Histological benefits of virological response to peginterferon alfa-2a monotherapy in patients with hepatitis C and advanced fibrosis or compensated cirrhosis

Authors


Dr G. T. Everson, Section of Hepatology, University of Colorado School of Medicine, University of Colorado Health Sciences Center, UCH AOP, Hepatology Section, 1635 N Ursula, B-154, PO Box 6510, Aurora, CO 80045, USA.
E-mail: greg.everson@uchsc.edu

Summary

Background  Patients with chronic hepatitis C virus and advanced fibrosis or cirrhosis are at risk for disease progression and hepatic decompensation.

Aim  To determine the effects on hepatic histology of treatment with peginterferon alfa-2a (90 or 180 μg/week) or interferon alfa-2a (3 million units three times weekly) for 48 weeks in patients with paired biopsies.

Methods  Liver biopsies were obtained at baseline and 6 months after end of treatment. Histological and virological responses were compared.

Results  Patients attaining sustained virological response (n = 40) demonstrated the greatest improvements in fibrosis (−1.0, P < 0.0001) and inflammation (−0.65, P < 0.0001). Patients who cleared hepatitis C virus during treatment, but later relapsed (n = 59), experienced less improvement in fibrosis (−0.04, P < 0.0001) and inflammation (−0.14, P = 0.0768). Nonresponders (n = 85) showed no significant improvement in inflammation or fibrosis. Multiple regression analysis showed that the only factors contributing to improvement in fibrosis were sustained virological response (vs. nonresponder, P = 0.0005; vs. relapse, P = 0.7525) and body mass index ≤30 kg/m2 (P = 0.0995).

Conclusions  These findings indicate that virological response to peginterferon alfa-2a improves inflammation and fibrosis in hepatitis C virus patients with advanced fibrosis or cirrhosis. Improving virological response and maintaining ideal body weight are critical for achieving optimal histological outcomes in hepatitis C virus patients.

Introduction

Chronic infection with the hepatitis C virus (HCV) is associated with variable rates of progression of hepatic fibrosis.1 Because of their improved pharmacokinetics, pegylated forms of interferon alfa have replaced unmodified interferon alfa, with or without the nucleoside analogue ribavirin, in the treatment of patients with chronic HCV infection.2 The primary goal of treatment has been to achieve a sustained virological response (SVR), defined as undetectable serum HCV-RNA 6 months after the end of treatment. Several studies, primarily in noncirrhotic patients, have shown that these treatment regimens may be effective in slowing, stopping or even reversing the progression of hepatic fibrosis.3–10 The antifibrotic effect of interferon may be related to viral clearance or suppression of HCV-RNA, or to a direct effect of interferon. There have been conflicting reports concerning whether improvements in fibrosis scores can occur in patients who do not have sustained viral clearance.

Various demographic and clinicopathological variables have been reported to affect the rate of hepatic fibrosis progression and the probability of virological and histological responses to treatment in HCV-infected individuals. These variables include age, race/ethnicity, alcohol consumption, weight, body mass index (BMI), the presence of diabetes or insulin resistance, and hepatic steatosis or iron deposition.1, 11–20 The contribution of each of these variables to the histological changes occurring after interferon therapy remains uncertain.

Patients with advanced fibrosis and compensated cirrhosis are candidates for treatment with pegylated interferon alfa-2a plus ribavirin, although they have lower overall SVR rates than patients with less advanced liver disease.9, 21 Our analysis evaluated the effects of treatment on detailed liver histology in chronically infected patients with compensated liver cirrhosis or bridging fibrosis, using the histological results from the only Phase III multinational clinical trial comparing monotherapy with standard interferon [3 million units (MU) three times weekly] vs. peginterferon alfa-2a (90 or 180 μg/week).22 The primary aim of this analysis was to assess the association between histological response and virological response. In addition, we examined the influence of selected covariates on these relationships.

Materials and methods

Patients

Patients in this analysis were chronically infected with HCV and had histological evidence of liver cirrhosis (76%) or bridging fibrosis (24%) and took part in an open-label, randomized, parallel-dose trial designed to compare the effects of 48 weeks of treatment with 90- or 180-μg peginterferon alfa-2a once weekly and 3 MU of standard interferon alfa-2a three times weekly (Protocol NV 15495).22 Although histology results were reported in the original article, this analysis represents an in-depth appraisal of the relationships between virological and histological responses. All patients were followed for 24 weeks after the end of treatment. Individuals in this study had not been previously treated with interferon therapies and had abnormal serum alanine aminotransferase (ALT) levels on at least two occasions during the preceding 6 months. Exclusion criteria, detailed previously, included patients with active alcohol and/or drug abuse within 1 year of entry.22 Patients who drank alcohol were instructed to consume <20 g daily during the study and were questioned during the study regarding alcohol consumption.

Histological evaluation of the liver

Pre-treatment liver biopsy samples were examined without blinding before randomization. These samples were subsequently coded and evaluated in parallel with those obtained at week 72 by a central pathologist (SG) who was blinded to treatment and time of biopsy. The median time between biopsies was 593 days. Fibrosis and activity scores were graded according to METAVIR and Histological Activity Index criteria.23, 24 Improvements in fibrosis and activity were defined as ≥1 stage or grade improvements respectively.

Virological and biochemical assays

HCV-RNA levels were measured by COBAS AMPLICOR HCV MONITOR Test, v. 2.0 (Roche Diagnostics, Branchburg, NJ, USA) (limit of quantitation 100 copies/mL) at baseline and every 4 weeks after initiation of treatment, until week 72. Early viral response (EVR) was defined as undetectable plasma HCV-RNA (<100 copies/mL) or ≥2-log10 decrease in plasma HCV-RNA at week 12. SVR was defined as undetectable HCV-RNA at week 72.22 Patients with an end-of-treatment response (ETR), defined as undetectable HCV-RNA at week 48, but who experienced re-emergence of HCV-RNA by week 72, were defined as relapsers.22 Serum ALT levels were measured before treatment and at the end of follow-up. The ALT quotient was defined as the baseline ALT level divided by the upper limit of normal (ULN). A sustained biochemical response (SBR) was defined as ALT levels below ULN at 72 weeks. HCV genotyping was performed by sequence analysis of a portion of the 5′-untranslated region of the viral genome.25

Statistical methods

Only METAVIR results were used for statistical assessments of histology. Baseline characteristics, including sex, age, race, weight, BMI (≤30 kg/m2 vs. >30 kg/m2), ALT level (≤4× ULN vs. >4× ULN), serum HCV-RNA level (≥2 × 106 copies/mL vs. <2 × 106 copies/mL), HCV genotype (genotype 1 vs. non-1) and baseline fibrosis scores for all patients with paired liver biopsies were summarized as mean or as frequency and percentage where indicated. The relationship between each baseline characteristic and histological (fibrosis and activity) response to treatment was tested using single- and multiple-variable logistic regression analyses. In multiple-variable analyses, odds ratio (ORs) were determined for particular characteristics (risk factors) and adjusted for confounding variables, and 95% confidence intervals were determined. A significance level of α = 0.10 was used for logistic regression for analysis of predictors. All other tests were performed using α = 0.05. Fibrosis change from baseline for each group based on viral response was compared using the Wilcoxon signed rank test. The proportion of patients with histological improvement in each of the three treatment arms was assessed by the Mantel-Haenszel χ2 test. All patients with paired biopsies were evaluated.

Results

Baseline characteristics

Of the 271 patients enrolled in the study, 184 (68%) had paired liver biopsy specimens.22 Relative to the intent-to-treat (ITT) population in each group, those with paired biopsies constituted 55 of 88 (63%) patients in the interferon alfa-2a 3 MU three times weekly group, 61 of 96 (64%) in the peginterferon alfa-2a 90-μg/week group and 68 of 87 (78%) in the peginterferon alfa-2a 180-μg/week group. The demographic and clinical characteristics of these 184 patients were similar to those of the 271 patients in the ITT population (Table 1). At baseline, 140 (76%) had cirrhosis and 44 (24%) had extensive bridging fibrosis. With respect to treatment arm, of the 55 patients randomized to interferon alfa-2a 3 MU three times weekly, 39 (71%) exhibited cirrhosis and 16 (29%) exhibited bridging fibrosis; of the 61 patients randomized to peginterferon alfa-2a 90 μg/week, 45 (74%) exhibited cirrhosis and 16 (26%) exhibited bridging fibrosis; and of the 68 patients randomized to peginterferon alfa-2a 180 μg/week, 56 (82%) exhibited cirrhosis and 12 (18%) exhibited bridging fibrosis.

Table 1.   Demographic and clinical characteristics at baseline in patients with paired liver biopsies
 Interferon alfa-2a (3 MU tiw; n = 55)Peginterferon alfa-2a (90 μg/week; n = 61)Peginterferon alfa-2a (180 μg/week; n = 68)
  1. * Defined as baseline ALT divided by the ULN.

  2. MU, million units; BMI, body mass index; ALT, alanine aminotransferase; ULN, upper limit of normal; HCV, hepatitis C virus.

Male gender, n (%)35 (64)45 (74)49 (72)
Mean age (years)47.847.546.6
Race, n (%)
 Caucasian48 (87)58 (95)60 (88)
 Non-Caucasian7 (13)3 (5)8 (12)
Weight (kg)82.285.081.4
BMI >30 kg/m2, n (%)16 (29)18 (30)18 (27)
ALT
 Mean ALT quotient*3.33.34.2
 >4× ULN30 (34)26 (27)31 (36)
Geometric mean plasma HCV-RNA (×106 copies/mL)2.42.41.7
HCV genotype 1, n (%)31 (56)33 (54)38 (56)
METAVIR grade, mean2.02.12.0
METAVIR stage, mean3.73.73.8
 F3, n (%)16 (29)16 (26)12 (18)
 F4, n (%)39 (71)45 (74)56 (82)

Virological response

The overall rates of SVR, relapse and nonresponse for these 184 patients by treatment arm are shown in Table 2. The SVR rate in the peginterferon alfa-2a 180-μg/week group was 37%, compared with 16% in the peginterferon alfa-2a 90-μg/week group and 9% in the interferon alfa-2a 3 MU three times weekly group.

Table 2.   Virological response rates
  Interferon alfa-2a (3 MU tiw; n = 55)Peginterferon alfa-2a (90 μg/week; n = 61) Peginterferon alfa-2a (180 μg/week; n = 68)
  1. SVR, sustained virological response.

SVR, n (%)5 (9)10 (16)25 (37)
Relapse, n (%)14 (26)29 (48)16 (24)
Nonresponse, n (%)36 (66)22 (36)27 (40)

Biochemical response

The SBR rate was 34% in the peginterferon alfa-2a 180-μg/week group, 20% in the peginterferon alfa-2a 90-μg/week group and 15% in the interferon alfa-2a 3 MU three times weekly group.

Histological response

Histological response (≥1 point reduction in either fibrosis or necroinflammatory score) following treatment correlated with SVR.22 Histological improvement after SVR was observed in 80%, 100% and 88% of patients randomized to interferon alfa-2a 3 MU three times weekly, peginterferon alfa-2a 90 μg/week and peginterferon alfa-2a 180 μg/week respectively. In contrast, only 26%, 33% and 35% of patients without SVR demonstrated histological improvement with these respective treatments.

A summary of the percentages of patients with liver fibrosis improvement, grouped according to treatment arm, is shown in Table 3. Improvement in fibrosis stage, defined as post-treatment score less than baseline score, was observed in 54 patients overall (29%). Fibrosis improved in 24 of 68 (35%) patients receiving peginterferon alfa-2a 180 μg/week, 15 of 61 (25%) patients receiving peginterferon alfa-2a 90 μg/week and 15 of 55 (27%) patients receiving standard interferon alfa-2a 3 MU three times weekly.

Table 3.   Summary of proportional of patients showing fibrosis stage and inflammatory activity grade improvements
Histological responseInterferon alfa-2a (3 MU tiw; n = 55)Peginterferon alfa-2a (90 μg/week; n = 61)Peginterferon alfa-2a (180 μg/week; n = 68)
  1. MU, million units.

Liver fibrosis stage improvement, n (%)15 (27)15 (25)24 (35)
P-value (vs. interferon) 0.74290.3438
Activity grade improvement, n (%)6 (11)19 (31)19 (28)
P-value (vs. interferon) 0.00840.0201

Table 3 also shows the percentages of patients with improvement in grade of inflammatory activity, grouped according to treatment arm. Inflammatory activity improved in 44 patients overall (24%). Reduced inflammation was observed in 11% of patients receiving unmodified interferon alfa-2a, and 31% and 28% of those receiving peginterferon alfa-2a 90 μg/week and 180 μg/week respectively (P = 0.0084 and P = 0.0201 compared with interferon alfa-2a 3 MU three times weekly).

Variables associated with fibrosis improvement

Histological improvements by virological response are summarized in Figure 1. Histological benefit occurred in patients with virological response, either SVR or on-treatment clearance of HCV-RNA followed by relapse. Fibrosis improvement, defined as a ≥1 stage improvement in METAVIR score, was observed in 50% of patients who achieved SVR and in 34% of virological relapsers but in only 16% of virological nonresponders (Figure 1). Fibrosis scores decreased significantly from baseline in patients with SVR (mean change, −1.0; P < 0.0001) or virological relapse (mean change, −0.44; P < 0.0001); but there was only a trend towards a decrease in nonresponders (mean change, −0.15; P = 0.0669) (Figure 2a).

Figure 1.

 Histological improvement by virological response. SVR, sustained virological response.

Figure 2.

 Change from baseline in fibrosis stage (a) and necroinflammatory activity grade (b) grouped by virological response (Wilcoxon signed rank test). SVR, sustained virological response.

The percentage of patients with improvement in necroinflammatory grade was also related to virological response, with a ≥1 grade improvement observed in 63% of patients with SVR and 20% of virological relapsers, but in only 8% of nonresponders (Figure 1). The degree of inflammation improved significantly from baseline in patients who achieved SVR (mean change, −0.65; P < 0.0001), showed a trend towards improvement in virological relapsers (mean change, −0.14; P = 0.0768), but worsened in nonresponders (mean change, +0.12; P = 0.0383) (Figure 2b).

Single variable logistic regression analyses showed that the characteristics associated with improvement in fibrosis were SVR (vs. nonresponse, OR = 5.1; P =0.0022; vs. relapse, OR = 2.0; P = 0.6802); EVR (vs. non-EVR, OR = 3.1; P = 0.0014); SBR (vs. non-SBR, OR = 2.4; P = 0.0162) and BMI (≤30 vs. >30 kg/m2, OR = 2.1; P = 0.0619) (Table 4). Characteristics associated with improvement in inflammation were SVR (vs. nonresponse, OR = 18.6, P < 0.0001; vs. relapse, OR = 6.5, P = 0.31); EVR (vs. non-EVR, OR = 4.2; P = 0.0004); genotype non-1 (vs. genotype 1, OR = 3.9; P = 0.0003) and baseline ALT ≤4× ULN (vs. >4× ULN, OR = 2.0; P = 0.049) (Table 5).

Table 4.   Summary of single variable logistic regression analyses of fibrosis stage improvement
 Improvement in fibrosis stage
OR (95% CI)P-value*
  1. * A significance level of 0.10 was used for logistic regression for analysis of predictors.

  2. OR, odds ratio; CI, confidence interval; BMI, body mass index; EVR, early virological response; SBR, sustained biochemical response; SVR, sustained virological response.

BMI (≤30 vs. >30 kg/m2)2.1 (1.0–4.6)0.0619
EVR (yes vs. no)3.1 (1.6–6.3)0.0014
SBR vs. non-SBR2.4 (1.2–4.9)0.0162
SVR (d.f. = 2) 0.0007
 SVR vs. relapse2.0 (0.9–4.4)0.6802
 SVR vs. nonresponse5.1 (2.2–11.8)0.0022
Table 5.   Summary of single variable logistic regression analyses of activity grade improvement
 Improvement in activity grade
OR (95% CI)*P-value
  1. OR, odds ratio; CI, confidence interval; ALT, alanine aminotransferase; ULN, upper limit of normal; EVR, early virological response; SVR, sustained virological response.

ALT (≤4× ULN vs. >4× ULN)2.0 (1.0–4.0)0.049
Activity grade (3 vs. 2)9.3 (2.8–31.6)0.0003
EVR (yes vs. no)4.2 (1.9–9.5)0.0004
Genotype (non-1 vs. 1)3.9 (1.9–8.1) 0.0003
SVR (d.f. = 2) <0.0001
 SVR vs. relapse6.5 (2.7–16.1)0.31
 SVR vs. nonresponse18.6 (6.8–50.7)<0.0001

Multiple regression analysis was used to define independent predictors of histological response. Of baseline characteristics, only genotype non-1 (P =0.0003) and baseline activity grade ≥3 (P = 0.0002) were independent predictors of activity grade improvement, while BMI ≤30 kg/m2 tended to predict fibrosis improvement (P = 0.0619). In multiple regression analyses that included both significant baseline characteristics and SVR, SVR (P = 0.001) and BMI ≤30 kg/m2 (P = 0.0995) significantly predicted improvement in fibrosis. In similar analyses, baseline activity grade ≥3 (vs. grade 2; P = 0.0003) and SVR (vs. non-SVR; P < 0.0001) were significant independent predictors of improvement in necroinflammatory activity.

Discussion

One of the potential benefits of interferon alfa-based therapies in patients with advanced chronic HCV is reduction in hepatic fibrosis. Hepatic fibrosis may improve or stabilize after treatment, primarily in patients who attain SVR.3, 5, 7, 8, 26 A majority of subjects in these studies, however, did not have advanced fibrosis or cirrhosis. In our analysis of patients with advanced fibrosis, 76% of whom had cirrhosis, virological response characterized by either SVR or relapse, were associated with improvements in fibrosis and inflammation.

Our study has confirmed that virological response to interferon therapy is the key factor driving histological improvement. The only other characteristic at baseline that tended to predict fibrosis improvement was BMI ≤30 kg/m2 (P = 0.0619). By single-variable analyses, SVR, EVR and SBR were all significantly associated with improvement in fibrosis stage. By multiple regression analyses, the predictors of improvement in hepatic fibrosis were SVR (vs. nonresponder, P = 0.0005; vs. relapse, P = 0.7525) and BMI ≤30 kg/m2 (P = 0.0995). Independent predictors of improvement in inflammation were higher baseline necroinflammation (P = 0.0003) and SVR (vs. nonresponder, P < 0.0001; vs. relapse, P = 0.312).

The greater benefit on activity in patients with genotype non-1 may be because of the greater sensitivity to interferon. Indeed, patients infected with HCV genotypes 2 and 3 have been shown to have a higher rate of rapid virological response (RVR), defined as undetectable HCV-RNA at week 4, than patients infected with HCV genotype 1.27, 28 Indeed, we observed a significant association between EVR and improvements in inflammation, suggesting that inflammation may decrease as soon as the virus disappears.

Not all patients attaining SVR showed histological improvement, whereas some patients who did not attain SVR showed improvement in fibrosis stage. Among the variables that may account for this discordance between SVR and histological improvement are possible sampling error and the time interval between biopsies. A longer follow-up period may be needed to assess the reversibility of fibrosis or inflammatory activity fully.29 Sampling error could be especially crucial in patients with advanced fibrosis, many of whom were at stage F4 at initial biopsy. Patients with initial and follow-up biopsies at stage F4 were scored as stable as they cannot be given higher stages with the METAVIR system. The inability to measure progression beyond stage F4, as well as sampling error, could account for the trend towards improvement in fibrosis score in nonresponders, indicating that this finding may be an artefact.

Other studies also suggest that SVR improves outcomes in patients with HCV-related cirrhosis. In a retrospective analysis of 920 patients with cirrhosis treated with interferon-alfa monotherapy, the incidence rates of hepatic decompensation, hepatocellular carcinoma (HCC) and liver-related death were 0, 0.66 and 0.19 per 100 person-years, respectively, in the 124 patients who achieved SVR. In contrast, incidence rates were 1.9, 2.1 and 1.4 per 100 person-years, respectively, in patients who did not achieve SVR (SVR vs. non-SVR, P < 0.001).30 In another retrospective analysis of 479 patients with advanced fibrosis or cirrhosis (Ishak score, 4–6) followed up for a median 2.1 years after treatment, the hazard of events (death, liver failure and HCC) was significantly lower in the 143 patients who achieved SVR than in the 336 who did not (adjusted hazard ratio, 0.21; 95% CI: 0.07–0.58; P = 0.003).31 This difference was largely attributable to a reduction in liver failure, which occurred in none of the patients with SVR and 42 patients without SVR, indicating that SVR prevents liver failure in patients with chronic HCV and advanced fibrosis.

Other studies have shown that, relative to no treatment, virological response to interferon may reduce the cumulative incidence of HCC.32, 33 Additionally, in patients with fibrosis stage 3 or 4 and elevated baseline hepatic venous pressure gradient (HVPG >5 mmHg), SVR reduced HVPG.34 Reductions in HVPG were the greatest in those achieving ETR and ≥2-point decrease in inflammation.

A recent large study has confirmed that SVR prevents progression to cirrhosis.35 In an analysis of 1386 noncirrhotic chronic HCV patients, 892 treated with interferon-alfa and 494 untreated, the 15-year cumulative incidence of cirrhosis was significantly lower in treated patients (10% vs. 40%, respectively; P =0.0008). Furthermore, the 14.5-year cumulative incidence of cirrhosis was significantly lower in treated patients who achieved SVR (5%) compared with treated patients without SVR (22%; P = 0.0007) or untreated patients (37%; P < 0.0001).

Efficacy of maintenance therapy in preventing disease progression may also be linked to virological response. The effect of continued treatment on histological response was assessed in a trial in which 53 patients who showed a histological response to 6 months of interferon treatment but remained positive for HCV-RNA were randomized to no further treatment (n = 26) or to maintenance treatment with interferon for additional 24 months (n = 27).26 While both groups showed significant reductions in serum ALT level, viral titre and hepatic inflammation after 6 months of re-treatment with interferon, these improvements were maintained in patients who continued treatment, whereas these measurements returned to baseline in patients who stopped treatment. At the end of 30 months of treatment, mean fibrosis score declined from 2.5 to 1.7, and 80% of patients showed histological improvement (P < 0.03). In contrast, patients who stopped therapy showed an increase in mean fibrosis score from 2.2 to 2.4, and 30% of patients showed worsening of histology (P < 0.01).

In contrast, a recent trial in 102 patients with biopsy-proven cirrhosis randomized to treatment with interferon-alfa for 24 months or no treatment showed that complication-free survival rates were similar in the two groups.36 Results from the Hepatitis C Anti-viral Long-term Treatment against Cirrhosis (HALT-C) trial have indicated that maintenance therapy with 90-μg/week peginterferon alfa-2a did not reduce the rate of disease progression.37 The patients in the HALT-C trial were characterized by advanced hepatic fibrosis or compensated cirrhosis and nonresponse to prior therapy with peginterferon alfa-2a and ribavirin. In this trial, 1050 patients were randomized to treatment with 90-μg/week peginterferon alfa-2a for 3.5 years (n = 517) or to no treatment (n = 533), with patients stratified by fibrosis score. Although treatment significantly reduced mean serum ALT and HCV-RNA levels and necroinflammatory changes (P < 0.0001 each), it had no effect on the rates of any of the primary outcomes (i.e. liver-related death, hepatic decompensation and HCC). Increases in fibrosis were observed in 28% of patients in the treated group vs. 32% in the untreated group. Subanalyses of the relationships of degree of virological suppression to histological response and clinical outcomes in the HALT-C trial are yet to be presented or published.

An earlier retrospective study, in which fibrosis progression rates were pooled from four randomized-controlled trials of HCV-infected patients receiving one of 10 possible drug regimens, including therapies with standard and pegylated interferons with or without ribavirin, found that six baseline characteristics were associated with improved fibrosis or activity response.7 Among these were age <40 years, BMI <27 kg/m2 and low viral concentration at baseline,7 findings consistent with those of the meta-analysis by Camma et al.8 In our study, of the baseline variables considered, only BMI ≤30 kg/m2 was a predictor of improvement in hepatic fibrosis even in the presence of SVR.

Although histological improvement was attained by some patients in all three of our treatment arms, patients in the high-dose peginterferon treatment arm tended towards greatest benefit. This finding is in agreement with results showing that treatment with peginterferon alfa-2a 180 μg/week was more effective than 135 μg/week in inducing histological improvements, although the SVR rates in the two groups were identical.6

Several potential mechanisms have been hypothesized for the antifibrotic effect of interferon, including that interferon-alfa may directly reduce fibrogenesis.3, 10, 38, 39 However, the results of our study and the published literature suggest that most, if not all, of the antifibrotic effect is mediated mainly by suppressing HCV replication.40 Indeed, we found that patients who cleared virus during treatment, whether or not they achieved SVR, showed significant improvements in fibrosis, whereas patients who never cleared virus did not show improvement.

Our study had several limitations. The primary end point, change in histology required assessment of liver biopsy results. This may have introduced uncertainties related to sampling error, errors that may be amplified in patients with advanced fibrosis. In addition to sampling error, current histological staging does not allow determination of fibrosis progression in patients with fibrosis (stage F4) at baseline. Finally, a longer follow-up period may be needed to assess fully the reversibility of fibrosis or inflammatory activity.29

In conclusion, we report two key findings. First, the results support the concept that virological response to peginterferon alfa-2a is the key determinant of histological improvement. Secondly, obesity, defined as BMI >30 kg/m2, may impair histological improvement after interferon or peginterferon therapy. In patients with advanced fibrosis or compensated cirrhosis, efforts to improve rates of SVR and maintain ideal body weight are critical for achieving optimal histological outcomes.

Acknowledgements

We thank Dr Janet S. Lee for her contributions to the analysis of this study and her critical review of the manuscript, and Dr Richard J. Alexander for medical writing and editing assistance.

Declaration of personal interests: G. T. Everson receives research grant support, is a consultant, and on the speakers’ bureau for Roche Laboratories. L. Balart is an advisory board member for Roche Laboratories and Gilead, a speaker for Roche Laboratories and Schering-Plough, and has received research grants from Roche Laboratories, Schering-Plough, Coley, Vertex, Bayer, Sciclone, Gilead, and Ocera. S. S. Lee has received research support from Gilead, Human Genome Science, Microgenix, Roche Laboratories, Schering-Plough, Idenix, Bristol Myers Squibb, Virochem, GlaxoSmithKline, and Novartis. He is also a consultant for Genentech, Idenix, Microgenix, Roche Laboratories, Bristol Myers Squibb, Novartis, and Virochem as well as on the speakers’ bureau for Roche Laboratories, Gilead and Bristol Myers Squibb. R. W. Reindollar receives research funding from Hoffman-LaRoche and Schering-Plough, and is also on the speakers’ bureau for Hoffman-LaRoche. M. L. Shiffman has served as a speaker, a consultant, and an advisory board member for Roche Laboratories, and has received research funding from Roche Laboratories and Schering-Plough. G. Y. Minuk has no personal interests. P. J. Pockros has served as a speaker, a consultant, and an advisory board member for Hoffman-LaRoche. S. Govindarajan has served as a consultant and the central Pathologist for Roche Laboratories and has also received research funding from Roche Laboratories. E. Lentz is an employee of Roche Laboratories. E. J. Heathcote is an advisory board member for Axcan Pharma, Gilead Sciences, Hoffman-LaRoche, Novartis, and Schering-Plough; receives grant/research funding from Axcan Pharma, Bristol-Myers Squibb, Gilead Sciences, Hoffman-LaRoche, Novartis, Pharmasset, Schering-Plough, and Vertex, is a speaker for Axcan Pharma, Gilead Sciences, Hoffman-LaRoche, and Schering-Plough, and is also on the board of Axcan Pharma. Declaration of funding interests: Supported by Roche Laboratories.

Ancillary