| |
Risk factors for hepatic decompensation in patients with HIV/HCV coinfection and liver cirrhosis during interferon-based therapy
|
| |
| |
AIDS: Volume 18(13) 3 September 2004
Mauss, Stefan; Valenti, Williama; DePamphilis, Jeanb; Duff, Frankb; Cupelli, Lisab; Passe, Sharonb; Solsky, Jonathanb; Torriani, Francesca Jc; Dieterich, Douglasd; Larrey, Dominiquee
From the Center for HIV and Hepatogastroenterology, Düsseldorf, Germany, the aCommunity Health Network, Rochester, New York, bRoche Nutley, New Jersey, the cUniversity of California, San Diego, California, dMount Sinai Medical Center, New York, USA, and the eLiver Unit, Saint Eloi Hospital, Montpellier School of Medicine and INSERM 632, Montpellier, France.
ABSTRACT
Objective: Hepatic decompensation was reported from two recent trials (APRICOT and RIBAVIC) assessing interferon (IFN)-based treatment of hepatitis C virus (HCV) in HIV/HCV-coinfected patients. This paper identifies risk factors associated with hepatic decompensation in APRICOT.
Methods: APRICOT is a randomized, partially-blinded, controlled trial comparing treatment with peg-IFN a-2a 180 ug once weekly plus ribavirin/placebo 400 mg twice daily with IFN a-2a 3 million units three times weekly plus ribavirin 400 mg twice daily for 48 weeks in a total of 859 patients. Multiple logistic regression analysis was performed comparing the baseline characteristics of those cirrhotic patients who experienced decompensation with those of the other cirrhotic patients enrolled.
Results: Fourteen patients, all cirrhotic, experienced hepatic decompensation during the study. The incidence in the cirrhotic subgroup of the study was 10.4% (14/134). Six of the 14 patients died as a result of hepatic decompensation. The risk factors associated with hepatic decompensation were increased bilirubin, decreased haemoglobin, increased alkaline phosphatase or decreased platelets, and treatment with didanosine. Markers of viral replication, histological activity, cellular immune status or HCV-therapy, treatment with ribavirin and pegylated versus non-pegylated IFN were not associated with hepatic decompensation.
In total, 14 cases of hepatic decompensation were observed in 859 participants (1.6%). All cases were in patients with cirrhosis. The incidence of hepatic decompensation in this cirrhotic subgroup (14/134) was 10.4%. Their distribution reflects the randomization ratio 1 : 1 : 1 with five patients of 45 (11%) on PEG-IFN a-2a (40KD) with ribavirin, 5 of 44 (11%) on PEG-IFN a-2a (40KD) without ribavirin, and four of 45 (9%) on IFN a-2a plus ribavirin. The time to decompensation was <= 12 weeks in nine patients and <= 24 weeks in 13 patients. All decompensated patients had previously received antiretroviral treatments and 13 patients were receiving antiretroviral drugs at the onset of decompensation. In only two of 14 patients an increase in ALT levels preceded hepatic decompensation. There was no increase of HCV RNA in any of the 14 patients before hepatic decompensation. A decrease of HCV RNA of > 1 log10 before hepatic decompensation was observed in six of 14 patients. Three of the six patients achieved a decrease of HCV RNA below the limit of detection (< 50 IU/ml). There was no difference in HCV genotype distribution between cirrhotic patients with hepatic decompensation (HCV genotypes 1 and 4, 10/14; HCV genotypes 2 and 3, 4/14) and cirrhotic patients without hepatic decompensation (HCV genotypes 1 and 4, 85/120; HCV genotypes 2 and 3, 35/120). At the end of the 24-week treatment free follow-up period, hepatic decompensation had resolved in four patients, remained unresolved in four patients, and six patients had died as a result of hepatic decompensation.
Our analysis evaluated the baseline conditions of only the 13 patients with reports of deterioration within the first 24 weeks of treatment. Decompensation in the remaining patient after 24 weeks was felt to reflect natural progression of liver disease and was excluded from the analysis.. A sensitivity analysis conducted with this patient included did not alter the results.
Multivariate modelling resulted in two best-fitting models containing four variables: higher total bilirubin, lower haemoglobin, didanosine treatment, and higher alkaline phosphatase or, alternatively, lower platelets.
Variables associated with HIV or HCV viral replication, histological activity, cellular immune status or treatment of chronic hepatitis C, such as HIV-RNA, HCV-RNA, HAI score, CD4-positive cells, treatment with ribavirin and pegylated versus non-pegylated interferon were not associated with hepatic decompensation.
Conclusions: The results from APRICOT indicate that the overall risk of hepatic decompensation in HIV/HCV-coinfected patients without cirrhosis receiving IFN-based treatment is low. In contrast, patients with markers of advanced cirrhosis, despite the absence of a history of hepatic decompensation, should be monitored closely during IFN-based therapy, because they are at risk of hepatic decompensation. Treatment with antiretrovirals such as didanosine may increase the risk further.
AUTHOR DISCUSSION
Our analysis was aimed at identifying risk factors shared at baseline by those cirrhotic patients in whom hepatic decompensation occurred. Most risk factors associated with hepatic decompensation in this study are markers of advanced cirrhosis. Thrombocytopenia, anaemia, and elevated bilirubin are well established predictors of advanced cirrhosis. Elevated alkaline phosphatase as a marker of cholestasis resulting from intrahepatic cholangolitis has been reported in HIV/HCV-coinfected individuals. The association between didanosine and hepatic decompensation suggests that an interaction between antiretroviral drugs and HCV treatment must be considered. Although the sample size of this study is rather small and may not allow definitive conclusions, for the time being didanosine should be avoided or replaced in patients with frank or incomplete cirrhosis.
In the multivariate analysis, treatment with ribavirin, a drug with potential interaction with didanosine on the level of intracellular phosphorylation, was not associated with an increased risk of hepatic failure. In addition treatment with a specific formulation of interferon (pegylated versus conventional) was not detected by multiple logistic regression analysis as a risk factor to these patients.
Because of the lack of an untreated control arm, which was not considered standard of care at the time of the trial design, a comparison to the natural progression of HIV/HCV infected individuals is not possible.
The early occurrence of hepatic decompensation in the course of interferon-based therapy suggests the possibility of a direct treatment effect such as hepatotoxicity or immune activation resulting in the clearance of HCV-infected liver cells. However the majority of patients experiencing decompensation in this series did not experience an increase in ALT, a finding that would have been expected in cases of an immune mediated flare similar to those reported for hepatitis B infected patients clearing the virus.
Hepatic decompensation has been recently reported in another large study (RIBAVIC) evaluating HCV treatment in HIV/HCV coinfected patients. In the RIBAVIC study, a trial that compared PEG-IFN a-2b (12KD) plus ribavirin with conventional IFN a-2b plus ribavirin, a rate of hepatic decompensation of 7.9% (32/416) was reported for the overall study population. About 15% of the patients in RIBAVIC had liver cirrhosis. However the proportion of patients with liver cirrhosis in the decompensating group has not been reported so far. While a detailed analysis of these patients is not currently available, these findings suggest that hepatic decompensation in this population is not limited to a particular interferon-based combination. In addition this observation from RIBAVIC corroborates that the observation of hepatic decompensation in APRICOT is not an isolated phenomenon and that hepatic decompensation is not uncommon in interferon-based trials in this population, wherein liver disease can progress rapidly.
In conclusion, in APRICOT hepatic decompensation was observed only in HIV/HCV-coinfected patients with markers of advanced cirrhosis while receiving interferon-based treatment. In contrast, no hepatic decompensation was noted in HIV/HCV-coinfected patients without cirrhosis. Consequently, HIV/HCV-coinfected patients with markers of advanced liver cirrhosis, despite the absence of a history of hepatic decompensation, should be monitored closely during interferon-based therapy, because they are at risk of hepatic decompensation. Treatment with specific antiretrovirals such as didanosine may increase the risk further. Didanosine should be avoided in patients with advanced liver disease, particularly when alternative antiretroviral treatment options are available. Because of the lack of other treatment options for hepatitis C, and the limited access of HIV/HCV-coinfected patients to liver transplantation, patients with liver cirrhosis should not be excluded from interferon-based therapy in general. Considering that most risk factors for hepatic decompensation in this study are markers of progressive cirrhosis, coinfected patients with histological progression of their chronic hepatitis C should be considered for interferon-based treatment before they develop late stage liver disease.
INTRODUCTION
Co-infection with hepatitis C virus (HCV) and HIV is common because of similarities in the routes of transmission. Co-infection with HIV accelerates the progression of HCV liver disease to cirrhosis and increases the mortality rate compared to patients monoinfected with HCV or HIV. In the era of effective antiretroviral combination therapy for HIV, hepatitis C is one of the leading causes of death in populations with a substantial proportion of HIV/HCV-coinfected patients. Successful treatment of chronic hepatitis C may improve the life expectancy of people living with HIV.
HCV mono-infected patients with advanced liver fibrosis or cirrhosis have a substantial morbidity and mortality. The rate of spontaneous hepatic decompensation in patients with chronic hepatitis C and asymptomatic cirrhosis is 3-4% per year. Therefore, treating the HCV infection in these patients with a high risk of disease progression is clinically indicated. The side-effect profile of interferon (IFN)-based therapy is well characterized, but hepatic decompensation has not been reported from recent large trials in HCV mono-infected patients with early-stage cirrhosis (Child-Pugh A) treated with IFN-based therapy.
In contrast, we observed hepatic decompensation in HIV/HCV-coinfected patients during a randomized, partially blinded, controlled trial comparing treatment with PEG-IFN a-2a once weekly with or without ribavirin versus conventional IFN a-2a three times weekly plus ribavirin. The present analysis was conducted after unblinding of treatment in order to identify baseline risk factors, including antiretroviral and specific HCV therapy, associated with hepatic decompensation within the first 24 weeks of therapy.
METHODS
In the AIDS Pegasys Ribavirin International Coinfection Trial (APRICOT) HIV/HCV-coinfected patients were randomized into three groups: treatment with PEG-IFN a-2a 180 ug once weekly plus placebo/ribavirin 400 mg twice daily or conventional IFN a-2a 3 million units three times weekly plus ribavirin 400 mg twice daily for 48 weeks. Liver disease stage, classified according to the Ishak Modified Histological Activity scoring system, was assessed by a liver biopsy performed <= 15 months prior to entry. Patients without cirrhosis, and those with incomplete or frank cirrhosis (Child-Pugh grade A), were eligible for enrolment (n = 134).
A review of all adverse events and laboratory abnormalities that were classified as serious or resulted in treatment discontinuation was conducted for all patients who received study medication and had at least one post-dose observation (n = 859). The review focused on identifying signs and symptoms of liver disease progression.
HCV RNA was determined quantitatively with Cobas AMPLICOR HCV Monitor v2.0 (Roche Molecular Systems, Branchburg, New Jersey, USA) with a lower limit of detection of < 600 IU/ml and qualitatively with AMPLICOR HCV v2.0 (Roche Molecular Systems) with a lower limit of detection of < 50 IU/ml.
Statistical analyses of patient baseline risk factors are based on the 13 patients who decompensated within the first 24 weeks of the study. A univariate analysis, followed by a multiple logistic regression analysis, was performed while ribavirin treatment was still blinded, comparing the characteristics of the cirrhotic patients experiencing early hepatic decompensation (n = 13) with those of cirrhotic patients without evidence of decompensation (n = 120).
Baseline variables chosen for statistical analysis were age, sex, weight, histological activity of hepatitis [hepatitis activity index (HAI) grading score and HAI staging score], serum HCV RNA (log, linear), bilirubin, albumin, prothrombin time normalized ratio (PT-INR), platelets, haemoglobin, alanine aminotransferase (ALT), aspartate aminotransferase, alkaline phosphatase, CD4-positive cell count, CD8-positive cell count, HIV-RNA (log, linear), antiretroviral drugs and classes of antiretroviral drugs. Multiple logistic regression analysis was performed on all variables from the univariate analysis with P < 0.20. The forward selection procedure [P <= 0.25 to enter model (adjusted chi-square statistic)] was kept to equal to or fewer than four variables because of a limited sample size. Then, a best subsets procedure [likelihood score (chi-square) statistic] identified two best-fitting models of n-variable combinations, where n was the number of variables from the forward selection procedure. The model calibration was assessed with the Hosmer-Lemeshow test and discrimination was characterized by c-statistic.
After the study was unblinded, a second multiple logistic regression analysis was performed, after adjusting for the baseline risk factors identified in the original analysis, to assess the effect of HCV treatment on hepatic decompensation. Wald chi-square statistics were used to test the effect of ribavirin versus placebo and the effect of PEG-IFN a-2a versus conventional IFN a-2a.
|
|
| |
| |
|
|
|
