Bianchi G, Tomassetti S, Bugianesi E, Lenzi M et al. Nonalcoholic fatty liver disease: a feature of the metabolic syndrome. Diabetes 2001; 50: 1844–1850.

  5. 5

    Vuppalanchi R, Chalasani N . Nonalcoholic fatty liver disease and nonalcoholic steatohepatitis: selected practical issues in their evaluation and management. Hepatology 2009; 49: 306–317.

  6. 6

    Tsochatzis EA, Papatheodoridis GV, Archimandritis AJ . Adipokines in nonalcoholic steatohepatitis: from pathogenesis to implications in diagnosis and therapy. Mediators Inflamm 2009; 2009: 831670.

  7. 7

    Socha P, Horvath A, Vajro P, Dziechciarz P, Dhawan A, Szajewska H . Pharmacological interventions for nonalcoholic fatty liver disease in adults and in children: a systematic review. J Pediatr Gastroenterol Nutr 2009; 48: 587–596.

  8. 8

    Sanyal AJ, Mofrad PS, Contos MJ, Sargeant C, Luketic VA, Sterling RK et al. A pilot study of vitamin E versus vitamin E and pioglitazone for the treatment of nonalcoholic steatohepatitis. Clin Gastroenterol Hepatol 2004; 2: 1107–1115.

  9. 9

    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: 1176–1184.

  10. 10

    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: 188–196.

  11. 11

    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: 2297–2307.

  12. 12

    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: 100–110.

  13. 13

    Wang CC, Huang HT . Voice aerodynamic analysis of normal Taiwanese adults. J Formos Med Assoc 2005; 104: 868–872.

  14. 14

    Neuschwander-Tetri BA, Brunt EM, Wehmeier KR, Oliver D, Bacon BR . Improved nonalcoholic steatohepatitis after 48 weeks of treatment with the PPAR-gamma ligand rosiglitazone. Hepatology 2003; 38: 1008–1017.

  15. 15

    Marchesini G, Brizi M, Bianchi G, Tomassetti S, Zoli M, Melchionda N . Metformin in non-alcoholic steatohepatitis. Lancet 2001; 358: 893–894.

  16. 16

    Nair S, Diehl AM, Wiseman M, Farr Jr GH, Perrillo RP . Metformin in the treatment of non-alcoholic steatohepatitis: a pilot open label trial. Aliment Pharmacol Ther 2004; 20: 23–28.

  17. 17

    Uygun A, Kadayifci A, Isik AT, Ozgurtas T, Deveci S, Tuzun A et al. Metformin in the treatment of patients with non-alcoholic steatohepatitis. Aliment Pharmacol Ther 2004; 19: 537–544.

  18. 18

    Bugianesi E, Gentilcore E, Manini R, Natale S, Vanni E, Villanova N et al. A randomized controlled trial of metformin versus vitamin E or prescriptive diet in nonalcoholic fatty liver disease. Am J Gastroenterol 2005; 100: 1082–1090.

  19. 19

    Nobili V, Manco M, Ciampalini P, Alisi A, Devito R, Bugianesi E et al. Metformin use in children with nonalcoholic fatty liver disease: an open-label, 24-month, observational pilot study. Clin Ther 2008; 30: 1168–1176.

  20. 20

    Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC . Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412–419.

  21. 21

    Ascaso JF, Pardo S, Real JT, Lorente RI, Priego A, Carmena R . Diagnosing insulin resistance by simple quantitative methods in subjects with normal glucose metabolism. Diabetes Care 2003; 26: 3320–3325.

  22. 22

    Palmentieri B, de Sio I, La Mura V, Masarone M, Vecchione R, Bruno S et al. The role of bright liver echo pattern on ultrasound B-mode examination in the diagnosis of liver steatosis. Dig Liver Dis 2006; 38: 485–489.

  23. 23

    Roldan-Valadez E, Favila R, Martinez-Lopez M, Uribe M, Mendez-Sanchez N . Imaging techniques for assessing hepatic fat content in nonalcoholic fatty liver disease. Ann Hepatol 2008; 7: 212–220.

  24. 24

    National Institutes Of Health. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults—the evidence report National Institutes Of Health. Obes Res 1998; 6 (Suppl 2): 51S–209S.

  25. 25

    Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA et al. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation 2005; 112: 2735–2752.

  26. 26

    Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo Jr JL et al. The seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure: the JNC 7 report. JAMA 2003; 289: 2560–2572.

  27. 27

    Graham I, Atar D, Borch-Johnsen K, Boysen G, Burell G, Cifkova R et al. European guidelines on cardiovascular disease prevention in clinical practice: executive summary. Fourth joint task force of the European Society Of Cardiology and other societies on cardiovascular disease prevention in clinical practice (constituted by representatives of nine societies and by invited experts). Eur J Cardiovasc Prev Rehabil 2007; 14 (Suppl 2): E1–E40.

  28. 28

    Expert Panel on Detection, Evaluation, Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the national cholesterol education program (NCEP). Expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 2001; 285: 2486–2497.

  29. 29

    Marchesini G, Bugianesi E, Forlani G, Cerrelli F, Lenzi M, Manini R et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003; 37: 917–923.

  30. 30

    Kotronen A, Westerbacka J, Bergholm R, Pietilainen KH, Yki-Jarvinen H . Liver fat in the metabolic syndrome. J Clin Endocrinol Metab 2007; 92: 3490–3497.

  31. 31

    Huang MA, Greenson JK, Chao C, Anderson L, Peterman D, Jacobson J et al. One-year intense nutritional counseling results in histological improvement in patients with non-alcoholic steatohepatitis: a pilot study. Am J Gastroenterol 2005; 100: 1072–1081.

  32. 32

    Park HS, Kim MW, Shin ES . Effect of weight control on hepatic abnormalities in obese patients with fatty liver. J Korean Med Sci 1995; 10: 414–421.

  33. 33

    Hickman IJ, Jonsson JR, Prins JB, Ash S, Purdie DM, Clouston AD et al. Modest weight loss and physical activity in overweight patients with chronic liver disease results in sustained improvements in alanine aminotransferase, fasting insulin, and quality of life. Gut 2004; 53: 413–419.

  34. 34

    Suzuki A, Lindor K, St Saver J, Lymp J, Mendes F, Muto A et al. Effect of changes on body weight and lifestyle in nonalcoholic fatty liver disease. J Hepatol 2005; 43: 1060–1066.

  35. 35

    Petersen KF, Dufour S, Befroy D, Lehrke M, Hendler RE, Shulman GI . Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes. Diabetes 2005; 54: 603–608.

  36. 36

    Kirpichnikov D, McFarlane SI, Sowers JR . Metformin: an update. Ann Intern Med 2002; 137: 25–33.

  37. 37

    Loomba R, Lutchman G, Kleiner DE, Ricks M, Feld JJ, Borg BB et al. Clinical trial: pilot study of metformin for the treatment of nonalcoholic steatohepatitis. Aliment Pharmacol Ther 2008; 29: 172–182.

  38. 38

    He L, Sabet A, Djedjos S, Miller R, Sun X, Hussain MA et al. Metformin and insulin suppress hepatic gluconeogenesis through phosphorylation of CREB binding protein. Cell 2009; 137: 635–646.

  39. 39

    Kohjima M, Higuchi N, Kato M, Kotoh K, Yoshimoto T, Fujino T et al. SREBP-1c, regulated by the insulin and AMPK signaling pathways, plays a role in nonalcoholic fatty liver disease. Int J Mol Med 2008; 21: 507–511.

  40. 40

    Musso G, Gambino R, Bo S, Uberti B, Biroli G, Pagano G et al. Should nonalcoholic fatty liver disease be included in the definition of metabolic syndrome? A cross-sectional comparison with Adult Treatment Panel III criteria in nonobese nondiabetic subjects. Diabetes Care 2008; 31: 562–568.

  41. 41

    Haukeland JW, Konopski Z, Eggesbo HB, von Volkmann HL, Raschpichler G, Bjoro K et al. Metformin in patients with non-alcoholic fatty liver disease: a randomized, controlled trial. Scand J Gastroenterol 2009; 44: 853–860.

  42. 42

    Mofrad P, Contos MJ, Haque M, Sargeant C, Fisher RA, Luketic VA et al. Clinical and histologic spectrum of nonalcoholic fatty liver disease associated with normal ALT values. Hepatology 2003; 37: 1286–1292.

  43. 43

    Tobari M, Hashimoto E, Yatsuji S, Torii N, Shiratori K . Imaging of nonalcoholic steatohepatitis: advantages and pitfalls of ultrasonography and computed tomography. Intern Med 2009; 48: 739–746.

  44. 44

    Lilly M, Godwin M . Treating prediabetes with metformin: systematic review and meta-analysis. Can Fam Physician 2009; 55: 363–369.

  45. 45

    Yamauchi T, Kamon J, Minokoshi Y, Ito Y, Waki H, Uchida S et al. Adiponectin stimulates glucose utilization and fatty-acid oxidation by activating AMP-activated protein kinase. Nat Med 2002; 8: 1288–1295.

  46. 46

    Wieckowska A, McCullough AJ, Feldstein AE . Noninvasive diagnosis and monitoring of nonalcoholic steatohepatitis: present and future. Hepatology 2007; 46: 582–589.

  47. 47

    Angulo P . Nonalcoholic fatty liver disease. N Engl J Med 2002; 346: 1221–1231.

  48. 48

    Gaidos JK, Hillner BE, Sanyal AJ . A decision analysis study of the value of a liver biopsy in nonalcoholic steatohepatitis. Liver Int 2008; 28: 650–658.

  49. 49

    Hamaguchi M, Kojima T, Takeda N, Nakagawa T, Taniguchi H, Fujii K et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med 2005; 143: 722–728.

  50. 50

    Castera L . Non-invasive diagnosis of steatosis and fibrosis. Diabetes Metab 2008; 34: 674–679.

  51. 51

    Dasarathy S, Dasarathy J, Khiyami A, Joseph R, Lopez R, McCullough AJ . Validity of real time ultrasound in the diagnosis of hepatic steatosis: a prospective study. J Hepatol 2009; 51: 1061–1067.

  52. 52

    Duseja A, Das A, Dhiman RK, Chawla YK, Thumburu KT, Bhadada S et al. Metformin is effective in achieving biochemical response in patients with nonalcoholic fatty liver disease (NAFLD) not responding to lifestyle interventions. Ann Hepatol 2007; 6: 222–226.

  53. 53

    de Oliveira CP, Stefano JT, de Siqueira ER, Silva LS, de Campos Mazo DF, Lima VM et al. Combination of N-acetylcysteine and metformin improves histological steatosis and fibrosis in patients with non-alcoholic steatohepatitis. Hepatol Res 2008; 38: 159–165.

  54. 54

    Idilman R, Mizrak D, Corapcioglu D, Bektas M, Doganay B, Sayki M et al. Clinical trial: insulin-sensitizing agents may reduce consequences of insulin resistance in individuals with non-alcoholic steatohepatitis. Aliment Pharmacol Ther 2008; 28: 200–208.

Download references

Author information

Correspondence to A Belfiore.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Garinis, G., Fruci, B., Mazza, A. et al. Metformin versus dietary treatment in nonalcoholic hepatic steatosis: a randomized study. Int J Obes 34, 1255–1264 (2010) doi:10.1038/ijo.2010.40

Download citation


  • dietary treatment
  • metformin
  • metabolic syndrome

Further reading