Individualizing treatment duration in hepatitis C virus genotype 2/3-infected patients
Alessandra Mangia, MD, Liver Unit, IRCCS ‘Casa Sollievo della Sofferenza’, 71013 San Giovanni Rotondo, Italy
Tel: +01139 882416270
Fax: +01139 882416271
The drugs currently licensed for the treatment of hepatitis C are Peg-Interferon (PEG-IFN) and ribavirin. In recent years, the recommendation to treat hepatitis C virus genotype 2- and 3-infected patients with a fixed 800 mg/day dose of ribavirin in combination with PEG-IFN and for just 24 weeks has been challenged by the concept of tailoring the length of therapy according to on-treatment viral response. Therefore, the objective of the present review was to highlight the different designs of the studies on short treatment duration and the role of wk4-R as a predictor of sustained virological response after an abbreviated course of treatment. The secondary aim was to verify whether we had enough evidence to support the implementation of a short treatment course in subsets of patients with genotype 2 and 3 infection. We will also focus on how drug dosing may have influenced the outcome of treatment. To clarify reasons for discrepant results in the studies so far published, the recently discovered genetic variant near the interleukin 28B gene will be presented and its predictive role will be discussed. Finally, we will face the debated issue of whether the subset of patients with genotype 2 or 3 requires an extended treatment duration.
The approved therapeutic regimen for patients with hepatitis C virus (HCV) genotype 2 and 3 infection includes Peg-Interferon (PEG-IFN) and ribavirin at a fixed 800 mg/day dose for 24 weeks. However, several modifications of this regimen have been directed at an increasing tolerability and compliance and reducing costs. To define which kind of patients with genotype 2 and 3 can be treated for <24 weeks, we should bear in mind that the aim of an individualized short course of treatment is to avoid unnecessary side effects in patients with a very high expectancy of response after a standard 24-week course. Therefore, the essential condition for a successful short-course treatment is to attain, after 12–16 weeks, rates of sustained virological response (SVR) not inferior to those attained with the standard 24 weeks. The current literature has so far provided a number of studies with heterogeneous characteristics that support the efficacy of short treatment in patients with undetectable HCV RNA after 4 weeks of treatment [wk4-R or rapid virological response (RVR)] (1–5) and three studies, one of them very large, named Accelerate, in which patients were randomized to a short course, independently of wk4-R (6–8). The Accelerate study enrolled a large number of patients of different ethnicity and showed that the efficacy of treatment may be compromised when patients are treated for <24 weeks as the relapse is unacceptably high (6). Although it has been shown that the occurrence of relapse after short therapy did not impact negatively on the outcome of a repeat 24-week course (9), the challenge we are facing now is to better refine features of patients with a lower likelihood of relapse after a short course of treatment. In our protocol, evaluating so far 496 individuals who were treated with PEG-IFN α2b (1.5 μg/kg) and weight-based ribavirin for 12 weeks after RVR, 96% attained an end-of-treatment response, but 14% relapsed after the end of treatment (9). This rate of relapse, lower than the rate reported in the Accelerate, can also be considered unacceptable when compared with the 2–7% rates observed after 24 weeks of treatment (2, 4, 6). The question, therefore, is: how to reduce the number of patients with HCV genotype 2 and 3 who experience a relapse after a short course and need to be retreated?
Pooled analysis and meta-analysis combining data derived from different trials have also attempted to clarify the controversial aspects of this issue (10, 11), but this debated question is not definitely answered. Indeed, several differences in the existing trials not only in the design of the studies but also in the regimen of treatment and in the characteristics of the patients enrolled have raised the debate on the suitability of adopting a short treatment in current clinical practice.
The aim of the present review was to evaluate if there was enough evidence to support a personalized treatment in patients with genotype 2 and 3.
Is rapid virological response the single predictor?
A relevant aspect that needs to be clarified when discussing the efficacy of the short course of treatment is the role of wk4-R as a main predictor of SVR. As a premise, we should remember that although the lower detection limit of commercial HCV RNA assays has been changing over the time, most of the studies on short treatment used the same assay with the sensitivity of 50 international units (IU)/ml. Therefore, the definition of wk4-R, at least in this setting refers to samples with HCV RNA levels below this cut off.
Plentiful evidence supports the wk4 response as a primary determinant of treatment duration and outcome, for both patients with easy- and difficult-to-cure genotypes (12, 13); however, the occurrence of relapse after an end-of-treatment response suggests that other negative predictors reduce SVR rates after wk4-R. With an aim to select for a short course patients to whom an over-treatment must be spared because they have a very high likelihood of response after 12–16 weeks and to exclude patients at higher risk of relapse, we should identify these modulators.
Only few studies were powered to evaluate factors potentially associated with relapse in patients with genotype 2 and 3 treated for a course of short duration. In a cohort of 718 patients who received PEG-IFN α2b and weight-based ribavirin for 12 weeks, in accordance with wk4-R, we demonstrated, by comparing baseline features of patients with or without relapse, that platelet counts and body mass index (BMI) are the only two independent predictors of a favourable outcome (9). Indeed, patients with BMI >30 and/or low platelet count (<140 000 mm3) had a rate of relapse as high as 27.5%. The role of these two negative predictors was confirmed in a subgroup analysis of the Accelerate study recently published by Diago et al. (14). In this analysis, the authors demonstrated that, after 16 weeks of treatment, the rates of relapse are 17% higher than after a standard 24-week course in patients with cirrhosis, 11% in patients with higher body weight and 11.5% in patients with older age (14). Finally, evidence derived by the NORDynamiC supports the concept of an association between severity of liver damage and increased rates of relapse after a short course. In that study, including 53% of cases with advanced liver damage, a relapse occurred in up to 33% of patients when the duration of treatment was reduced irrespective of RVR (7). Therefore, higher body weight and the presence of advanced liver damage may be considered strong predictors of relapse and if we face an obese patient or a patient with evidence of advanced liver damage it is advisable not to treat him/her for <24 weeks unless side effects do not occur. Of interest, we have shown that of 43 patients with RVR who relapsed after a short course of therapy, 70% achieved an SVR after a 24-week course of retreatment (9).
Strictly related to the issue of relapse is the relevance of the economical cost-effectiveness of an abbreviated course of treatment. Considering that a 24-week retreatment is feasible and efficient in patients who relapse after a short course of therapy (9), two treatment strategies for patients with genotype 2 and 3 and wk-4R could be implemented. The first one would advice standard treatment duration for all the individuals, whereas the second would implement short therapy only for patients at low risk of relapse. Different economical evaluation of these two strategies have been performed (4, 15) and the reduction of costs has been unequivocally demonstrated with the latter approach. Consequently, treating all HCV genotype 2 and 3 patients for 24 weeks is more expensive than treating for patients with RVR for 12–14 weeks and retreating the 10–14% of patients who experience a relapse for 24 additional weeks (4, 15).
Another relevant issue may be whether levels of viraemia also represent a predictor of relapse in patients with RVR. Thresholds used to define high or low viraemia levels and rates of patients with low viral load in the different studies are reported in Table 1, which shows baseline features of patients enrolled in the different trials on short treatment duration. It was shown by the Accelerate study that patients with high viraemia levels achieve a lower rate of SVR when treated for 16 weeks independent of their infection with genotype 2 or 3 (6). SVR rates were, respectively, 83 and 82% in low viraemic (≤400 000 IU/ml) genotype 2 patients treated for the standard or the abbreviated therapy, and 73 vs. 58% in high viraemic individuals. Among genotype 3 patients, the overall SVR rates after a short or standard course of therapy were 62 vs. 66% respectively. These rates increased up to 81 vs. 81% in the subset of patients with low viraemia, and decreased to 52 vs. 59% in those with high viraemia (6). By contrast, in the North-C study, in patients with genotype 2 infection no difference could be appreciated in SVR rates attained after standard or short duration both in low (100% in each group) or high viraemic patients (90 and 95% respectively) (4). Results, in genotype 3 patients, were slightly lower independently of viraemia levels. Indeed, in patients with viraemia levels below 400 000 IU/ml, rates of response were comparably high (80 and 100%) after a course of short or standard duration. Similarly, in patients with higher levels of viraemia, no differences were observed in the rates of response (86% for both treatment arms) (4). The different conclusions reached by the two studies may be reconciled by analysing separately SVR response in patients with or without wk4-R in the Accelerate. As shown in a meta-analysis of data on 2275 patients, once wk4-R has been achieved, genotype 3- as well as genotype 2-infected patients respond to a short course of treatment irrespective of basal viraemia (16). By contrast, viraemia results as a negative predictor of SVR in studies in which the impact of HCV RNA levels has been analysed independently of wk4-R (6, 7).
Table 1. Features of patients and schedules of treatment in the studies evaluating the efficacy of short treatment after wk4-R (2–4) or independently of wk4-R (5, 6)
|Years of publications||2005||2005||2008||2007||2008|
|Total number of patients||283||153||428||1469||382|
|Genotype 2 no.||213||40||85||701||104|
|Genotype 3 no.||70||123||343||728||276|
|Mean age (years)||48.1||40||39.6||46.6||42|
|Male gender (%)||56||68||63||51||58|
|Mean body weight (kg)||69.4||76.6||78||81.6||78|
|Mean body mass index||26||–||25.3||27.7||26|
|Bridging fibrosis/cirrhosis (%)||24||NA||15||24||37|
|Viraemia cut off level (IU/ml)||800 000||800 000||400 000||400 000||400 000|
|Low viraemic patients (%)||35||–||28.3||21||27|
|Patients with RVR (no)||178||142||298||–||–|
|Type of Peg-Interferon||2b||2a||2b||2a||2a|
|Dosage of Peg-Interferon (mcg)||1.0||180||1.5||180||180|
|Dosage of ribavirin (mg/day)||1000–1200||800–1200||800–1400||800||800|
The principle of non-inferiority trials
Some peculiar aspects owing to the different design of the trials may also help in explaining discrepancies in the conclusions from one study to another. The choice of the non-inferiority margin represents the main critical issue in trials with a non-inferiority design. Within studies on different duration of treatment for patients with genotype 2 and 3, different non-inferiority margins have been considered: 7% in the study by Shiffman, 10% in the study by Dalgard and 12% in the study by Mangia (2, 4, 6). These differences might have probably influenced the discrepant conclusions reached by the Authors. In the North-C trial, for example, the wide 95% confidence interval (CI) 1.7–17.7 of the 9.6% difference in SVR rates between standard and short treatment did overlap the planned 10% non-inferiority margin and prevented the author from concluding that 14 weeks is equivalent to 24 weeks of treatment (4). The negative results attained in this study, related to the fact that the lower boundary of CI felt outside the a priori hypothesis of a non-inferiority, also drive attention to the appropriateness of performing in these studies an intention-to-treat (ITT) rather than a per-protocol (PP) analyis.
In a non-inferiority trial, the role of ITT population should be considered carefully. Rates of patients discontinuing treatment may differ after an abbreviated treatment course or after a course of standard 24 week. As in the ITT analysis, patients lost to follow-up are considered as patients not achieving SVR; the higher the rate of patients lost to follow-up in the different studies, the higher the number of patients for whom we are assuming a final outcome that is probably not true. Practically, we do not expect short treatment to give more side effects than standard treatment and because we need guidance on what to do with patients with RVR who actually took treatment for 12–16 weeks without discontinuing during these first weeks of treatment, we should consider that a PP may be more appropriate in evaluating the effectiveness of a reduced treatment duration in HCV genotype 2- and 3-infected patients. If we consider the North-C study, the results of PP analysis showed a difference in SVR rates between the two different treatment duration of only 7%. This difference lower than the 9.6% reported after the ITT analysis, would have supported the conclusions that the efficacy of short treatment course is not inferior to a 24-week standard duration course (4).
One of the most relevant aspects that differs in the studies on short treatment is the dosage of both PEG-IFN and ribavirin. Weight-based or flat 800 mg doses of ribavirin daily have been variably used. In general, weight-based dosages have been used in combination with PEG-IFN α2b (1, 2, 4) whereas flat dosages have been combined with PEG-IFN α2a (6, 7). There are no studies on short treatment evaluating the dosage of 800 mg in combination with PEG-IFN α2b; by contrast, three studies investigated the efficacy of a weight-based dosage in combination with PEG-IFN α2a (3, 5, 8).
In the initial studies on abbreviated treatment in patients with genotype 2 and 3 (1, 2), PEG-IFN α2b was combined with weight-based dosages of ribavirin, at 1.5 mcg/kg/week in the study by Dalgard and at 1.0 mcg/kg/week in the study by Mangia. At that time, it had not been shown that 1.0 mcg/kg of PEG-IFN α2b attained results comparable with those attained using 1.5 mcg/kg. Now, the Ideal study has proven that, even in difficult to treat patients, the efficacy of 1.5 mcg/kg is not different from that of 1.0 mcg/kg (17). Therefore, we can assume that, in the contest of patients with genotype 2 and 3, the rate of SVR may be rather influenced by the exposure to ribavirin than to interferon. Results of the three studies realized with the use of weight-based ribavirin and PEG-IFN α2a, suggested that weight-based dosages may help in preventing a relapse (3, 5, 8) (Table 2). Of relevance, in the study by Yu et al. (8) conducted in Asian patients infected with genotype 2 only, SVR up to 95% were shown with relapse rates lower than 5%. At variance, in the two studies performed with PEG-IFN α2a and fixed dosages of ribavirin (5, 6), the rates of relapse amounted to 31 and 33% after 16 weeks and to 18 and 12% after 24 weeks respectively.
Table 2. Relapse rates according to wk4-R in the studies evaluating the efficacy of short treatment
Recently, a higher ribavirin concentration at day 29 during treatment has been shown associated with an increased likelihood of achieving SVR in patients with HCV genotype 2/3 infection (18).
The results of the North-C study adopting the combination of PEG-IFN α2b and 800–1400 mg of ribavirin are apparently disappointing. The authors showed 81.1% rates of SVR after a 14-week course of treatment and of 90.7% after a standard course. Their conclusions were that short treatment was inferior to standard treatment (4). Although, as reported previously, this conclusion reflects a wide CI whose lower boundary was felt outside the a priori hypothesis of a non-inferiority, the difference between the standard and the short treatment duration, in the rates of relapse, was not higher than 5.5%. This finding is in keeping with our results showing only 5.7% difference in the rate of relapse between standard and short duration (2), and therefore it can be assumed to indirectly confirm the need of weight-based dosage of ribavirin in patients undergoing short treatment course. Definitely, out of six randomized trials published as full papers on this topic, four used weight-based dosages of ribavirin (2–4, 9) and two flat dosages (6, 7). In three out of four studies using a weight-based dosage, the equivalence between standard and short course of treatment was demonstrated.
Genetic host factors
Another possible moderator of outcome that has recently gained consideration, appears to be the genetic variability of interleukin (IL)28B gene coding for interferon λ. Higher frequencies of T/G single nucleotide polymorphism at rs809917 or C/T at rs12979860 near IL28B gene were observed in patients with genotype 1 infection who respond to interferon treatment (19–21). Of interest, these genetic variants can explain the lower rates of SVR reported in many clinical trials in patients of African American origin as compared with Caucasian (22, 23).
We have recently investigated whether this polymorphism adds to the already known predictors of SVR in patients with genotype 2 and 3 as well as in patients with genotype 1. In a cohort of patients infected with genotype 2 and 3 already characterized, the CC polymorphism at the IL28B gene resulted associated with SVR, although the effect size appears attenuated when compared with genotype 1 (24). In detail, the CC genotype was found in 82% of patients with SVR whereas the TT genotype was detected in 57% and the CT genotype at an intermediate frequency of 75% [odds ratio (OR) 1.8, 95% CI=1.2–2.7, P=0.0046]. The role of IL28B has also been investigated in patients with genotype 2 and 3 by others (25, 26), but the original findings of our study was the discovery that the genetic variants influence SVR in non-RVR patients (24). Indeed, differences in SVR rates between patients with favourable or unfavourable genotype were more relevant among patients who failed to attain RVR. In this subgroup of slow responders, among patients with SVR, CC was found in 87% of patients, whereas TT was observed in only 29% and CT in 67% (OR=4, 95% CI 1.9–8.5). We can speculate that the evaluation of the genetic variability of IL28B gene will help in future decisions of how to personalize treatment in patients with HCV RNA still detectable at week 4 after the start of treatment. It would also be interesting to evaluate the impact of this genetic polymorphism on the Accelerate study with an aim to explain the 18% rate of relapse registered after standard treatment higher than the 2–5% rate reported in patients treated for 24 weeks in European studies (2–6).
Which patients require an extended treatment
The final question is about who should not be absolutely treated for <24 weeks. The answer to this question is easily given. An RVR is usually obtained in nearly two-thirds of patients with genotype 2 and 3, although this rate may vary accordingly with the sensitivity of the assay used to assess whether HCV RNA is undetectable or not at week 4 (3). In the remaining one-third of patients with detectable HCV RNA at week 4, SVR is not higher than 50% (2–7). In particular, in the studies randomizing patients at baseline independently of RVR, the rates of SVR range from 26 to 41% (6, 7). Whether there are chances of increasing rates of wk4-R in this subset of slow responders is not clear, but there are no doubts that these patients need 24 weeks of treatment (27). One approach to increase SVR for patients with a slow on-treatment response has been the intensification of antiviral treatment. For patients infected with genotype 1 who did not attain wk4-R, prolonging treatment duration up to 72 weeks increased SVR rates by reducing the occurrence of relapse (28). A retrospective analysis of a pooled database of non-RVR patients infected with genotype 2 and 3 by registration studies showed that, in patients treated for 24 or 48 weeks, prolonging the duration of treatment may be of some advantage (29). In patients with genotype 3 only, the role of an extended treatment up to 36 weeks has been recently, prospectively, evaluated. Among patients with no wk4-R, rates of SVR were numerically but not significantly higher in the longer treatment duration arm as compared with the standard 24-week arm (30). These results are only apparently in contrast with the evidences generated by Willems (29), because as in that study genotype 2 and 3 were combined and analysed together, a direct comparison between the two studies is not feasible.
A second option to increase the rates of response in patients without wk-R has been the administration of higher doses of ribavirin. As shown by the results of Win-R study, rates of SVR were significantly lower in patients with genotype 2 and 3 receiving a flat 800 mg/day as compared with those attained with weight-based ribavirin dosages combined with PEG-α2b (31). In this respect, we have also recently shown that only dosages equal to or higher than 15 mg/kg/day are associated with rates of RVR in patients with genotype 2 and 3 (32). Whether the genetic evaluation of patients with genotype 2 and 3 infection for IL28B gene variants will better address the issue of treatment of patients without wk4-R is under investigation by our group.
In conclusion, for patients with genotype 2 and 3 infection, considerable progress has been made to reduce side effects and costs, and to increase tolerability and compliance of antiviral treatment. These aims were pursued modifying the approved fixed regimen of PEG-IFN and ribavirin given for 24 weeks accordingly with on treatment response and we currently have all the means to individualize treatment for genotype 2- and 3-infected patients. Slightly lower SVR rates, after a short course of treatment in comparison with the standard in patients with wk4-R and an absence of factors predictive of relapse, are of minor clinical relevance because the first objective of our work was to cure the highest number of patients with the lower number of side effects using the currently available combination treatment. The advent of genetic factors may now allow us to select patients who can still be cured with the current standard of PEG-IFN and ribavirin combination and to spare wane treatment to the others, waiting for more effective treatments. The task of curing patients who are not responding at week 4 with a combination of small molecules acting as direct antiviral inhibitors may represent the next challenging objective for patients with negative genetic predictors.