| |
Utility of week-4 viral response to tailor treatment duration in hepatitis C virus genotype 3/HIV co-infected patients
|
| |
| |
AIDS: Volume 21(4) 19 February 2007 p 477-481
Crespo, Manuela; Esteban, Juan Ib; Ribera, Estebana; Falco, Vicenca; Sauleda, Silviac; Buti, Mariab; Esteban, Rafaelb; Guardia, Jaimeb; Ocana, Inmaa; Pahissa, Alberta
From the aInfectious Diseases Department, Spain.
bLiver Unit, Department of Medicine, Spain.
cCentre de Transfusio i Banc de Teixits, Institut Catala de la Salut, Hospital Universitari Vall d'Hebron, Universitat Autonoma de Barcelona, Spain.
Abstract
Objective: To investigate the utility of a week-4 virological response for sustained response prediction in hepatitis C virus (HCV) genotype 3/HIV-co-infected patients treated with interferon and ribavirin for 24 weeks.
Methods: Using a real-time polymerase chain reaction-based quantitative assay (COBAS AmpliPrep-COBAS-TaqMan 48; Roche Diagnostics) we retrospectively analysed samples obtained at baseline and weeks 4 and 12 from a subset of 35 HCV genotype 3-HIV co-infected patients enrolled in a randomized comparative trial of peginterferon α-2b versus interferon α-2b both in combination with ribavirin.
Results: In an intention-to-treat analysis, 78% of patients treated with peginterferon and 53% of those receiving standard interferon achieved a sustained virological response (SVR) Overall, at 4 weeks, 49% of patients had HCV RNA < 50 IU/ml and 63% had < 600 IU/ml. Of these rapid responders 88 and 86% achieved a SVR, respectively, with only one patient relapsing among end-of-treatment responders. In contrast, only 44 and 31% of patients with a week-4 HCV RNA ≥ 50 or ≥ 600 IU/ml achieved an SVR, respectively, with relapse rates of 33 and 50%, respectively. In multivariate logistic regression analysis a serum HCV RNA level below 600 IU/ml at week 4 was the strongest independent predictor of SVR (odds ratio, 11.3; 95% confidence interval, 1.7 to 75.0; P = 0.012).
Conclusion: Monitoring early viral response may be useful to tailor the duration of treatment among patients with HCV genotype 3/HIV-co-infection. Patients whose HCV RNA falls below 600 IU/ml at 4 weeks are at low risk of relapse after 24 weeks of combination therapy.
Introduction
Well-designed studies to assess the optimal duration of treatment for hepatitis C virus (HCV), genotypes 2 or 3, infection in HIV-co-infected patients have not yet been performed. Recently reported data, however, suggest higher relapse rates occur among co-infected patients with HCV genotypes 2/3 who have been treated for 24 weeks [1,2] than in those receiving 48-week treatment schedules [3-5]. On the basis of these data, a statement from a European Consensus Conference recommended that all HIV-co-infected patients be treated for 48 weeks, irrespective of the HCV genotype [6]. However, this recommendation might underestimate the higher risk of toxicity, the lower compliance and the cost associated with extended treatment regimens and ignores the fact that a significant number of HCV genotype 2/3 co-infected patients who clear the virus during the first month of treatment achieve a sustained viral response (SVR) with a 24-week treatment schedule [7]. In the present study we reanalysed the utility of the week-4 virologic response through retrospective testing of stored samples with a commercially available real-time polymerase chain reaction (PCR) quantitative test, to individualize treatment duration for HCV genotype 3 infection in HIV-co-infected patients.
Methods
We have recently conducted a randomized trial in our hospital with the aim of comparing the efficacy of interferon (IFN)α-2b versus pegylated interferon (peg-IFN)α-2b both in combination with ribavirin (RBV) for the treatment of HCV/HIV-co-infected patients. The methodology has been reported elsewhere [7]. Briefly, 121 HCV/HIV-co-infected patients, naive for HCV treatment, were stratified by sex and genotype and randomized, between January 2001 and April 2003, to receive peg-IFNα-2b (1.5 μg/kg per week) or IFNα-2b [3 million International Units/three times in a week (MU/tiw)], both in combination with a fixed dose of RBV (800 mg/day). Patients with genotype 2 (n = 1) or 3 (n = 41) were treated for 24 weeks. The study protocol was approved by the Institutional Review Board and conducted in accordance with the Declaration of Helsinki and Good Clinical Practices guidelines. Blood samples were handled and stored under optimal conditions for HCV RNA quantization as previously recommended [8]. Serum HCV RNA level at baseline and at 4, 8, and 12 weeks of treatment was determined with a non-commercial TaqMan-based real-time reverse transcriptase (RT)-PCR test, with a detection limit of 100 IU/ml and a dynamic range of 5 logs [9]. A commercial qualitative PCR test with a detection limit of 50 IU/ml (Cobas Amplicor HCV v2.0; Roche Molecular Systems, Branchburg, New Jersey, USA) was used at end of treatment and end of follow-up to assess sustained response.
In order to investigate the most appropriate cut-off to define a rapid viral response (RVR) at week 4 of treatment and assess its utility for SVR prediction in genotype 3 HCV/HIV-co-infected patients treated with interferon and RBV combination therapy for 24 weeks, we have retested stored serum samples obtained at baseline, and at weeks 4 and 12 of treatment, with a commercially available real-time PCR test combined with an automated platform for RNA extraction (COBAS AmpliPrep-COBAS-TaqMan 48; Roche Molecular Diagnostics, Pleasanton, California, USA), with a detection limit of 15 IU/ml and a dynamic range of 43 to 68 000 000 IU/ml.
Levels of HCV RNA at weeks 4 and 12 of treatment were used to construct receiver operating characteristic (ROC) curves, and establish the best cut-off point of predicted probability for SVR. Stepwise and backward multiple logistic regression analyses were used to identify baseline variables that were predictive of week-4 viral response and of SVR. Non-normally distributed baseline continuous variables underwent logarithmic transformation. Additionally, week-4 viral response was entered into the model both as a categorical value (< 50, < 600 or ∇HCV RNA ≥ 2 log10 IU/ml) and as a continuous variable to assess it as independent predictor of SVR.
Results
Trial results on both efficacy and safety for the entire cohort have been reported elsewhere [7]. Samples for the current testing were available from 35 of 42 genotype 3 patients enrolled in the trial. Baseline characteristics of these patients are summarized in Table 1. Most patients were on HAART regimens according to currently recommended guidelines. Remarkably, 80% of them had baseline viral loads above 800 000 IU/ml. Three patients withdrew from treatment because of intolerance or were lost to follow up. The RBV dose was reduced because of anaemia in three patients and was definitively stopped in another three because of symptomatic hyperlactataemia. The peg-IFN dose was reduced in four patients because of severe flu-like symptoms. In an intention-to-treat analysis SVR was achieved in 23 (66%) patients; 78% (14/18) of those in the peg-IFNα-2b group and in 53% (9/17) of those in the standard IFNα-2b group [relative risk (RR), 1.47; 95% confidence interval (CI), 0.88-2.45; P = 0.164].
Overall, irrespective of the treatment arm, 49% (17/35) of the patients retested had HCV RNA levels < 50 IU/ml at week 4 of treatment. Of these rapid viral responders (RVR) 88% (15/17) achieved a SVR, and only one (6%) of 16 with an end-of-treatment response (ETR) relapsed. In contrast, 12 (67%) of 18 patients with HCV RNA at week-4 ≥ 50 IU/ml, achieved an ETR and 4 (33%) of the latter relapsed (RR, 2.0; 95% CI, 0.28-14.2; P = 0.85) accounting for a significantly lower SVR rate of 44% (8/18) (RR, 0.50; 95% CI, 0.29-0.87; P = 0.012) The corresponding figures of ETR and SVR when RVR was defined using the cut-off value of 600 IU/ml, are shown in Fig. 1.
Sensitivity, specificity and positive and negative predictive values of the week-4 response for SVR prediction were 0.65, 0.83, 0.88 and 0.56, respectively, for the 50 IU/ml cut off value and 0.83, 0.75, 0.86 and 0.69, respectively, for the 600 IU/ml cut off. The area under the ROC curve constructed using the week-4 HCV RNA levels was 0.812 (95% CI, 0.650-0.973).
Of the baseline variables [age, sex, treatment group, HCV RNA level, aspartate aminotransferase (AST), alanine amintransferase (ALT), ϒ-glutamil transpeptidase, RBV dose in mg/kg body weight, and CD4 T-cell count] entered in a multiple logistic regression analysis none was a predictor of the viral response at week 4 of treatment (using both cut-off levels 50 and 600 IU/ml as dependent variables) whereas baseline AST level was a weak predictor of SVR [odds ratio (OR), 0.987; 95% CI, 0.976-0.998; P = 0.026]. When the week-4 viral response was included in the multivariate analysis, both AST level and week-4 viral response were independent predictors of SVR. The best final model included AST level (OR, 0.983; 95% CI, 0.969-0.997; P = 0.020) and week-4 viral response with the cut-off of 600 IU/ml (OR, 11.3; 95% CI, 1.70-75.0; P = 0.012).
Discussion
Kinetics of early viral response to peg-IFN and RBV treatment among HIV-co-infected patients seem to be similar to those observed in mono-infected patients [10], and a 2-log drop at week 12 can be confidently used as stopping rule [3-5]. However, recommendations that co-infected patients be treated for 48 weeks irrespective of HCV genotype are based on cross comparisons between historical and prospective studies, which were not designed to address optimal treatment duration and it remains to be proven to what extent HIV-positive patients with HCV genotypes 2 or 3 benefit from schedules longer than 24 weeks. In fact, in a recent Italian multicenter trial, which randomized 128 HCV genotype 2 or 3-co-infected patients to treatment with peg-IFN and weight-adjusted RBV for either 28 or 48 weeks, failed to show an increased efficacy of prolonging therapy to 48 weeks (Zanini B, et al. personal communication) Indeed, in this study, relapse rates were similar in both groups because of the high drop out rate among patients assigned to the longer treatment group, resulting in identical SVR rates (61%) in both arms. These data underline the importance of compliance in treatment outcome and that adherence may be especially relevant among co-infected patients. Hence, strategies favouring compliance and avoiding unnecessary extension of treatment in very early responders might decrease side effects associated with pharmacological interactions and might be more effective than currently recommended strategies. In this regard, several studies in HCV genotypes 2 or 3 mono-infected patients have shown that patients with an RVR, defined as undetectable HCV RNA at treatment week 4, achieve similar SVR rates with 12-16-week treatment schedules of peg-IFN and weight-adjusted RBV doses, than those treated for 24 weeks [11-13].
We have recently reported sustained viral response in 71% of HCV genotype 2/3 HIV-co-infected patients treated with peg-IFNα-2b plus RBV for 24 weeks [7], a proportion similar to those reported in large trials in which such patients received treatment for 48 weeks [3-5], and showed that HCV RNA clearance at treatment week 4, as assessed with a non-commercial real-time PCR technique, was highly predictive of SVR (Crespo M, et al., personal communication).
Herein, we have confirmed our previous findings using a commercially available quantitative test. Co-infected patients with HCV genotype 3 showing an HCV RNA decrease below 600 IU/ml at 4 weeks of treatment, are at minimal risk for relapse and achieve a very high rate of sustained response after 24 weeks of therapy. Using the cut off value of commonly used qualitative tests (50 IU/ml) to define RVR was less sensitive with a similar positive predictive value. Consequently, in multivariate analysis, the best prediction model for SVR included, in addition to baseline AST level, the 600 IU/ml cut-off, which is the detection limit of a widely used quantitative test.
Our study has several limitations. First, we have performed a post hoc analysis in a small subset of patients from a randomized trial not designed to address the utility of the early viral response to individualize treatment duration. Second, because of lack of histological data, we were unable to evaluate the influence of fibrosis stage on the predictive value of the RVR. In this regard, it is possible that the negative correlation between higher baseline AST values and SVR might be related to more advanced liver fibrosis, as recently reported [14]. Third, because of the limited number of patients studied, the most appropriate cut off to define a RVR in these patients cannot be established. Nonetheless, because of their high sensitivity and wide dynamic range, real-time PCR tests, such as the one used in the present study, should allow accurate determination of the predictive values of different cut offs in larger patient cohorts. Finally, because RBV appears to have a direct antiviral effect that synergizes with that of IFN [15,16] and higher RBV doses have been shown to increase RVR rates in co-infected patients [17], the utility of early monitoring of viral response might have been underestimated in our study were fixed RBV doses were used.
In summary, monitoring early viral response might be useful for the management of genotype 3 HCV infection in HIV-co-infected patients. Patients showing an HCV RNA decrease below 600 IU/ml at 4 weeks are at low risk of relapse after 24 weeks of combination therapy. The reduced exposure to treatment will probably translate into a better benefit-risk ratio among co-infected patients. Our results on the utility of an RVR as a guide to tailor treatment duration in these patients warrant further investigation through well-designed randomized trials.
|
|
| |
| |
|
|
|
