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

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT Rosiglitazone for Nonalcoholic Steatohepatitis: One-Year Results of the Randomized Placebo-Controlled Fatty Liver Improvement With Rosiglitazone Therapy (FLIRT) Trial

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

VLAD RATZIU,* PHILIPPE GIRAL,‡ SOPHIE JACQUEMINET,§ FRÉDERIC CHARLOTTE,储 AGNÈS HARTEMANN–HEURTIER,§ LAWRENCE SERFATY,¶ PHILIPPE PODEVIN,# JEAN–MARC LACORTE,** CAROLE BERNHARDT,* ERIC BRUCKERT,‡ ANDRÉ GRIMALDI,§ and THIERRY POYNARD* for the LIDO Study Group *Université Pierre et Marie Curie Paris VI, Assistance Publique-Hôpitaux de Paris, Service d’Hépatogastroentérologie, §d’Endocrinologie, 储de Diabétologie, ¶ Laboratoire d’Anatomie Pathologique, **Fédération de Biochimie, Hôpital Pitié-Salpêtrière, Paris, France; ‡INSERM UMRS_893, CdR Saint-Antoine, Paris, France; # Assistance Publique-Hôpitaux de Paris, Service d’Hépatologie, Hôpital Saint-Antoine, and **Service d’Hépatogastroentérologie, Hôpital Cochin, Paris, France.

Background & Aims: Nonalcoholic steatohepatitis (NASH) is a liver disease that complicates insulinresistant states. This trial tested the efficacy and safety of rosiglitazone, an insulin-sensitizing agent, in patients with NASH. Methods: Sixty-three patients with histologically proven NASH were randomly assigned to receive rosiglitazone (4 mg/day for the first month and 8 mg/day thereafter; n ⴝ 32) or placebo (n ⴝ 31) for 1 year. Liver biopsy was performed at the end of treatment. End points were improvement in the histologic score of steatosis, normalization of serum transaminase levels, and improvement in necroinflammation and fibrosis. Results: More patients treated with rosiglitazone than receiving placebo had improved steatosis (47% vs 16%; P ⴝ .014) and normalized transaminase levels (38% vs 7%; P ⴝ .005), although only half of patients responded. There was no improvement in other histologic lesions, including fibrosis, and a composite score of activity, the nonalcoholic fatty liver disease activity score. Improvement of steatosis correlated with reduction of transaminase levels (r ⴝ 0.36; P < .005), improvement in insulin sensitivity (r ⴝ 0.34; P ⴝ .008), and increase in adiponectin levels (r ⴝ ⴚ0.54; P < .01) but not with weight variations. Independent predictors of response were rosiglitazone treatment, the absence of diabetes, and massive steatosis. Weight gain was the main adverse effect (mean gain of 1.5 kg in the rosiglitazone group vs ⴚ1 kg in the placebo group; P < .01), and painful swollen legs was the main reason for dose reduction/discontinuation. Serum hemoglobin level was slightly but significantly reduced. There was no hepatic toxicity. Conclusions: In patients with NASH, rosiglitazone improves steatosis and transaminase levels despite weight gain, an effect related to an improvement in insulin sen-

sitivity. However, there is no improvement in other parameters of liver injury.

N

onalcoholic steatohepatitis (NASH) is a liver disease complicating insulin resistance,1 which is highly prevalent due to the fast-growing rate of progression of obesity and diabetes. NASH results from excessive hepatic fat deposition and is characterized by the presence of steatosis coexisting with hepatic inflammation and hepatocellular injury. Liver fibrosis is the main hepatic complication in this condition,2 and 10%–20% of patients with NASH will develop cirrhosis.3 Patients with NASH have increased liver-related mortality4,5 and NASH-induced cirrhosis can result in end-stage liver disease,6,7 including the development of hepatocellular carcinoma.6,8 Insulin resistance is the major driving force behind excessive fat accumulation in the liver and development of steatohepatitis and its progression.9 Therefore, correction of insulin resistance is a relevant therapeutic target in this condition. Diet-induced weight loss, lifestyle modifications including physical exercise, and qualitative changes in diet can achieve this objective in some patients. However, only a small percentage of patients with NASH can implement these measures efficiently. For instance, only 6% of Japanese employees diagnosed with nonalcoholic fatty liver disease were able to lose ⬎5% of body weight during a 1-year follow-up.10 Even within clinical trials, there is a 30%– 40% dropout rate for various dietary regimens or nutritional counseling approaches.11 Moreover, the feasibility of regular physical training as a Abbreviations used in this paper: EOT, end of treatment; HOMA, homeostasis model assessment; NAS, nonalcoholic fatty liver disease activity score; NASH, nonalcoholic steatohepatitis; QUICKI, quantitative insulin-sensitivity check index. © 2008 by the AGA Institute 0016-5085/08/$34.00 doi:10.1053/j.gastro.2008.03.078

long-term treatment is severely hampered by comorbidities related to the metabolic syndrome or by low adherence rates.12 Therefore, insulin-sensitizing agents could be of therapeutic benefit in this condition. Two uncontrolled studies with rosiglitazone13 and pioglitazone14 have suggested biochemical and histologic improvement in patients with NASH. The only placebocontrolled trial available showed encouraging results with pioglitazone, both for metabolic control and improvement of the liver disease.15 However, treatment duration in this study was only 6 months, and some reports with other insulin-sensitizing agents such as metformin have shown a breakthrough after an initial liver biochemical response.16 Importantly, none of the available trials identify nonresponders to glitazones. We performed the current trial to assess whether rosiglitazone therapy improves liver disease in unselected patients with NASH, assess if this improvement is associated with reduction of insulin resistance, and identify a potential patient profile of response to glitazones.

Patients and Methods Study Design This was a randomized, double-blind, placebocontrolled trial in patients with histologically proven NASH. Patients were included between January 2003 and November 2004 in a single institution. Randomization (presealed envelopes) was conducted by blocks of 4 and stratified on metformin use. Patients received either placebo or rosiglitazone for 12 months and were followed up for 4 months after the end of treatment (EOT) until the end of follow-up. Rosiglitazone (GlaxoSmithKline, Marly le Roi, France) was administered at a dosage of 4 mg daily for the first month, 8 mg daily until EOT, and then discontinued. Patients in both groups were instructed to lose weight if they were obese or overweight, to follow a healthy diet, and to exercise at least twice a week. However, no specialized nutritional counseling was implemented and no information on changes in diet or physical activity was collected throughout the study. During treatment, patients were followed up at months 1 and 2 and then every 2 months until EOT when liver biopsy was performed. After the end of follow-up, all patients were asked to participate in a 2-year, open-label, extension trial of rosiglitazone, irrespective of what they received during the 1-year double-blinded phase of the trial. This extension trial will be reported on subsequently. All patients gave written informed consent, and the study protocol was approved by the Ethics Committee of the Pitié-Salpêtrière Hospital.

Inclusion Criteria Patients aged 18 –75 years with a histologic diagnosis of NASH and elevated alanine aminotransferase (ALT) level were included. The histologic diagnosis of

ROSIGLITAZONE FOR NONALCOHOLIC STEATOHEPATITIS

101

NASH was defined as steatosis ⬎20% and the presence of either hepatocyte ballooning or intralobular hepatocyte necrosis, all signs with centrolobular accentuation, on a liver biopsy performed within 12 months before inclusion. Elevated ALT level was defined as a baseline level ⬎28 IU/L in women and 35 IU/L in men and at least 2 other abnormal values in the 6 months before baseline.

Exclusion Criteria Exclusion criteria were the presence of bland steatosis on liver biopsy or steatosis with nonspecific inflammation; daily alcohol consumption ⬎30 g in men and 20 g in women whether current or in the past; any cause of liver disease other than NASH,17 including suspicion of drug-induced liver injury (introduction of a new drug in the past 6 months without prior documentation of elevated ALT level); treatment with insulin for diabetes or with ursodeoxycholic acid; cardiac insufficiency (New York Heart Association Class ⬎1); current or past treatment with drugs that can induce steatohepatitis; neoplastic disease; Child class B or C cirrhosis; pregnancy; organ transplantation; hemoglobin level ⬍10 g/dL; polymorphonuclear count ⬍750/mm3; and platelet count ⬍50,000/mm3.

Efficacy End Points The primary end point was a reduction in steatosis ⬎30% between baseline and EOT or disappearance of steatosis at EOT. The 2 secondary end points were (1) normalization of serum ALT level at EOT and (2) improvement in activity grade and/or liver fibrosis between baseline and EOT.

Analytical Methods Pathology. Liver samples were paraffin embedded and stained with H&E-safran and Sirius red. Steatosis was semiquantitatively assessed both as the percent of hepatocytes involved within a lobule (0%–100%, steatosis score) and by using a 4-grade classification modified from Kleiner et al18: 0, absent; 1, ⬍5%; 2, 5%–33%; 3, 33%– 66%; 4, ⬎66%. Staging and grading were performed according to Brunt et al19: F0, absence of any fibrosis; F1, isolated perisinusoidal or portal/periportal fibrosis; F2, perisinusoidal and portal/periportal fibrosis; F3, bridging fibrosis; F4, cirrhosis. The nonalcoholic fatty liver disease activity score (NAS) was calculated according to Kleiner et al.18 The effect of treatment was assessed on both overall fibrosis and, specifically, perisinusoidal fibrosis. Two different pathologic analyses were performed by a single pathologist who was unaware of the clinical data or treatment assignation. The first was a reading of all slides in random order and establishing scores for all elementary histologic lesions at EOT and baseline. The second was a ranked pair assessment in which, for each patient, pretreatment and EOT liver biopsy specimens were compared and the severity of each histologic lesion was delineated as worse, unchanged, or improved.

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

July 2008

102

RATZIU ET AL

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Clinical and biochemical data. Diabetes was considered present if patients were taking medications for diabetes (metformin or sulfamides) or if the fasting serum glucose level was ⬎7 mmol/L in untreated patients. Arterial hypertension was considered present if patients were taking antihypertensive drugs. At baseline, month 4, month 8, and EOT and end of follow-up, biochemistry results were obtained from a centralized laboratory after a 12-hour overnight fast. The homeostasis model assessment (HOMA) index, a marker of insulin resistance, was calculated as follows: Glucose (mmol/L) ⫻ Insulin (␮UI/ L)/22.5. The quantitative insulin-sensitivity check index (QUICKI), a marker of insulin sensitivity, was calculated as follows: 1/log (HOMA ⫻ 405). Serum leptin and adiponectin levels were determined using a radioimmunoassay kit from Linco Research (St Charles, MO). The sensitivity of these assays was 0.5 ng/mL and 1 ng/mL, respectively.

Statistical Methods The proportion of patients with a histologic response (defined as the achievement of the primary end point) and a biochemical response (achievement of the secondary end point on transaminases) was compared between the 2 groups. The decline in transaminase values during treatment for each patient was expressed as the percentage from the baseline value of ALT values at each time point. For the histologic ranked assessment analyses, a 2-class comparison was performed using the Fisher exact test between the proportion of patients with progression (worsening) of a histologic feature and that of patients without progression (grouping those with no change and improvement). Numerical variables were compared using analysis of variance for those normally distributed and nonparametric tests such as the Mann– Whitney U test for those without a normal distribution. Proportions were compared using the Fisher exact test. Independent factors for histologic response were tested by logistic regression. Correlations between efficacy end points and numerical clinical and biologic variables were tested by the Spearman rank correlation matrix. The assumption for sample size calculations was that the primary end point would be met in 50% of patients treated with the active drug and 10% of patients receiving placebo. The inclusion of 29 patients per arm was deemed necessary for rejecting the null hypothesis with 80% power and a type I error rate of 0.05. The total number of included patients was set at 32 per arm to account for a 10% lost to follow-up rate.

Results Study Population Thirty-two patients were randomized in each arm. One patient from the placebo arm retracted his consent after randomization and was not enrolled. Another pa-

GASTROENTEROLOGY Vol. 135, No. 1

tient from the rosiglitazone arm decided to stop treatment after 10 months because of side effects. Because this patient received more than 80% of the assigned amount and duration of therapy, he was included in the analysis as a nonresponder both for the histologic and biochemical efficacy criteria. Thus, we report on the results of 32 patients in the rosiglitazone arm and 31 patients in the placebo arm. Characteristics of study participants are listed in Tables 1 and 2. Most patients were asymptomatic except for asthenia, which was present at baseline in 20 rosiglitazone-treated patients and 16 patients receiving placebo. The delay between liver biopsy and inclusion was longer than 6 months in 6 rosiglitazone-treated patients and 7 patients receiving placebo (P ⫽ .76). The proportion of patients with ballooning, lobular inflammation, perisinusoidal fibrosis, or significant fibrosis (stages ⬎F2) was not different between groups.

Efficacy Improvement in steatosis. A ⬎30% reduction in steatosis occurred more frequently in the rosiglitazone group (47%; 15/32) than in the placebo group (16%; 5/31; P ⫽ .014). Table 3 shows the mean reduction in the steatosis histologic score, and Figure 1 shows the distribution of loss/gain of steatosis in each treatment arm. In the rosiglitazone arm, only 1 patient progressed (by 50%) and 16 had a minor variation (unchanged or more or less 20%), while among the 15 histologic responders 6 had a reduction of 30%, 6 of 40%, and 3 of ⬎60%. In the placebo arm, 1 patient lost 50%, 1 lost 40%, and 3 lost 30%. Improvement in transaminase levels. An EOT normal ALT level was achieved in 12 rosiglitazone-treated patients (38%) and 2 patients receiving placebo (7%; P ⫽ .005; Figure 2). Figure 2A shows that at all time points during treatment, the decrease in ALT values was higher for rosiglitazone-treated patients than for those receiving placebo. An additional 8 rosiglitazone-treated patients had normal transaminase levels at some point during treatment but abnormal levels at month 12, while this was the case for only 3 patients in the placebo group. Thus, overall, 20 patients in the rosiglitazone group and 5 patients in the placebo group normalized transaminase levels during treatment (P ⬍ .001). The reduction in ALT levels was documented early on during therapy, averaged 30% from baseline, and was maintained throughout treatment. In contrast, in the placebo group, mean variation in ALT levels was only ⫾10%. Other histologic end points. There was no significant improvement in fibrosis, hepatocyte ballooning, or lobular inflammation/necrosis in rosiglitazone-treated patients (Table 3). There was no significant difference in the mean variation of the composite NAS. However, the ranked assessment of paired pretreatment and EOT liver biopsy specimens revealed several changes suggesting a favorable effect of rosiglitazone (Table 4). The proportion of patients

July 2008

ROSIGLITAZONE FOR NONALCOHOLIC STEATOHEPATITIS

103

Variable

All (n ⫽ 63)

Rosiglitazone (n ⫽ 32)

Placebo (n ⫽ 31)

P value

Age (y), mean (SD) Male sex, n (%) Body mass index (kg/m2), mean (SD) Diabetes, n (%) Metformin treated, n Sulfamide treated, n Untreated, n Arterial hypertension, n (%) Waist circumference (cm), mean (SD) Hemoglobin level (g/dL), mean (SD) AST level (IU/L), mean (SD) ALT level (IU/L), mean (SD) ␥-glutamyltransferase level (IU/L), mean (SD) Bilirubin level (␮mol/L), median (IQR) Serum glucose level (mmol/L), median (IQR) Hemoglobin A1c level (%), median (IQR) Insulin level (IU/mL), median (IQR)a HOMA, median (IQR)a QUICKI, median (IQR)a Leptin level (ng/mL), median (IQR)b Adiponectin level (␮g/mL), median (IQR)c Triglycerides level (mmol/L), median (IQR) High-density lipoprotein cholesterol level (mmol/L), median (IQR) Metabolic syndrome (%) Serum glucose level ⬎5.5 mmol/L (%) Triglycerides level ⬎1.7 mmol/L (%) Waist circumference, n (%)d Arterial hypertension High-density lipoprotein cholesterol levele Delay between biopsy and inclusion (mo), median (IQR)

53.6 (10.9) 37 (59) 31 (5.2) 20 (32) 16 10 3 22 (35) 105 (12) 14.2 (1.31) 54 (38) 77 (39) 92 (76) 10 (8) 5.3 (1.8) 5.6 (0.9) 16 (13.6) 4.39 (4.3) 0.31 (0.04) 15.1 (11) 5.56 (1.1) 1.45 (1) 1.23 (0.45) 31 (49) 27 (43) 23 (37) 49 (78) 23 (37) 27 (43) 1.6 (4.3)

53.1 (11.5) 19 (59) 31.5 (6) 9 (28) 7 7 1 10 (31) 106 (13) 14.1 (1.4) 46 (26) 69 (40) 71 (48) 9.5 (4.8) 5.3 (1.9) 5.6 (1.03) 17.2 (14.1) 4.57 (4.66) 0.31 (0.04) 15.2 (14.6) 5.66 (1.11) 1.34 (0.9) 1.22 (0.42) 14 (44) 13 (41) 10 (31) 22 (69) 10 (31) 15 (47) 1.55 (4.1)

54.1 (10.4) 18 (58) 30.5 (4.4) 11 (35) 9 3 2 13 (42) 104 (10) 14.3 (1.2) 61 (46) 84 (38) 113 (93) 10 (11) 5.4 (2.2) 5.6 (0.9) 15.2 (11.6) 4.21 (4.32) 0.31 (0.04) 14.1 (11.4) 5.47 (1) 1.65 (1.21) 1.27 (0.49) 17 (57) 14 (45) 13 (43) 27 (87) 13 (42) 12 (39) 1.9 (4.4)

.71 1.00 .44 .60

.44 .53 .57 .12 .14 .08 .39 .58 .77 .50 .70 .70 .80 .35 .74 .69 .45 .80 .44 .13 .44 .61 .84

IQR, interquartile range. nondiabetic patients only. bAvailable for 51 patients only (23 rosiglitazone and 28 placebo). cAvailable for 52 patients only (24 rosiglitazone and 28 placebo). d⬎88 cm for women and ⬎102 cm for men. e⬍1 mmol/L for men and ⬍1.3 mmol/L for women. aIn

experiencing progression of hepatocyte ballooning, portal inflammation, and overall fibrosis was significantly lower with rosiglitazone than with placebo. A similar trend was observed for perisinusoidal fibrosis (P ⫽ .1).

Improvement in Insulin Sensitivity and Histologic Correlates Rosiglitazone but not placebo significantly improved serum glucose levels, both in the whole cohort and in diabetic patients, as well as surrogate markers of insulin resistance, hyperinsulinemia, HOMA, and QUICKI (Table 3). There was a significant correlation between reduction in the histologic score of steatosis and reduction in transaminase levels (r ⫽ 0.36, P ⬍ .005) and fasting insulin levels (r ⫽ 0.34, P ⫽ .008) and improvement in HOMA (r ⫽ 0.34, P ⫽ .008) and QUICKI indices (r ⫽ ⫺0.27, P ⫽ .035) but no correlation with variations in weight or waist circumference during treatment (r ⫽ 0.04, P ⫽ NS). Interestingly, the reduction in insulin levels was significantly higher in responders to rosiglita-

zone as compared with nonresponders (⫺6.9 vs ⫺3.4 IU/mL; P ⫽ .02), indicating that the loss of liver fat was associated with a more marked improvement in insulin sensitivity. The same trend, although not significant, was noted for serum glucose and HOMA levels.

Adipocytokines and Histologic Response Serum adiponectin but not leptin levels increased during rosiglitazone treatment 1.91-fold versus 0.96-fold in patients receiving placebo (P ⫽ .04) (Figure 3A). This increase was greater in patients who had improved steatosis than in those who did not (2.35-fold vs 1.21-fold, respectively; P ⫽ .04). Changes in serum adiponectin levels correlated negatively with reduction in steatosis (r ⫽ ⫺0.54, P ⬍ .01; Figure 3B) but not with changes in body weight.

Predictors of Response Responders had lower median ␥-glutamyltransferase levels (48 vs 71 UI/L; P ⫽ .03), lower median serum glucose levels (5.1 vs 5.4 mmol/L; P ⫽ .02), less frequently

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Table 1. Clinical and Biological Characteristics at Baseline

104

RATZIU ET AL

GASTROENTEROLOGY Vol. 135, No. 1

Table 2. Histologic Characteristics at Baseline

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Size/portal spaces, median (IQR) Steatosis (%), median (IQR) Steatosis grade, n (%) 0 : 0% 1 : ⬍5% 2 : 5%–33% 3 : 33%–66% 4 : ⬎66% Hepatocyte ballooning None Mild Moderate Lobular necrosis and inflammation None Mild Moderate Marked Mallory bodies Polymorphonuclear infiltrate Perisinusoidal fibrosis None Minimal Moderate Fibrosis (stage) Absent (F0) Perisinusoidal only (F1) Portal/periportal only (F1) Perisinusoidal and portal/periportal (F2) Bridging (F3) Cirrhosis (F4) NAS, median (IQR)

All (n ⫽ 63)

Rosiglitazone (n ⫽ 32)

Placebo (n ⫽ 31)

P value

23 (12)/20 (13) 50 (40)

25 (14)/20 (15) 50 (40)

20 (10)/18 (15) 50 (40)

.26/.6 .88 .81

0 0 21 (33) 22 (35) 20 (32)

11 (34) 10 (31) 11 (34)

10 (32) 12 (39) 9 (29)

25 (40) 24 (38) 14 (22)

13 (41) 14 (44) 5 (16)

12 (39) 10 (22) 9 (29)

9 37 16 1 13 (21) 1 (2)

4 (13) 23 (72) 4 (13) 1 (3) 5 (16) 0

5 (16) 14 (45) 12 (39) 0 8 (26) 1 (2)

5 (8) 37 (59) 21 (33)

4 (13) 21 (66) 7 (22)

1 (3) 16 (52) 14 (45)

4 (6) 13 (21) 1 (2) 25 (40) 18 (29) 2 (3) 4 (2)

3 (9) 8 (25) 1 (3) 15 (47) 4 (13) 1 (3) 4 (2.8)

1 (3) 5 (16) 0 10 (32) 14 (45) 1 (3) 3 (2)

.40

.06

.44 .31 .09

.10

.08

IQR, interquartile range.

diabetes (10% vs 42%; P ⬍ .02), and higher amounts of steatosis (65% vs 45%; P ⬍ .0001) than nonresponders. There was a trend toward a higher level of serum adiponectin (11.3 vs 6.5 ␮g/mL, respectively; P ⫽ .06). No other baseline factors were associated with response. We next analyzed the association between steatosis response and improvement in insulin sensitivity. In the rosiglitazone arm, a reduction of 0.5 or more in HOMA index was seen in 14 of the 15 responders (93%) and 10 of the 17 nonresponders (59%; P ⫽ .02). In the placebo arm, the respective proportions were 20% and 27% (P ⫽ NS). Different kinetic profiles were seen according to the steatosis response. In responders, there was a more marked and prolonged decrease in mean ALT levels than in nonresponders at months 8 and 12 of treatment (P ⬍ .006 and P ⬍ .05, respectively; Figure 2B). There was a trend toward a higher reduction in ␥-glutamyltransferase levels. There was no significant difference in mean weight variation between EOT and baseline in responders (⫺0.3% of baseline weight or ⫺0.05 kg) and nonresponders (⫹0.3% of baseline weight or 0.3 kg). Logistic regression analysis identified the following independent predictors of histologic response: drug treatment with rosiglitazone (odds ratio, 4.93; 95% con-

fidence interval, 1.3–19.8; P ⫽ .024), absence of diabetes (odds ratio, 0.14; 95% confidence interval, 0.02– 0.79; P ⫽ .026), and a high histologic score of steatosis (⬎66%; odds ratio, 8.98; 95% confidence interval, 2.12–38; P ⫽ .003).

Durability of Response In the rosiglitazone group, 4 months after EOT, the following variables returned to baseline levels: serum transaminases, glucose, glycosylated hemoglobin, cholesterol and low-density lipoprotein cholesterol, and hemoglobin. In contrast, there was a still significant improvement in insulin levels and HOMA values 4 months after EOT.

Safety Analysis There were no new or unexpected side effects associated with rosiglitazone (Table 5). No cardiovascular events occurred. A few patients (n ⫽ 3), but all in the rosiglitazone group, experienced painful swollen legs while others reported muscular cramps (without concomitant use of statins or fibrates). There was a slight but significant increase in serum cholesterol and low-density

July 2008

ROSIGLITAZONE FOR NONALCOHOLIC STEATOHEPATITIS

105

Table 3. Changes in Metabolic and Histologic Parameters Between Baseline and EOT ⫺0.93 (1.72) ⫺0.18 (0.98) ⫺5.5 (9.1) ⫺1.4 (2.1) 0.026 (0.022) 0.45 (0.67) 0.12 (1) 0.05 (0.27) 0.41 (0.79) 1.91 (1.39) 0.96 (0.2) ⫺20 (26) 0.03 (0.95) ⫺0.03 (0.54) 0.13 (0.71) ⫺0.09 (0.73) ⫺1 (4)

Placebo 0.55 (1.74) 0.17 (0.53) 2.5 (12.8) 0.59 (4.5) ⫺0.007 (0.036) ⫺0.24 (0.75) 0.2 (1.26) 0.01 (0.27) ⫺0.32 (0.87) 0.94 (0.23) 1.07 (0.2) ⫺5 (23) ⫺0.18 (1.14) ⫺0.06 (0.63) 0.23 (0.8) ⫺0.13 (0.81) 0 (4)

P value .001 .078 ⬍.001 ⬍.001 ⬍.001 .0013 .94 .72 .0014 .04 .21 .02 .43 .83 .61 .86 .60

NOTE. Values are expressed as mean (SD) unless otherwise stated. the whole cohort (diabetic and nondiabetic patients) the values rosiglitazone- and placebo-treated patients were as follows, respectively: ⫺4.8 (9.2) and 2.1 (14.8) for insulin (P ⬍ .002); ⫺1.4 (2.6) and 0.61 (5.5) for HOMA index (P ⬍ .001); and 0.02 (0.03) and ⫺0.007 (0.03) for QUICKI index (P ⬍ .001). bPaired measurements at baseline and EOT were available in 22 patients (12 rosiglitazone and 10 placebo) for adiponectin and in 21 patients (11 rosiglitazone and 10 placebo) for leptin. aIn

lipoprotein cholesterol levels in rosiglitazone-treated patients (Table 3). Dose reduction due to side effects was necessary in 5 rosiglitazone-treated patients (for headache in 1 patient, an episode of malaise in another patient, and swollen legs in 3 patients, including 2 with leg pain and difficulties walking) and 1 patient receiving placebo (headache with nausea). One patient from the rosiglitazone group de-

Figure 1. Difference in the histologic score for steatosis between EOT and baseline liver biopsy specimens in both rosiglitazone (black circles) and placebo (white circles). Steatosis is expressed as the percentage of hepatocytes containing fat droplets. Responders for steatosis (ie, patients having lost ⬎30% during treatment) are those on or below the dashed line.

cided to stop treatment at month 10 because of pain associated with swollen legs.

Discussion In this randomized placebo-controlled study, a 1-year course of rosiglitazone significantly improved steatosis and normalized transaminase levels but did not impact on any other liver histologic lesions in patients with NASH. Almost half of treated patients experienced a considerable reduction in liver fat, most of these with a decrease of 40% or more. Using stringent criteria,20 40% of rosiglitazone-treated patients normalized ALT levels at EOT. The reduction in transaminase levels occurred rapidly, in the first 4 months, and was maintained throughout the treatment period, which is consistent with data from uncontrolled trials.13,14 However, shortly after drug discontinuation, transaminases returned to baseline levels and part of the gain on insulin sensitivity was lost. This short-lived durability was reported in an uncontrolled trial of pioglitazone, along with a loss of the histologic benefit of treatment 1 year after drug discontinuation.21 Therefore, long-term treatment with this class of drugs will be needed and safety will become an important issue. In this trial, rosiglitazone was safe in most patients with no cardiovascular events or hepatotoxicity, although some patients experienced an asymptomatic low-grade reduction in hemoglobin levels as well as weight gain. In some patients, however, dose reduction or treatment discontinuation was necessary mainly because of painful swollen legs.

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Rosiglitazone Serum glucose level (mmol/L) Hemoglobin A1c level (%) Fasting insulin level in nondiabetic patients (␮UI/mL), median (interquartile range)a HOMA in nondiabetic patients, median (interquartile range)a QUICKI in nondiabetic patients, median (interquartile range)a Cholesterol level (mmol/L) Triglycerides level (mmol/L) High-density lipoprotein cholesterol level (mmol/L) Low-density lipoprotein cholesterol level (mmol/L) Adiponectin level (␮g/mL)b Leptin level (ng/mL)b Steatosis, mean reduction (%) Fibrosis Perisinusoidal fibrosis Hepatocyte ballooning Lobular necrosis/inflammation NAS

106

RATZIU ET AL

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Figure 2. ALT course during therapy. Results are expressed as means (⫾SD) of fold elevation over baseline values at each time point during treatment and at EOT (month 12) and end of follow-up (month 16). (A) Rosiglitazone (black circles) and placebo (white circles); mean ALT values were significantly lower in the rosiglitazone group at months 4, 8, and 12. (B) Responders for steatosis (black circles) and nonresponders (white circles); mean ALT values were significantly lower in responders at months 8 and 12.

Contrary to other reports, there was no improvement in liver necroinflammation, the overall NAS, or liver fibrosis. A placebo-controlled trial of pioglitazone of smaller sample size and shorter duration15 reported a significant improvement in the necroinflammatory activity and a nearly significant trend for fibrosis improvement.15 Whether these discrepant results may be explained by differences in the study population or by differences between the 2 drugs cannot be anticipated at this point. Both glitazones are equally effective in lowering serum glucose levels, but divergent actions on the lipid profile (serum triglycerides and low-density lipoprotein cholesterol) exist. Whether this or other yet-unidentified mechanisms could impact on liver histology needs to be studied. Although fibrosis regression has been doc-

GASTROENTEROLOGY Vol. 135, No. 1

umented,22 it is a slow process conditioned by prior disappearance of causative lesions, in this case, steatosis and necroinflammation.17,23 Therefore, it can hardly be reliably shown in short and small-sized studies with grossly semiquantitative histologic scores. The ranked assessment of pretreatment and EOT biopsy specimens could be a more sensitive method for detecting early histologic changes. Data provided here suggest that glitazones could improve the whole spectrum of liver injury, although this needs to be confirmed in larger and longer trials and might take longer than previous studies suggested.15 Insulin resistance is a central mechanism in the pathogenesis of NASH. Clinical data and clamp studies have shown a gradual loss of insulin sensitivity during the transition from normal liver to steatosis and steatohepatitis,1,9 while, in the latter stage, insulin resistance could also be a main determinant of fibrosis progression. In the present study, patients treated with rosiglitazone but not placebo experienced a significant reduction in fasting glucose level, glycosylated hemoglobin level, and surrogate serum markers of insulin resistance (hyperinsulinemia, HOMA index), which were strongly correlated with improvement in steatosis. Moreover, rosiglitazone also significantly increased serum levels of adiponectin, a known effect of glitazones in diabetic patients,24 which is mediated both at the transcriptional25 and posttranscriptional levels.26 Adiponectin has strong antisteatogenic and antidiabetogenic properties,27 and the ability of glitazones to increase adiponectin expression is a theoretical advantage over metformin, which lacks this effect, at least in the short run.28 It has been shown that the increase in adiponectin levels and the improvement in insulin resistance observed with glitazones are causally related through exclusively hepatic effects, namely a decrease in hepatic glucose production and an increase in hepatic adenosine monophosphate–activated protein kinase.29 In this study, the increase in adiponectin levels was significantly correlated with the loss of liver fat. Taken together, these data provide a biological framework for the beneficial effect of rosiglitazone in NASH by suggesting that the improvement in steatosis is mediated by an insulin-sensitizing effect of the drug. Notably, rosiglitazone improved steatosis and transaminase levels despite weight gain in 40% of responders. This is in line with the previous demonstration that weight gain occurring during glitazone therapy has no deleterious effect on insulin sensitivity, because it is mainly due to an expansion of peripheral and not central fat depots, which are less metabolically active. Reduction in liver fat has important clinical implications, both hepatic and extrahepatic. Excess liver fat is the prerequisite in the sequential multistep process that leads to NASH and its progression. Indeed, steatosis fuels lipid peroxidation and oxidative stress, which contributes to the necroinflammatory lesions associated with steato-

July 2008

ROSIGLITAZONE FOR NONALCOHOLIC STEATOHEPATITIS

107

Table 4. Ranked Assessment of Pretreatment and EOT Liver Biopsy Pairs

Hepatocyte ballooning Lobular necrosis/inflammation Portal inflammation Fibrosis Perisinusoidal fibrosis Steatosis NAS

Placebo (n ⫽ 31)

Worsened

Unchanged

Improved

Worsened

Unchanged

Improved

1a (3) 3 (9) 0b 1c (3) 1 (3) 1 (3) 11 (33)

23 (72) 13 (41) 28 (88) 26 (81) 27 (84) 16 (50) 2 (6)

8 (25) 16 (50) 4 (13) 5 (16) 4 (13) 15d (47) 19 (60)

7a (16) 5 (16) 5b (16) 6c (19) 5 (16) 3 (10) 15 (48)

17 (48) 15 (48) 22 (71) 20 (65) 22 (71) 23 (74) 3 (10)

7 (36) 11 (36) 4 (13) 5 (16) 4 (13) 5d (16) 13 (42)

NOTE. Values are expressed as n (%). ⫽ .026. bP ⫽ .02. cP ⫽ .05. dP ⫽ .01. aP

hepatitis. Conversely, reversal of NASH after weight loss is always associated with a massive reduction in liver fat.30 Steatosis also increases the vulnerability of the liver to noxious challenges such as endotoxin31 and is associated with increased hepatic expression of proinflammatory cytokines,32–34 mitochondrial dysfunction,35 hepatic alterations in innate immunity,36 increased apoptosis,37 and decreased energy stores.38 Moreover, hepatic stellate cell activation by reactive oxygen species and lipid peroxidation products is an essential trigger for fibrogenesis. This strongly suggests that reduction in liver fat will result in improvement of steatohepatitis and NASH-induced fibrosis, although this needs to be specifically investigated in larger trials with longer treatment duration. Extrahepatic consequences of liver fat reduction are also meaningful.39 It has been shown that liver fat induces hepatic insulin resistance independent of central fat or free fatty acid flux in the liver.40 Excess liver fat impairs insulin signaling, and this induces the expression of clinical features of the metabolic syndrome in nondiabetic patients.41 Steatosis is also a predictor of type 2 diabetes, independent of major confounders. Finally, in diabetic patients, steatosis is an independent predictor of insulin requirements.42 Therefore, we believe that improvement of steatosis has significant clinical relevance in patients with NASH. Although rosiglitazone was more effective than placebo, an important finding of this study is that half of treated patients were nonresponders on steatosis. None of the previous trials in NASH13–15 provide data on nonresponders to glitazones. While the vast majority of histologic responders in the rosiglitazone arm had an improvement in insulin resistance (93%), so did a significantly smaller, although still substantial, proportion of nonresponders (59%). Because of the small sample size of the current trial, data interpretation should be prudent, but 2 main implications emerge. On a cognitive level, this suggests that mechanisms other than insulin sensitization might be involved in the improvement of

liver injury by thiazolidinediones. Animal models have shown that, independent of its insulin-sensitizing effects, rosiglitazone improves liver injury and reduces liver fibrosis in both toxic and cholestatic rodent models.43 Another implication, on a clinical level, is that in addition to reversing insulin resistance, other therapeutic approaches might be required in some patients. Agents with potent antioxidant, antiapoptotic, or hepatoprotective properties44 might open the way to future individualized therapies in patients with NASH, and it is improbable that a single molecule will be the answer to all patients with this disease. Because therapy for NASH is likely to be prolonged, the early identification of predictors of nonresponse either at baseline or early on during therapy is important. At baseline, responders had lower ␥-glutamyltransferase levels and less frequently diabetes. Serum ␥-glutamyltransferase levels are positively correlated with the HOMA index45 and with the risk of incident type 2 diabetes.46 Therefore, it is possible that the beneficial effect of rosiglitazone is optimal at earlier stages of insulin resistance. Along the same lines, an interesting finding was the higher baseline adiponectin levels in histologic responders. Recently, it has been shown that rodents with a low adiponectin level have a blunted response to glitazones, which suggests that adiponectin is an important mediator of the insulin-sensitizing effects of glitazones.47 In the current study, responders had a more marked decrease in ALT level than nonresponders, although this was best evidenced at the eighth month of treatment, a late time point for the prediction of response to therapy. Future studies in NASH should identify clinically useful early predictors of the metabolic and hepatic responses. Because this and other studies21 have shown that once glitazones have been stopped the durability of the response is very limited, long-term therapy may be anticipated as long as the underlying insulin resistance is not corrected by dietary or lifestyle modifications. Therefore,

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Rosiglitazone (n ⫽ 32)

108

RATZIU ET AL

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

any potential hepatic benefits of this class of drugs need to be weighed against long-term safety issues, foremost of which is cardiovascular toxicity. Significant concerns have been raised about the potential of both glitazones to induce congestive heart failure, especially in patients with this preexisting condition. Despite methodological limitations of available studies, a strong signal suggestive of an increased cardiac ischemic risk with rosiglitazone48 was acknowledged through a black box warning for this drug. Bone loss is another adverse long-term effect for both glitazones, although whether this translates to an increased risk of fractures is unknown.49 This study has several strengths and also some limitations. Although randomized, there were some nonsignificant differences in baseline biochemical and histologic parameters whose impact on the therapeutic response cannot be excluded. However, contrary to other reports,

GASTROENTEROLOGY Vol. 135, No. 1

Table 5. Side Effects Rosiglitazone Weight change from baseline ⫹1.5 (5.2) (kg), mean (SD) Weight gain ⬎3 kg (%) 10 (31) Weight gain ⬎5% (%) 6 (19) Swollen legsa 10 (31) Muscular cramps, n (%)b 5 (16) Asthenia, n (%) 6 (19) Congestive heart failure 0 Hemoglobin reduction (g/dL), ⫺0.65 (0.77) mean (SD)c Clinically significant, n (%)d 3 (9) Hepatotoxicitye 0

Placebo ⫺1 (3.5) 4 (13) 2 (6.5) 8 (26) 1 (3) 11 (35) 0 ⫺0.14 (0.75) 1 (3) 0

P value .03 .13 .73 .78 .20 .16 .01 .60

NOTE. Other side effects were nausea, headache, abdominal bloating, and diarrhea with no differences between groups. aAt any time point during the study period. bNot explained by the use of statins or fibrates. cBetween end of treatment and baseline. dEOT hemogloblin level ⬍12 g/dL together with a decrease of ⬎1 g/dL. eElevation of ALT level ⬎3 times the baseline value at any time point during the study.

the placebo group showed a very low rate of histologic and biochemical response.50 We believe this strengthens the perception of NASH as a real and constant disease process and should encourage the development of trials with other therapeutic agents. A shortcoming of this study is the lack of specific nutritional intervention and of a systematic assessment of changes in diet or physical activity. Nonetheless, these patients had long-standing liver biochemistry and metabolic abnormalities and most had already tried, unsuccessfully, lifestyle changes including weight loss diets. Although this is the largest and longest-duration randomized trial of a glitazone reported so far in NASH, it is still a pilot study. Future larger trials testing more prolonged treatment are needed to establish longer-term efficacy. References

Figure 3. Changes in serum adiponectin levels during treatment. (A) Serum adiponectin level at baseline (month 0 [M0]) and at EOT in patients treated with rosiglitazone (n ⫽ 12) and those receiving placebo (n ⫽ 10); P ⫽ .04; error bars represent SD. (B) Correlation between changes in serum adiponectin level (ratio EOT/M0) and changes in histologic score of steatosis during treatment (difference between steatosis on the EOT biopsy and baseline biopsy, EOT-M0); r ⫽ ⫺0.54, P ⬍ .01, Spearman ranked test).

1. Marchesini G, Bugianesi E, Forlani G, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003;37:917–923. 2. Ratziu V, Poynard T. NASH: a hidden and silent fibroser finally revealed? J Hepatol 2005;42:12–14. 3. Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: from steatosis to cirrhosis. Hepatology 2006;43:S99 –S112. 4. Ekstedt M, Franzen LE, Mathiesen UL, et al. Long-term follow-up of patients with NAFLD and elevated liver enzymes. Hepatology 2006;44:865– 873. 5. Adams LA, Lymp JF, St Sauver J, et al. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005;129:113–121. 6. Ratziu V, Bonyhay L, Di Martino V, et al. Survival, liver failure, and hepatocellular carcinoma in obesity-related cryptogenic cirrhosis. Hepatology 2002;35:1485–1493. 7. Sanyal AJ, Banas C, Sargeant C, et al. Similarities and differences in outcomes of cirrhosis due to nonalcoholic steatohepatitis and hepatitis C. Hepatology 2006;43:682– 689.

8. Bugianesi E, Leone N, Vanni E, et al. Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002;123:134 – 140. 9. Sanyal AJ, Campbell-Sargent C, Mirshahi F, et al. Nonalcoholic steatohepatitis: association of insulin resistance and mitochondrial abnormalities. Gastroenterology 2001;120:1183–1192. 10. Suzuki A, Lindor K, St Saver J, et al. Effect of changes on body weight and lifestyle in nonalcoholic fatty liver disease. J Hepatol 2005;43:1060 –1066. 11. Foster GD, Wyatt HR, Hill JO, et al. A randomized trial of a low-carbohydrate diet for obesity. N Engl J Med 2003;348:2082– 2090. 12. Skarfors ET, Wegener TA, Lithell H, et al. Physical training as treatment for type 2 (non-insulin-dependent) diabetes in elderly men. A feasibility study over 2 years. Diabetologia 1987;30: 930 –933. 13. Neuschwander-Tetri BA, Brunt EM, Wehmeier KR, et al. Improved nonalcoholic steatohepatitis after 48 weeks of treatment with the PPAR-gamma ligand rosiglitazone. Hepatology 2003;38: 1008 –1017. 14. Promrat K, Lutchman G, Uwaifo GI, et al. A pilot study of pioglitazone treatment for nonalcoholic steatohepatitis. Hepatology 2004;39:188 –196. 15. Belfort R, Harrison SA, Brown K, et al. A placebo-controlled trial of pioglitazone in subjects with nonalcoholic steatohepatitis. N Engl J Med 2006;355:2297–2307. 16. Nair S, Diehl AM, Wiseman M, et al. Metformin in the treatment of non-alcoholic steatohepatitis: a pilot open label trial. Aliment Pharmacol Ther 2004;20:23–28. 17. Ratziu V, Giral P, Charlotte F, et al. Liver fibrosis in overweight patients. Gastroenterology 2000;118:1117–1123. 18. Kleiner DE, Brunt EM, Van Natta M, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005;41:1313–1321. 19. Brunt EM, Janney CG, DiBisceglie AM, et al. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999;94:2467–2474. 20. Prati D, Taioli E, Zanella A, et al. Updated definitions of healthy ranges for serum alanine aminotransferase levels. Ann Intern Med 2002;137:1–10. 21. Lutchman G, Modi A, Kleiner DE, et al. The effects of discontinuing pioglitazone in patients with nonalcoholic steatohepatitis. Hepatology 2007;46:424 – 429. 22. Friedman SL, Bansal MB. Reversal of hepatic fibrosis—fact or fantasy? Hepatology 2006;43:S82–S88. 23. Day CP. Pathogenesis of steatohepatitis. Best Pract Res Clin Gastroenterol 2002;16:663– 678. 24. Miyazaki Y, Mahankali A, Wajcberg E, et al. Effect of pioglitazone on circulating adipocytokine levels and insulin sensitivity in type 2 diabetic patients. J Clin Endocrinol Metab 2004;89:4312– 4319. 25. Iwaki M, Matsuda M, Maeda N, et al. Induction of adiponectin, a fat-derived antidiabetic and antiatherogenic factor, by nuclear receptors. Diabetes 2003;52:1655–1663. 26. Maeda N, Takahashi M, Funahashi T, et al. PPARgamma ligands increase expression and plasma concentrations of adiponectin, an adipose-derived protein. Diabetes 2001;50:2094 –2099. 27. Semple RK, Chatterjee VK, O’Rahilly S. PPAR gamma and human metabolic disease. J Clin Invest 2006;116:581–589. 28. Tiikkainen M, Hakkinen AM, Korsheninnikova E, et al. Effects of rosiglitazone and metformin on liver fat content, hepatic insulin resistance, insulin clearance, and gene expression in adipose tissue in patients with type 2 diabetes. Diabetes 2004;53: 2169 –2176. 29. Kubota N, Terauchi Y, Kubota T, et al. Pioglitazone ameliorates insulin resistance and diabetes by both adiponectin-dependent

ROSIGLITAZONE FOR NONALCOHOLIC STEATOHEPATITIS

30.

31.

32.

33.

34.

35.

36.

37.

38.

39. 40.

41.

42.

43.

44.

45.

46.

47.

48.

109

and -independent pathways. J Biol Chem 2006;281:8748 – 8755. Dixon JB, Bhathal PS, Hughes NR, et al. Nonalcoholic fatty liver disease: improvement in liver histological analysis with weight loss. Hepatology 2004;39:1647–1654. Yang SQ, Lin HZ, Lane MD, et al. Obesity increases sensitivity to endotoxin liver injury: implications for the pathogenesis of steatohepatitis. Proc Natl Acad Sci U S A 1997;94:2557–2562. Lin HZ, Yang SQ, Chuckaree C, et al. Metformin reverses fatty liver disease in obese, leptin-deficient mice. Nat Med 2000;6: 998 –1003. Ueki K, Kondo T, Tseng YH, et al. Central role of suppressors of cytokine signaling proteins in hepatic steatosis, insulin resistance, and the metabolic syndrome in the mouse. Proc Natl Acad Sci U S A 2004;101:10422–10427. Arkan MC, Hevener AL, Greten FR, et al. IKK-beta links inflammation to obesity-induced insulin resistance. Nat Med 2005;11: 191–198. Fromenty B, Robin MA, Igoudjil A, et al. The ins and outs of mitochondrial dysfunction in NASH. Diabetes Metab 2004;30: 121–138. Li Z, Lin H, Yang S, et al. Murine leptin deficiency alters Kupffer cell production of cytokines that regulate the innate immune system. Gastroenterology 2002;123:1304 –1310. Feldstein AE, Canbay A, Guicciardi ME, et al. Diet associated hepatic steatosis sensitizes to Fas mediated liver injury in mice. J Hepatol 2003;39:978 –983. Cortez-Pinto H, Chatham J, Chacko VP, et al. Alterations in liver ATP homeostasis in human nonalcoholic steatohepatitis: a pilot study. JAMA 1999;282:1659 –1664. Yki-Jarvinen H. Fat in the liver and insulin resistance. Ann Med 2005;37:347–356. Samuel VT, Liu ZX, Qu X, et al. Mechanism of hepatic insulin resistance in non-alcoholic fatty liver disease. J Biol Chem 2004; 279:32345–32353. Seppala-Lindroos A, Vehkavaara S, Hakkinen AM, et al. Fat accumulation in the liver is associated with defects in insulin suppression of glucose production and serum free fatty acids independent of obesity in normal men. J Clin Endocrinol Metab 2002;87:3023–3028. Ryysy L, Hakkinen AM, Goto T, et al. Hepatic fat content and insulin action on free fatty acids and glucose metabolism rather than insulin absorption are associated with insulin requirements during insulin therapy in type 2 diabetic patients. Diabetes 2000; 49:749 –758. Galli A, Crabb DW, Ceni E, et al. Antidiabetic thiazolidinediones inhibit collagen synthesis and hepatic stellate cell activation in vivo and in vitro. Gastroenterology 2002;122:1924 – 1940. McCullough AJ. Thiazolidinediones for nonalcoholic steatohepatitis—promising but not ready for prime time. N Engl J Med 2006;355:2361–2363. Marchesini G, Avagnina S, Barantani EG, et al. Aminotransferase and gamma-glutamyltranspeptidase levels in obesity are associated with insulin resistance and the metabolic syndrome. J Endocrinol Invest 2005;28:333–339. Lee DH, Silventoinen K, Jacobs DR Jr, et al. Gamma-glutamyltransferase, obesity, and the risk of type 2 diabetes: observational cohort study among 20,158 middle-aged men and women. J Clin Endocrinol Metab 2004;89:5410 –5414. Nawrocki AR, Rajala MW, Tomas E, et al. Mice lacking adiponectin show decreased hepatic insulin sensitivity and reduced responsiveness to peroxisome proliferator-activated receptor gamma agonists. J Biol Chem 2006;281:2654 –2660. US Food and Drug Administration. FDA briefing document: joint meeting of the Endocrinologic and Metabolic Drug Advisory Committee and the Drug Safety and Risk Management Committee.

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

July 2008

110

RATZIU ET AL

Available at: http://www.fda.gov/ohrms/dockets/ac/07/minutes/ 2007-4308m1-final.pdf. Accessed September 6, 2007. 49. Murphy CE, Rodgers PT. Effects of thiazolidinediones on bone loss and fracture. Ann Pharmacother 2007;41:2014 –2018. 50. Lindor KD, Kowdley KV, Heathcote EJ, et al. Ursodeoxycholic acid for treatment of nonalcoholic steatohepatitis: results of a randomized trial. Hepatology 2004;39:770 –778.

CLINICAL ADVANCES IN LIVER, PANCREAS, AND BILIARY TRACT

Received November 13, 2007. Accepted March 20, 2008. The authors wish to thank Janet Ratziu for manuscript editing and Gilles Brami and Denis Vallée from GSK Pharmaceuticals for their support during the trial. This was an investigator-initiated trial, and GlaxoSmithKline Pharmaceuticals had no direct or indirect involvement in the design of the trial, data collection, or preparation or submission of the manuscript. GlaxoSmithKline provided rosiglitazone and placebo for this trial and partly funded the trial. None of the authors have a personal conflict of

GASTROENTEROLOGY Vol. 135, No. 1

interest with the manufacturer of any of the marketed thiazolidinediones. V.R. is a consultant to Astellas, Gilead, Pfizer, Sanofi-Aventis, and Trophos. T.P. is a consultant for and owns 15% of BioPredictive, a company that markets FibroTest and SteatoTest. The members of the Liver Injury in Diabetes and Obesity (LIDO) Study Group are as follows: André Grimaldi, Philippe Giral, Eric Bruckert, Arnaud Basdevant, Karine Clement, Agnès Hartemann-Heurtier, Sophie Gombert, Francine Lamaison, Dominique Simon, Joseph Moussalli, Pascal Lebray, Christiane Coussieu, Djamila Messous, Françoise Imbert-Bismut, Yves Benhamou, Cecilia D’Arrondel, Carole Bernhardt, Isabelle Ravalet, Hôpital Pitié-Salpêtrière; Philippe Podevin, Hôpital Cochin; Christian Boitard, Etienne Larger, Hôpital Hotel-Dieu; Lawrence Serfaty, Chantal Housset, Jacqueline Capeau, Hôpital Saint Antoine, all in Paris, France.. Address requests for reprints to: Prof Vlad Ratziu, Hôpital PitiéSalpetriére, 47-83 Bd de l’Hôpital, Paris, France. e-mail: [email protected]; fax: (33) 1 42 16 10 49.