| |
Mitochondrial Toxicity and HCV Liver Disease Progression in Coinfection
|
| |
| |
Progression of Fibrosis in HIV and Hepatitis C Virus-Coinfected Patients Treated with Interferon plus Ribavirin-Based Therapy: Analysis of Risk Factors
Clinical Infectious Diseases March 1, 2008;46:768-774
Firouze Bani-Sadr,1 Nathanael Lapidus,1 Pierre Bedossa,2 Corinne Merle De Boever,5 Christian Perronne,6 Philippe Halfon,7 Stanislas Pol,3 Fabrice Carrat,1 Patrice Cacoub,4 and the French National Agency for Research on AIDS and Viral Hepatitis-HC02-Ribavic Study Team
1Groupe Hospitalier Universitaire Est, Universite Paris 6, INSERM U707, 2Groupe Hospitalier Universitaire Nord, Universite Paris 7, 3Groupe Hospitalier Universitaire Ouest, Universite Paris 5, INSERM U370, and 4Groupe Hospitalier Universitaire Est, UMR 7087, Universite Pierre et Marie Curie, Paris-6, Paris, 5Hopital Gui de Chauliac, Montpellier, 6Centre Hospitalier Universitaire Raymond Poincare, Universite de Versailles, Garches, and 7Laboratoire Alphabio-CDL Pharma et Hopital Ambroise Pare, Marseille, France
"....To our knowledge, this is the first study to analyze risk factors for worsening of fibrosis during anti-HCV therapy in HIV-HCV-coinfected patients. Fibrosis worsened within 2 years in 34 (17.1%) of the patients who started a 48-week course of IFN-_-2b plus ribavirin-based therapy, including pegylated IFN and conventional IFN. Thirteen (38.2%) of these 34 patients had at least a 3-grade increase in fibrosis on the Ishak scale.....
..... Worsening of fibrosis was related both to failure to have a sustained viral response (OR, 9.05 p=.003) and to receipt of didanosine-based antiretroviral therapy (OR, 3.34; p=.007)......
.....In conclusion, the ultimate benefit of anti-HCV therapy is an improvement in the histologic status of the liver. However, in HIV-HCV-coinfected patients, worsening of fibrosis can occur rapidly, despite HCV therapy and HCV infection cure. Mitochondrial toxicity-involving a synergistic, deleterious interaction between medications for HCV and HIV infections and, particularly, didanosine-seems to play a major role. Therefore, anti-HCV therapy for HIV-HCV-coinfected patients should ideally be administered before antiretroviral treatment is initiated. If anti-HCV and anti-HIV treatments must be administered concomitantly, then the NRTIs with the lowest mitochondrial toxicity should be preferred....
.... A large proportion of our patients had immunovirological control of HIV disease while receiving antiretroviral treatment. The mean CD4 cell count was >500 cells/mm3, and the HIV load was <400 copies/mL in two-thirds of patients....
.....potential additive effect on mitochondrial toxicity of ribavirin and antiretroviral drugs [23]. Therefore, worsening of fibrosis during anti-HCV therapy could occur because of a synergistic detrimental effect involving the lack of any reduction in mitochondrial depletion in nonresponders and an enhancement of mitochondrial damage caused by combined antiretroviral therapy and anti-HCV therapy...."
ABSTRACT
Background. We determined the prevalence and determinants of worsening fibrosis in patients coinfected with human immunodeficiency virus (HIV) and hepatitis C virus (HCV) who were receiving anti-HCV therapy.
Methods. Among 383 HIV-HCV-coinfected patients who received at least 1 dose of anti-HCV treatment (weekly subcutaneous injections of 1.5 _g/kg pegylated interferon-_-2b plus daily ribavirin or thrice-weekly subcutaneous injections of 3 MU of interferon-_-2b plus daily ribavirin for 48 weeks), paired pretreatment and posttreatment liver biopsy specimens were available and interpretable for 198 cases. Hepatic necroinflammation and fibrosis were graded with Ishak's classification. Histological worsening of fibrosis was defined as a score increase of >2 points in patients with fibrosis stage of <4 and as a score increase of 1 point in patients with stage-5 fibrosis.
Results. The mean deviation between the 2 biopsies was weeks. Fibrosis worsened in 34 patients (17.1%). In univariate analysis, ongoing antiretroviral therapy, failure to achieve a sustained viral response, nucleoside reverse-transcriptase inhibitor therapy, didanosine therapy, and stavudine therapy were significantly associated with worsening of fibrosis. Didanosine (odds ratio, 3.34; 95% confidence interval, 1.39-7.96; p=.007) and failure to have a sustained viral response (odds ratio, 9.05; 95% confidence interval, 2.06-39.66; p=.003) remained significantly associated with worsening of fibrosis.
Conclusion. The mitochondrial toxicity of antiretrovirals, such as didanosine, seems to play a major role in worsening of fibrosis during HCV therapy. Therefore, anti-HCV therapy should ideally be administered before antiretroviral treatment initiation. If anti-HCV and anti-HIV treatments have to be administered concomitantly, then nucleoside reverse-transcriptase inhibitors with the lowest mitochondrial toxicity should be preferred.
BACKGROUND
Compared with hepatitis C virus (HCV) monoinfection, concurrent infection with HIV and HCV results in more rapid progression towards cirrhosis and liver failure [1, 2]. This progression is not linear over time and may occur after only a few years [3]. In the pre-HAART era, a low CD4 cell count, longer estimated duration of HCV infection, heavy alcohol intake, and a high necroinflammation score were associated with fibrosis progression [1, 4, 5]. The effect of HAART on liver fibrosis is controversial [6]. For unknown reasons, rapid worsening of fibrosis is still described in some HIV-infected patients with high CD4 cell counts [7-9]. Anti-HCV treatment has been associated with improvement of fibrosis in patients who achieve a sustained viral response and in HCV-monoinfected patients who do not achieve viral response [10, 11]. However, we observed a disturbingly high rate of worsening fibrosis among HIV-HCV-coinfected patients enrolled in a randomized, controlled trial (French National Agency for Research on AIDS and Viral Hepatitis-HC02-Ribavic Study) comparing pegylated IFN-ƒ¿-2b plus ribavirin with conventional IFN-ƒ¿-2b plus ribavirin. Because the rate of sustained viral response to pegylated IFN and ribavirin is lower (<40%) among HIV-HCV-coinfected patients than among HCV-monoinfected patients, a better knowledge of the factors associated with worsening of fibrosis during anti-HCV therapy is important to optimize the safety of this treatment. The aim of the present study was to identify biological and/or clinical features associated with worsening of fibrosis in HIV-HCV-coinfected patients treated for HCV infection.
Methods
This study involved 205 HIV-HCV-coinfected patients who were enrolled in the French National Agency for Research on AIDS and Viral Hepatitis-HC02-Ribavic Study and for whom both pretreatment and posttreatment liver biopsy specimens were available. The results of this trial have been reported in detail elsewhere [12]. In brief, HIV-HCV-coinfected patients who had never received IFN or ribavirin were randomly assigned to receive either weekly subcutaneous injections of pegylated IFN-_-2b (1.5 _g/kg) plus daily ribavirin (800 mg) or thrice-weekly subcutaneous injections of IFN-_-2b (3 MU) plus daily ribavirin (800 mg) for 48 weeks. Patients were eligible for the trial if they had a detectable serum HCV RNA level, liver biopsy performed within the previous 18 months (with examination of the specimen revealing at least mild activity or fibrosis), a CD4 cell count >200 cells/mm3, stable HIV RNA load (<1 log variation in the previous 3 months) at randomization, and stable or no antiretroviral treatment during the previous 3 months. The main ineligibility criteria were active injection illicit drug use and/or self-reported daily alcohol intake exceeding 40 g (women) or 50 g (men) within 3 months before study entry (alcohol consumption before this period was not recorded), as well as decompensated cirrhosis and hepatitis B surface antigenemia.
Biochemical and hematological tests were performed at local laboratories. HCV RNA assays, viral genotyping, and pathological evaluation of liver biopsy specimens were performed at central laboratories.
HCV RNA was detected with a PCR assay (Amplicor 2.0 HCV Monitor; Roche Diagnostics Systems) with a detection limit of 50 IU/mL. HCV RNA levels were measured with a branched-chain DNA assay (bDNA3.0; Bayer Diagnostics) with a detection limit of 615 IU/mL. HCV genotyping was performed by sequence analysis of the 5' untranslated region.
The following data were collected at enrollment: age, sex, body mass index (calculated as the weight in kilograms divided by the square of the height in meters), antiretroviral drug regimen, duration of HIV and HCV infections (defined as the date of first blood transfusion or the date of first injection drug use), Ishak scores, biochemical test results, liver enzyme activities, CD4 lymphocyte count, plasma HIV RNA load, and plasma HCV RNA load. The patients were also evaluated 24 weeks after treatment completion. A sustained viral response (the primary end point for efficacy in the Ribavic trial) was defined as an undetectable serum HCV RNA level by a qualitative PCR-based assay (Amplicor 2.0 HCV Monitor; Roche Diagnostics Systems) 24 weeks after treatment completion (study week 72).
Pretreatment biopsy specimens were compared with those obtained at week 72 by 2 experienced pathologists blinded to the clinical and laboratory findings. Hepatic necroinflammation and fibrosis were evaluated using Ishak's classification, in which activity ranges from 0 to 12 and stage of fibrosis ranges from 0 (none) to 6 [13]. Steatosis was graded as follows: 0, none; 1, <30% of hepatocytes containing fat; 2, 30%-70% of hepatocytes containing fat; and 3, >70% of hepatocytes containing fat. Histological worsening of fibrosis was defined as a score increase of 2 points in patients with a fibrosis stage of <4 and as a score increase of 1 point in patients with stage-5 fibrosis.
The _2 test or Fisher's exact test was used to analyze qualitative variables, and the Mann-Whitney U test was used for quantitative variables. Logistic regression models were used to test possible associations between histological worsening of fibrosis (outcome variable) and pretreatment characteristics (input variables). Characteristics with P values <.10 in univariate analysis were included in multivariate models based on a backward elimination procedure. All statistical tests were 2-sided, with a type-I error of 5%. Mean values are shown with SDs.
Results
A total of 383 patients received at least 1 dose of study medication; 194 patients received pegylated IFN-_-2b plus ribavirin, and 189 received conventional IFN-_-2b plus ribavirin. Paired pretreatment and posttreatment liver biopsy specimens were available for 205 patients, and both were interpretable for 198 cases. The reasons for missing posttreatment results were refusal of biopsy in 124 cases (71.3%), failure to return for follow-up in 46 cases (25.8%), and clotting disorders in 8 cases (4%). There were no significant differences in baseline characteristics between patients who did and did not have posttreatment biopsy performed. The Ishak fibrosis stage was also similar for these groups (mean Ishak fibrosis score, 2.6 +/- 1.2 vs. 2.0 +/- 4.0; p=.88). The rate of sustained viral response was significantly higher among patients with paired biopsy specimens than among the others patients (30.3% vs. 16.7%; p=.002).
The mean duration from the pretreatment liver biopsy to study entry was weeks. The second biopsy was obtained a mean of 30 +/- 16 weeks after the end of treatment. The mean duration between the 2 biopsies was 109 +/- 34 weeks. The mean sizes of the pretreatment and posttreatment liver biopsy specimens were 17.6 +/- 8.2 mm and 18.2 +/- 8.6 mm, respectively.
The clinical and biological characteristics of these 198 patients are summarized in table 1. Most patients were male (70.7%), and the mean age was 39.9 +/- 5.2 years. Injection drug use was the risk factor for HCV transmission in 80.3% of cases. One hundred sixty-one patients (81.3%) were receiving antiretroviral therapy at study enrollment. The mean Ishak scores for necroinflammation and fibrosis were 4.5 +/- 1.4 and 2.6 +/- 1.2, respectively. Steatosis was present in 127 patients (64.1%). One hundred six patients (53.5%) received conventional IFN-_-2b plus ribavirin, and 92 patients (46.5%) received pegylated IFN-_-2b plus ribavirin. The mean duration of anti-HCV therapy was 314 +/- 60 days. Sixty patients (30.3%) had a sustained viral response.
Worsening of fibrosis occurred in 34 patients (17.1%). Pretreatment and posttreatment histologic stages of fibrosis are summarized in table 2. None of the patients had stage-6 fibrosis at baseline. Thirteen patients had a score increase of 3 points. The necroinflammatory grade remained stable from pretreatment to posttreatment in patients who experienced worsening of fibrosis (mean necroinflammatory grade, 4.62 +/- 1.6 and, respectively) but decreased in the other patients (mean necroinflammatory grade, 4.46 +/- 1.31 and 3.97 +/- 1.11, respectively).
In univariate analysis, patients who experienced worsening of fibrosis were more likely than the other patients to be receiving antiretroviral therapy (97% vs. 78%; p=.031) and less likely to have a sustained HCV response (5.9% vs. 35.4%; p<.001) (table 1). No association was observed between worsening of fibrosis and the use of nonnucleoside reverse-transcriptase inhibitor- or protease inhibitor-containing regimens at baseline. Receipt of nucleoside reverse-transcriptase inhibitors (NRTIs) and, specifically, didanosine and stavudine was significantly associated with worsening of fibrosis. The most frequent dual NRTI combinations in patients who experienced worsening of fibrosis were stavudine-lamivudine (11 patients), zidovudine-lamivudine (8), and didanosine-stavudine (8). Triple NRTI-only regimens were not associated with worsening of fibrosis.
The following factors were not associated with the risk of fibrosis worsening: age, sex, mean duration of HIV and HCV infections, AIDS status, baseline CD4 cell count, HIV load <400 copies/mL, mean duration of antiretroviral treatment, HCV transmission group, mean Ishak scores for necroinflammation and fibrosis, steatosis (yes vs. no), severe steatosis (<30% vs. >30%), HCV load, HCV genotype, serum levels of alanine aminotransferase and aspartate aminotransferase, daily ribavirin dose (per kg of body weight), anti-HCV treatment arm, and duration of anti-HCV therapy.
Because all patients who received antiretroviral therapy received NRTIs, the use of antiretroviral therapy was not introduced in multivariate analysis. Didanosine therapy (OR, 3.34; 95% CI 1.39-7.96; p=.007) and failure to have a sustained viral response (OR, 9.05; 95% CI, 2.06-39.66; p=.003) remained statistically significantly associated with worsening of fibrosis (table 3).
Discussion
To our knowledge, this is the first study to analyze risk factors for worsening of fibrosis during anti-HCV therapy in HIV-HCV-coinfected patients. Fibrosis worsened within 2 years in 34 (17.1%) of the patients who started a 48-week course of IFN-_-2b plus ribavirin-based therapy, including pegylated IFN and conventional IFN. Thirteen (38.2%) of these 34 patients had at least a 3-grade increase in fibrosis on the Ishak scale. Worsening of fibrosis was related both to failure to have a sustained viral response (OR, 9.05; 95% CI, 2.06-39.66; p=.003) and to receipt of didanosine-based antiretroviral therapy (OR, 3.34; 95% CI, 1.39-7.96; p=.007). Because the rate of sustained viral response was significantly higher in patients with paired biopsy specimens than in patients without paired biopsy specimens (30.3% vs. 16.7%; p=.002), and because the baseline demographic disease and histological characteristics were similar in the 2 groups, the rate of worsening of fibrosis among the study population was probably not overestimated.
The histological response correlates with sustained viral response status in both HCV-monoinfected and HCV-HIV-coinfected patients [11, 12, 14-16]. Progression of fibrosis slows during anti-HCV therapy in HCV-monoinfected patients, even when there is no sustained viral response, and maintenance therapy with low-dose pegylated IFN is currently under evaluation [11, 16]. In our study, even though anti-HCV therapy lasted a mean duration of >40 weeks, overall, 9 patients experienced an improvement of fibrosis (defined as a decrease in fibrosis stage of 2 points). A high percentage of nonresponders (23.2%; 32 of 138 nonresponders) experienced worsening of their fibrosis; only 6 (4.11%) of these nonresponders experienced an improvement in fibrosis. Therefore, IFN-based therapy does not appear to prevent rapid progression of fibrosis in HCV-HIV-coinfected patients in the absence of sustained viral response. Indeed, 2 HAART-era studies showed a similar rate of rapid fibrosis progression (about 25% of patients experienced worsening of fibrosis in <4 years) among HIV-HCV-coinfected patients who were not treated for HCV infection [7, 8]. Furthermore, our results are in line with the study by Sulkowski et al. [8] in which, among 37 HIV-HCV-coinfected patients treated for HCV infection, a high rate (38.2%) of worsening of fibrosis was seen among the nonresponders (13 of 34 patients experienced worsening of fibrosis).
Liver biopsy is the gold standard for evaluating liver fibrosis, despite the dual pitfalls of sampling error associated with biopsy and interobserver variability between pathologists. These pitfalls were not an issue in the present study, because the mean length of the biopsy specimen was >15 mm and biopsy specimens were evaluated by expert hepatologists.
Didanosine therapy was a major determinant of fibrosis worsening in this study. Studies of the impact of antiretroviral drugs on liver fibrosis have given conflicting results, but most have focused on protease inhibitors or on nonnucleoside reverse-transcriptase inhibitors and not on specific NRTIs [6, 17, 18]. In our study, both didanosine and stavudine therapies were associated with worsening of fibrosis in univariate analysis, and didanosine therapy remained independently associated with worsening of fibrosis in multivariate analysis, whereas the durations of antiretroviral therapy, protease inhibitor exposure, and nonucleoside reverse-transcriptase inhibitor exposure were not associated with fibrosis progression. The mitochondrial toxicity of NRTIs is attributable to inhibition of mitochondrial _ polymerase and indirect damage of some liver functions [19, 20]. In vitro, the mitochondrial toxicity of NRTIs for the Hep G2 cell line has been graded in descending order as follows: zalcitabine, didanosine, and stavudine [19]. In addition, ribavirin is known to enhance didanosine phosphorylation, thereby potentiating both the efficacy and the toxicity of this drug [21]. Since September 2002, didanosine-ribavirin combination therapy has not been recommended [22].
Our data point to the involvement of mitochondrial toxicity in the observed progression of fibrosis. HCV replication, which correlates with the extent of mitochodrial DNA depletion in PBMCs in HCV-HIV-coinfected patients, may be reversed by successful anti-HCV therapy [23]. In contrast, a significant depletion of PBMC mitochondrial content is observed when antiretroviral therapy is combined with anti-HCV therapy, supporting a potential additive effect on mitochondrial toxicity of ribavirin and antiretroviral drugs [23]. Therefore, worsening of fibrosis during anti-HCV therapy could occur because of a synergistic detrimental effect involving the lack of any reduction in mitochondrial depletion in nonresponders and an enhancement of mitochondrial damage caused by combined antiretroviral therapy and anti-HCV therapy. We previously found that concomitant treatment with didanosine and IFN-_-2b or pegylated IFN-_-2b plus ribavirin was associated with an adjusted 46-fold increase in the risk of symptomatic mitochondrial toxicity and an adjusted 8.8-fold increase in the risk of spontaneous hepatic decompensation [24, 25]. Of note, none of the 7 patients who had spontaneous hepatic decompensation in the Ribavic trial were included in this analysis, because posttreatment biopsy samples were not available. Prolonged didanosine exposure has also been implicated in cryptogenic liver disease leading to severe liver complications, particularly variceal bleeding and portal thrombosis in patients who are free of active HCV or hepatitis B virus infection and who have no other common causes of liver disease (e.g., use of alcohol and/or medications) [26]. In some cases, didanosine was given with hydroxyurea, which is known to enhance the intracellular metabolites of didanosine [26].
Dideoxynucleosides, such as didanosine and stavudine, have also been associated with microvesicular steatosis, which is a histological pattern of steatosis often related to mitochondrial toxicity [27, 28]. In our study, steatosis quantification was performed on routinely stained biopsy specimens. Staining allowed for an accurate estimation of macrovacuolar steatosis, which is the most common histopathologic feature of HCV-induced steatosis, but might have underestimated microvesicular steatosis. In HCV-monoinfected patients, the onset and progression of steatosis are strong independent predictors of both the severity and the progression of fibrosis [29, 30]. We previously reported (in patients included in the Ribavic trial) that severe steatosis is related to a high hepatic fibrosis score [31]. However, the presence of steatosis at baseline was not associated with a high risk of fibrosis worsening in our study. This could be explained by the relatively short interval between biopsies, the high prevalence of steatosis, and the absence of accurate information regarding microvesicular steatosis. Necroinflammatory activity has also been linked to fibrosis progression both in HCV-monoinfected patients and in HIV-HCV-coinfected patients [32-34]. In contrast, baseline necroinflammatory grade did not differ between the patients who experienced worsening of fibrosis and the other patients in our study. However, it should be noted that the necroinflammatory grade remained stable in patients who experienced worsening of fibrosis and improved in the other patients. An elevated serum aspartate aminotransferase level has also been associated with fibrosis progression in HIV-HCV-coinfected patients [8]. In contrast, we failed to detect a significant association between baseline serum aspartate aminotransferase level and fibrosis progression. The mean aspartate aminotransferase and alanine aminotransferase levels decreased between the beginning and the end of HCV treatment (week 48) both in patients who experienced progression of fibrosis and in the other patients. However, at the end of HCV treatment, mean aspartate aminotransferase and alanine aminotransferase levels were significantly higher in patients who experienced worsening of fibrosis than in the other patients (mean aspartate aminotransferase level, 1.56 +/- 0.92 times the normal upper limit vs. 1.16 +/- 0.66 times the normal upper limit [p<.001]; mean alanine aminotransferase level, times the normal upper limit vs. 1.09 +/- 1.41 times the normal upper limit [p=.020]).
An independent association has been reported between hepatic fibrosis progression and CD4 cell depletion and/or absence of HIV suppression [4, 5, 32, 34]. A large proportion of our patients had immunovirological control of HIV disease while receiving antiretroviral treatment. The mean CD4 cell count was >500 cells/mm3, and the HIV load was <400 copies/mL in two-thirds of patients. We found no association between the CD4 cell count, the Centers for Disease Control and Prevention class of HIV disease, and a lack of HIV suppression in patients who experienced worsening of fibrosis.
One limitation of this study is that alcohol consumption was not assessed during the treatment and follow-up period. In HIV-seronegative populations, similar to HIV-infected patients, heavy alcohol consumption is an independent predictor of rapid liver fibrosis progression [4]. In the current study, all of the patients should have had limited daily alcohol consumption during the 3 months preceding their enrollment, but changes in alcohol consumption may have influenced fibrosis progression.
In another large trial (AIDS Pegasys Ribavarin International Coinfection Trial) comparing pegylated IFN-_-2a (with or without ribavirin) with IFN-_-2a plus ribavirin in HIV-HCV-coinfected patients, histological worsening (defined as an increase of >2 points in the Ishak-modified histological activity index) was seen in 20% of patients treated with IFN-_-2b plus ribavirin and in 13% of those treated with pegylated IFN-_-2b plus ribavirin; the data regarding fibrosis worsening have not yet been reported. Interestingly, in the AIDS Pegasys Ribavarin International Coinfection Trial, fibrosis worsened in 10% of patients in whom HCV infection was eradicated, compared with 5% of patients in our study [14].
In conclusion, the ultimate benefit of anti-HCV therapy is an improvement in the histologic status of the liver. However, in HIV-HCV-coinfected patients, worsening of fibrosis can occur rapidly, despite HCV therapy and HCV infection cure. Mitochondrial toxicity-involving a synergistic, deleterious interaction between medications for HCV and HIV infections and, particularly, didanosine-seems to play a major role. Therefore, anti-HCV therapy for HIV-HCV-coinfected patients should ideally be administered before antiretroviral treatment is initiated. If anti-HCV and anti-HIV treatments must be administered concomitantly, then the NRTIs with the lowest mitochondrial toxicity should be preferred.
|
|
| |
| |
|
|
|
