Tuesday, May 24, 2011

Estimating the Likelihood of Sustained Virological Response in Chronic Hepatitis C Therapy

  • Abstract and Introduction

  • Methods

  • Results

  • Discussion

  •  ,
    From Journal of Viral Hepatitis

    Estimating the Likelihood of Sustained Virological Response in Chronic Hepatitis C Therapy

    S. Mauss; D. Hueppe; C. John; J Goelz; R. Heyne; B. Moeller; R. Link; G. Teuber; A. Herrmann; M. Spelter; S. Wollschlaeger; A. Baumgarten; K.-G. Simon; N. Dikopoulos; T. Witthoeft
    Posted: 05/24/2011; J Viral Hepat. 2011;18(4):e81-e90. © 2011 Blackwell Publishing
    Discussion Only See Links For Complete Text
    This well-documented large cohort study of patients treated with peginterferon plus RBV in clinical practice identified a number of independent positive and negative predictors of SVR including HCV genotype, baseline viral load, older age, and baseline GGT and platelet levels. Many of these factors have been identified previously under study conditions, and this cohort study confirms their relevance in daily clinical practice. In addition, the current study suggests that, in patients matched for potentially confounding baseline characteristics, treatment with peginterferon alfa-2a is a predictor of achieving SVR compared with peginterferon alfa-2b. The same predictive factors were also identified when period of treatment (pre- vs post-new treatment guidelines) was taken into account.

    HCV genotype 2 infection, HCV genotype 3 infection and low baseline viral load are important and well-known positive predictors of response in CHC identified in a number of clinical studies[2–4,15,16] and were also the main predictors of SVR in this analysis. In the current study, age greater than 40 years was a significant negative predictive variable for SVR. Age greater than 40–45 years has also been shown to be predictive of SVR following interferon-based therapy in clinical studies[2,3,17–19] and in cohort analyses.[20,21]

    In pivotal trials of peginterferon plus RBV, the presence of cirrhosis was associated with a lower rate of SVR,[2,3] and absence of cirrhosis has been shown to be an independent positive predictor of SVR.[22] In the current study, only 5.4% of patients with available biopsy data had cirrhosis. Although this is a relatively low proportion compared with some clinical trials[3,22] that included up to 15% of patients with cirrhosis, it is in line with other trials[4] and with other cohort studies reflecting routine clinical practice.[23–25] Overall, fewer patients with cirrhosis or moderate fibrosis at baseline in the current study achieved an SVR compared to those with no/mild fibrosis, although as baseline biopsies were only available for a limited number of patients the significance of these figures could not be determined. However, surrogate markers of fibrosis and cirrhosis included in the analysis did show significance in univariate and multivariate analysis. Elevated GGT levels are frequently seen in patients with CHC and have been shown to be related to hepatic steatosis, advanced fibrosis and insulin resistance.[26,27] Baseline GGT above ULN was a strong negative predictive factor for SVR in both analyses in the current study. Low pretreatment serum GGT levels have been shown to be predictors of response to treatment with conventional IFN[28,29] and have also been shown to be significantly and independently associated with SVR following peginterferon plus RBV in multivariate regression analysis.[16,30,31] Berg et al. [16] found a threshold baseline value of ≤0.93 times ULN yielded an odds ratio of 5.7 (95% CI, 3.2–10.0) making it one of the most significant baseline predictors of SVR along with genotype and viral load. Thrombocytopaenia has been shown to be a valid surrogate marker for portal hypertension, but also advanced liver fibrosis.[32,33] Low baseline platelet levels (<150 000/μL) were negatively associated with SVR in the current analysis. This association was also reported by Backus et al..[22] Although often attributed to hypersplenism, there is evidence to suggest that other mechanisms may be important in thrombocytopaenia associated with HCV infection[33] including impaired platelet production in the failing liver together with abnormalities in thrombopoietin degradation.[34]

    A low baseline level of cholesterol was found to be a negative predictive factor for SVR in univariate analysis in the current study, with a best cut-off for ROC of 187 mg/dL. Unfortunately, cholesterol could not be included as a variable in the multivariate analysis owing to the small number of patients with available data. Baseline cholesterol has been reported to be an independent predictor of SVR in other studies.[22,35] Low cholesterol levels are known to be associated with HCV infection and have been associated with more severe liver disease,[36–38] which may correlate with a poorer response to treatment.[39] Further investigation into the importance of cholesterol as a baseline predictor of response is warranted.

    Recently, several trials which aimed at directly comparing treatment response to peginterferon alfa-2a with that of peginterferon alfa-2b have been published. In the largest of such studies to date, the IDEAL (individualized dosing efficacy vs flat dosing to assess optimal pegylated interferon therapy) study,[40] SVR rates in genotype 1-infected patients were comparable between the two different peginterferons in combination with RBV.[41] As a result of major differences in RBV dosing and RBV dose adjustment in IDEAL, a direct comparison of the pegylated interferons alone in this study is impossible. However, the rates of rapid virological response (viral load below limit of detection at week 4), complete early virological response (viral load below limit of detection aft week 12) and week 24 response were consistently higher with peginterferon alfa-2a plus RBV than with peginterferon alfa-2b plus RBV in IDEAL, suggesting that peginterferon alfa-2a plus RBV may show greater antiviral efficacy. Despite the higher end of treatment response to peginterferon alfa-2a plus RBV, rates of SVR were similar between the two different peginterferons, because of a higher relapse rate in the peginterferon alfa-2a group. However, in a subanalysis of IDEAL looking at African American patients, a particularly difficult-to-treat population, treatment with peginterferon alfa-2a plus RBV seemed to result in a higher end of treatment and SVR rate compared with peginterferon alfa-2b plus RBV (45%vs 32% and 26%vs 23%, respectively).[42] In a recent prospective, randomized, independent Italian study including 320 patients, where initial dose of RBV and strategies for dose reduction were consistent between the two peginterferons, significantly more patients treated with peginterferon alfa-2a plus RBV achieved SVR compared with those treated with peginterferon alfa-2b plus RBV (68.7%vs 54.4%, respectively; P = 0.008).[43] Similarly, the randomized open-label Milan safety tolerability (MIST) study found that 66% of patients treated with peginterferon alfa-2a compared with 54% treated with peginterferon alfa-2b achieved an SVR (P = 0.02).[44] Of relevance to the current study, the PROBE observational study enrolled 1351 patients with genotype 1 virus at 167 treatment centres in Italy. Again, the trial found a greater chance for SVR for patients treated with peginterferon alfa-2a compared to those treated with peginterferon alfa-2b (41%vs 34%, respectively; P = 0.004).[45] In an earlier large observational study of US veterans, treatment with peginterferon alfa-2a was found to be a positive predictor of SVR compared with treatment with peginterferon alfa-2b, although overall SVR rates were relatively low (31%vs 24%, respectively).[22] Taken together, results from these studies and from the current study suggest that patients treated with peginterferon alfa-2a may be more likely to achieve an SVR compared with those who receive peginterferon alfa-2b. The reason for this apparent difference may be multifactorial. Differences in pharmacokinetics and viral kinetics may impact SVR,[46,47] while differences in adverse event profiles may affect adherence and thus SVR. In the current study, there was a significantly higher discontinuation rate with peginterferon alfa-2b compared with peginterferon alfa-2a. This reflected the significantly higher rate of withdrawal owing to virological nonresponse in the peginterferon alfa-2b group, which is likely to have contributed to the shorter mean duration of therapy in genotype 1/4 patients treated with peginterferon alfa-2b. Overall, a similar number of patients were withdrawn from the treatment for tolerability issues in both groups. Fewer patients treated with peginterferon alfa-2b received ≥80% of the recommended dose of both peginterferon and RBV compared with those treated with peginterferon alfa-2a. This was possibly a reflection of the higher discontinuation rate in patients treated with peginterferon alfa-2b, as the proportion of patients who received dose reductions was similar in both groups. However, a cumulative dose ≤80% of peginterferon or RBV was not significantly associated with SVR on multivariate analysis; therefore, lower dose alone does not seem to explain the difference in SVR achieved using peginterferon alfa-2a vs peginterferon alfa-2b.

    On-treatment decline in haemoglobin levels was not associated with treatment outcome in multivariate analysis, particularly in genotype 1 patients. However, the dosing regimens of ribavirin used in this analysis were completely at the discretion of the participating centres, and therefore, we cannot completely rule out the possibility that some genotype 1/4 patients did not receive or adhere to the recommended doses of RBV.
    Although retrospective studies are not a substitute for prospective studies, they provide an important source of information reflecting treatment under 'real-life' conditions, over longer time periods and in a more diverse patient and physician population. In particular, they provide the opportunity to assess whether success rates (in this case SVR) obtained in clinical trials can also be obtained under real-life conditions where patients are exposed to factors not encountered in clinical trials. The main strengths of the PRACTICE study are its large size and the long observation period covered. The study consisted of an unselected cohort of patients with chronic hepatitis C treated under routine, 'real-life' conditions in contrast to the highly selective and controlled randomized trial situation. The data derived from such cohorts may more closely reflect clinical practice than results obtained from the optimized setting of clinical studies. The use of matched pairs of patients to compare treatment regimes should help overcome problems of interpretation arising owing to confounding baseline characteristics. However, despite matching patients, it is not possible to completely rule out a selection bias in different centres when assessing suitability for treatment and in selecting the specific interferon for the treatment regimens. Clinicians were fully responsible for supplying data to the PRACTICE data set, and there was therefore a possibility that selection bias may have occurred.

    However, each treatment centre confirmed that all patients treated within the prespecified observational period were included in the data collection and that there was no patient selection for documented patients. As with all retrospective studies, another potential bias may be caused by missing data, as only data that were documented by the clinicians in the patient records were available for inclusion in the study. Although the core data for the evaluation of the HCV therapy were available in more than 89% of the patients, there were limited data for some parameters such as cholesterol. This reflects the difference between the clinical trial situation, where a wide range of parameters are systematically collected, and the 'real-world' situation, in which only the most relevant to treatment are generally routinely measured. Despite the more limited range of parameters measured, clinicians would appear to have managed their patients well in the PRACTICE cohort given the high rate of SVR achieved.

    In this large, multicentre study of patients treated with peginterferon plus RBV over seven years, baseline predictors of response on multivariate analysis were similar to those reported in prospective studies [2–4,15,16]. Significant positive baseline predictors of SVR were genotype 2/3 infection and low viral load, while negative predictive variables were age greater than 40 years, high baseline GGT and low baseline platelet levels. In addition, in patients matched in terms of baseline characteristics, this study suggests that treatment with peginterferon alfa-2a in routine daily practice is a positive predictor of SVR compared to peginterferon alfa-2b.

    No comments:

    Post a Comment