Meta-analysis: the outcome of anti-viral therapy in HCV genotype 2 and genotype 3 infected patients with chronic hepatitis


Dr A. Andriulli, Division of Gastroenterology, “Casa Sollievo della Sofferenza” Hospital, IRCCS, San Giovanni Rotondo, Italy.


Background  Anti-viral therapy seems more successful in HCV genotype 2 than genotype 3-infected patients.

Aim  To report sustained virological response (SVR) rates for HCV-2 and HCV-3 infection.

Methods  Meta-analyses were carried out on SVR data on 2275 patients treated for 24 weeks in eight individual trials and on 968 patients with rapid virological response (RVR) treated for 12–16 weeks or 24 weeks in four studies.

Results  After 24 weeks of therapy, SVR rates were 74% and 68%, respectively, for HCV-2 and HCV-3 genotype patients. Among high viraemics, SVR rate in HCV-2 infection (75%) differed from the 58% value in HCV-3 infection. Among low viraemic patients, respective rates were 79% and 75%. In RVR patients treated for 12–16 or 24 weeks, SVR rates in HCV-2 infection were 83% and 84%, respectively, and in HCV-3 infection 84% and 86%. In patients without RVR treated for 24 weeks, SVR was higher in HCV-2, with a 17.8% weighted difference.

Conclusions  Twenty-four weeks of therapy should remain standard duration for HCV-2 and low viraemic HCV-3 patients. In RVR patients, HCV-3 patients respond to short-treatment as well as HCV-2 patients, irrespective of basal viraemia. Patients without RVR may need longer treatment than the recommended 24 weeks.


In HCV-infected patients with chronic hepatitis, response rates to interferon-based therapy are heterogeneous and HCV genotypes have been concordantly shown as the most important baseline variable influencing treatment outcomes. For ‘difficult-to-treat’ genotypes (i.e. genotype 1), full dosage of anti-viral drugs for an extended treatment duration will eventuate in sustained virological response (SVR) in 40–45% of patients.1, 2 Conversely, in ‘easy-to-treat’ genotypes, such as HCV-2 and -3, SVR is observed in about 80% of patients with just 24 weeks of therapy, a lower dose of ribavirin, and either standard interferon or peg-interferon (Peg-IFN).3

Many clinical trials have assessed treatment response rates among HCV-2 and HCV-3 patients as a combined group and, thus, were unable to detect subtle differences in treatment efficacy between the two of them. In HCV-3 patients, reduced SVR rates after treatment duration of 24 were originally reported by Zeuzem et al.: in patients with HCV-3 or HCV-2 infection, therapy eventuated in 79% and 93% SVR rates, respectively, a 14%-difference that was associated with higher amount of hepatic steatosis among former patients.4 This finding would imply that virological response rates should be given for single genotypes rather than for any arbitrary combination of them. However, subsequent studies have not consistently reported a different outcome of therapy between the two genotypes.5, 6 Recently, treatment has been further shortened from 24 to 12–16 weeks on the concept of viral response during the initial weeks of therapy. Indeed, patients with a rapid clearance of the virus from the blood at treatment week 4 [rapid virological response (RVR)] achieved similar or marginally lower SVR rates after treatment duration of 12–16 or 24 weeks of combination therapy.7–10 Whether the short-treatment duration will benefit equally the two genotypes remains to be further assessed.

This meta-analytical review focuses on studies on combination therapy of Peg-IFN with ribavirin where 24 weeks of therapy was administered to patients with chronic HCV infection, with the aim to assess SVR data separately for HCV-2 and HCV-3 infected patients. In addition, a second meta-analysis was carried out and restricted to RVR patients who received treatment for a short duration of therapy, with the intent to verify whether this regimen would serve equally well patients infected with HCV genotypes 2 and 3.


Literature search

A search was conducted for clinical trials on the use of Peg-IFN and ribavirin combination therapy given for a 24-week or shorter treatment duration in patients with chronic HCV infection. To identify all relevant evidences, an a priori decision was made to search for all papers reported in peer reviewed journals regardless of results and publication status (full papers or abstracts). A two-stage search strategy was adopted. First, the PubMed database from January 2001 to July 2007 was searched under the search terms of Peg-IFN AND HCV infection, with randomized clinical trials as limits. Second, the reference lists of published reports and abstracts of two international meetings in hepatology (American Association for the Study of Liver Diseases and European Association for the Study of the Liver) were manually searched for additional citations.

Study selection

The search yielded information on eight clinical trials where duration of therapy lasted 24 weeks,4–6, 8, 10–13 and four trials presenting SVR data for RVR patients after a short (12–16 weeks) or standard (24 weeks) treatment duration.9, 10, 14, 15 Exclusion criteria were trials that used 48-week’ duration of therapy or a per-protocol analysis, if data were included in a subsequent publication or could not be retrieved separately for HCV-2 or HCV-3 patients, studies referring to patients with acute HCV infection, HIV co-infection, haemophilia, chronic renal failure, post-transplant setting, reports giving data for only one of the two genotypes, those with an insufficient sample size (<20 individuals) for each genotype, or retrospective collection of patients. The authors were contacted when necessary to obtain further detailed data as appropriate.

Data extraction

Original data collection forms were used to abstract from single studies the following information: sample size, patient characteristics, dosage and type of Peg-IFN administered, dosage of ribavirin. Two reviewers (A.A. and A.M.) screened papers and disagreements were resolved through discussion. Validity of retrieved studies was assessed and scored using the following modification of the Jadad’ scale,16 which considers three items: randomization (1 point if yes, 2 points if the method to generate the sequence of randomization was described and appropriate), double-blinding (1 point if yes, 2 points if the method of double-blinding was described and appropriate), and description of withdrawals and dropouts (1 point). To the original scale, we added the domain of sample size estimation, scored with 1 point if sample size was calculated, and 2 points if the estimated number of patients was eventually enrolled into the study. Therefore, the maximum possible score, which was 5 in the original Jadad’ scale, is 7 points in our extended scale. The validity assessment was performed by the same two reviewers who extracted the data from the original papers: reports with extensive flaws were those with a ≤2 score.


Formal statistical tests for heterogeneity of the odds ratios (ORs) were performed using the Cochrane Q test, homogeneity being assumed with a P value >0.05. Primary outcome was the estimation of SVR rates in patients with HCV-2 and HCV-3 infection. Secondary outcomes were the evaluation of SVR rates according to baseline viral load and type of Peg-IFN compounds being administered. A third outcome was to explore SVR rates in HCV-2 and HCV-3 patients with or without RVR.

Pooled ORs with 95% confidence intervals (CI) were calculated using the Peto method.17 A statistically significant result was assumed when the 95% CI of pooled OR did not include one. Whenever statistical significance was detected, absolute risk reduction with 95% CI and the number needed to treat with 95% CI were calculated. All meta-analytical data were analysed according to intention-to-treat analyses. Publication bias assessment (PBA), calculated by the Klein’s method,18 expresses how many unpublished studies with negative or null results are needed to influence the results of the meta-analysis. Moreover, funnel plots of individual study results were created by plotting the studies ORs against sample size to detect asymmetry in the distribution of trials, with the intent to establish whether additional small studies may have been conducted but not published because of negative results. All procedures and calculations used in the meta-analyses were made following the methodology reported elsewhere.19


SVR rates after standard-treatment duration

Outcome data after a 24-week course of anti-viral therapy in 2275 patients infected with HCV-2 or HCV-3 were retrieved from eight clinical trials.4–6, 8, 10–13 Similarities and differences in either designs and therapeutic regimens among the studies are summarized in Table 1. All selected studies had a quality score >5, according to our modification of the Jadad’ scale. In four studies,4, 8, 10, 11 only HCV-2 and HCV-3 patients had been enrolled, whereas in the remaining trials all genotypes were studied, but data could be retrieved separately for the two genotypes of interest. A majority of surveyed trials have taken as low viraemia a cut-off value of ≤600 000 IU/mL. For the meta-analysis, SVR results from original studies were taken irrespective of the type of Peg-IFN administered, the dose of the Peg-IFN alfa 2b (1.0 vs. 1.5 μg weekly) or the dose of ribavirin (weight-based or fixed dose of 800 mg).

Table 1.   Clinical trials of therapies in HCV-infected patients with genotypes 2 and 3
 ReferencePublication typeDesignPeg-interferonRibavirin (mg)No. of patients% of HCV-2 and HCV-3Low viraemia (IU/mL)Quality score
  1. RCT, randomized clinical trial.

  2. * Mean dosing.

  3. † 1.0 μg/kg.

  4. ‡ Post hoc analysis of data from Hadziyannis et al.3 and Fried et al.2 studies.

Working group5AbstractRetrospectiveα2a
800 weight-based277
≤600 0005
Mangia7FullRCTα2b†1000–120070100≤600 0005
Meyer-Wyss12  α2b80022740≤800 0005
Rizzetto‡6LetterRCTsα2a 800–1200240100≤800 000 
Shiffmann10FullRCTα2a 800731100≤400 0007
Zeuzem4FullProspectiveα2b 800–1400224100≤600 0001
Zeuzem13FullRCTα2a 80021227≤800 0005

Summary results for outcomes of therapy are presented in Table 2. Statistical tests failed to detect heterogeneity among studies (Q value = 7.334; P = 0.40). Patients with HCV-2 infection had a higher chance of attaining an SVR after 24 weeks of anti-viral therapy than those harbouring HCV-3 infection: 74.5% vs. 68.7% (Figure 1). Only three original trials4, 10, 11 produced significantly different SVR rates between the two genotypes. After pooling data from the eight studies, the weighted difference amounted to 8.7% (CI 5.1–12.3), with a pooled OR of 1.49 (CI 1.23–1.80). The PBA was 25, and the test for funnel plot asymmetry was nonsignificant (α = 0.10; P = 0.90). By restricting the analysis to the four reports where only HCV-2 and HCV-3 patients were enrolled,4, 8, 10, 11 the meta-analytic results were more robust with a weighted difference of 10.5% (CI 6.1–14.9), and a pooled OR of 1.62 (CI 1.30–2.04). All but one5 study presented SVR rates according to baseline viraemia: among the 900 low viraemic patients, SVR rates were homogeneous at the statistical test (Q = 5.82; P = 0.44), and higher in HCV-2 compared with HCV-3 patients: 79.1% vs. 75.4% (Table 2). No original trial produced significantly different rates, whereas significance emerged only after pooling of data: weighed difference of 7.1% (CI 1.9–12.3) and an OR of 1.50 (CI 1.08–2.09). The PBA was two, and the test for funnel plot asymmetry was nonsignificant (α = 0.13; P = 0.84). In 1001 high viraemic patients, the statistical test showed significant heterogeneity among the six considered trials (Q value = 17.4; P = 0.004). By using a random effect model, SVR was attained in 74.7% of HCV-2 and in 57.8% of HCV-3 patients, with a weighted difference of 24.9% (CI 12.8–37.0), an OR of 2.36 (CI 1.80–3.09), and a PBA of 46 (Figure 2). The test of funnel plot asymmetry approached the level of significance (α = 1.91; P = 0.07), suggesting the potential for publication bias.

Table 2.   Pooled odds ratios of and risk differences in sustained virological response rates between HCV genotype 2 and genotype 3 infected patients when treated with Peg-interferon and ribavirin for 24 weeks
 No. of trialsSVR/all
No. of pts (%)
Odds ratios (CI)*Differences in percentages (CI)
Genotype 3Genotype 2
  1. CI, confidence intervals.

  2. * Among the several subgroup analyses, summing up numbers of patients does not equal the number of all patients because information was missing from some studies.

All patients*8841/1225 (69)782/1050 (74)1.49 (1.23–1.80)8.7 (5.1–12.3)
Low viraemic patients7405/537 (75)287/363 (79)1.50 (1.08–2.09)7.1 (1.9–12.3)
High viraemic patients6289/500 (58)374/501 (75)2.36 (1.80–3.09)24.9 (12.8–37.0)
Peg-interferon alfa-2a*
 All patients4420/606 (69)418/556 (75)1.42 (1.10–1.84)7.1 (2.1–12.1)
 Low viraemic Patients3180/227 (79)106/127 (83)1.38 (0.80–2.40)5.6 (2.4–13.7)
 High viraemic patients3177/327 (54)269/359 (75)2.56 (1.86–3.52)22.5 (15.7–29.3)
Peg-interferon alfa-2b*
 All patients5411/619 (66)364/494 (74)1.71 (1.30–2.24)12.0 (6.7–17.2)
 Low viraemic Patients4225/310 (73)181/236 (77)1.57 (1.04–2.38)8.2 (1.4–15.0)
 High viraemic patients3112/173 (65)105/142 (74)1.91 (1.14–3.18)15.4 (5.4–25.5)
Figure 1.

 Forest plot of sustained virological responses in HCV-3 and HCV-2 patients after anti-viral therapy given for 24 weeks.

Figure 2.

 Forest plot of sustained virological responses in high viremic HCV-3 and HCV-2 patients after anti-viral therapy given for 24 weeks.

In combination with ribavirin, Peg-IFN α-2a was administered in four studies5, 6, 10, 13 and Peg-IFN α-2b in five studies;4–6, 11, 12 trial results were statistically homogeneous: Q values of 2.14 (P = 0.54) and of 1.86 (P = 1.00), respectively. The meta-analytical data are shown in Table 2: all ORs and weighted differences of percentages were statistically significant, when either considering the overall population of patients enrolled in these trials, and splitting patients according to baseline viraemia. The most relevant difference was noted among high viraemic patients, with SVR rates of 74% and 54% in HCV-2 and HCV-3 patients after Peg-IFN α-2a, and of 74% and 65% after Peg-IFN α-2b.

SVR rates in patients with rapid (week 4) virological response

In patients with RVR, outcome of therapy was assessed in four trials (Table 3). Three reports9, 10, 15 randomized patients to short (14–16 weeks) or standard 24-week treatment duration, whereas a single study14 reported SVR after an abbreviated course of therapy only. Among the 957 patients who received therapy for 12–16 weeks, SVR rates were similar when comparing HCV-3 and HCV-2 infected patients (Table 4, Figure 3). Meta-analytical data were homogeneous (Q value = 3.11; P = 0.38), with a nonsignificant test for funnel plot asymmetry (α = 0.52; P = 0.51). Similarly, no significant differences emerged after stratifying patients for either baseline viraemia levels (Table 4) or type of Peg-IFN administered in combination with ribavirin (data not shown). In four studies8–10, 15 where 760 RVR patients received treatment for 24 weeks, the test for heterogeneity approached the level of significance (Q value = 5.68; P = 0.06), with significant asymmetry in the funnel plot (α = 1.62; P = 0.02). Pooled SVR rates in HCV-2 patients were not different from those in HCV-3 patients (OR = 1.01, CI 0.66–1.54) when considering the overall population of patients or splitting data according to viraemia at baseline. Among 490 patients without RVR, SVR rates were higher in HCV-2 than in HCV-3 patients: 62.3% and 46.0%, respectively, with a 17.8% weighted difference (CI 8.7–27.0) and a pooled OR of 2.06 (CI 1.40–3.02).

Table 3.   Clinical trials of therapies in patients with HCV genotypes 2 and 3 chronic infection and rapid (week 4) virological response
ReferencePublication typeDesignPeg-interferonsRBV (mg)Therapy length (weeks)No. of patientsLow viremia (UI/mL)Quality score
  1. RCT, randomized clinical trial.

  2. * 1.0 or 1.5 μg/kg.

Andriulli et al.14FullMeta-analysisα2b*800–140012–14403<600 0005
Dalgard et al.15FullRCTα2a 800–140016
<400 0005
Shiffman et al.10FullRCTα2a 80016
<400 0007
VonWagner et al.9FullRCTα2a 800–120016
<800 0005
Table 4.   Sustained virological responses in HCV genotype 2 and genotype 3 patients with rapid (week 4) virological response, treated with Peg-interferons and ribavirin for short (12–16 weeks) or standard (24 weeks) duration, and for patients without rapid virological response who received treatment for 24 weeks
 No. of trialsNo. of pts (%)Odds ratiosDifferences of percentages (CI)
Genotype 3 SVR/allGenotype 2 SVR/all (CI)*
  1. CI, confidence intervals.

  2. * Among the several subgroup analyses, summing up numbers of patients does not equal the number of all patients because information was missing from some studies.

Rapid virological response/short-treatment duration
 All patients*4412/495 (84)384/462 (83)0.94 (0.66–1.36)2.7 (−6.9 to 1.6)
 Low viraemic patients399/113 (88)71/76 (93)1.88 (0.64–5.57)8.4 (2.2–14.5)
 High viraemic patients3108/131 (82)76/87 (87)1.47 (0.67–3.21)3.7 (−5.4 to 12.8)
Rapid virological response/standard-treatment duration
 All patients*4342/396 (86)310/364 (84)1.01 (0.66–1.54)1.0 (−5.5 to 3.5)
 Low viraemic patients265/69 (94)17/17 (100)1.26 (0.14–11.23)0.2 (−5.9 to 5–5)
 High viraemic patients275/93 (81)30/32 (94)2.94 (0.99–8.72)12.6 (1.3–23.9)
No rapid virological response
 All patients3134/291 (46)124/199 (62)2.06 (1.40–3.02)17.8 (8.7–27.0)
Figure 3.

 Forest plot of sustained virological responses in HCV-3 and HCV-2 patients who cleared the virus at treatment week 4, after a short-treatment course (12–16 weeks).


There has been considerable progress in customizing anti-viral therapy for patients with chronic HCV infection. So far, therapy has been tailored according to the genotypes of the infecting virus, broadly categorized into easy- and difficult-to-treat subtypes. However, recent evidence indicates a different outcome of therapy even among the former genotypes, with HCV-3 patients responding less well than HCV-2 patients.4 The results of the present meta-analysis re-affirm the original observation by Zeuzem et al.,4 and estimate an 8.7% (CI 5.1–12.3) difference in SVR rates after 24 weeks of treatment between the two genotypes. This finding seems to be robust as the PBA to revert the result was 25.The difference was not influenced by the type of Peg-IFN administered, but was considerably affected by viraemia levels at baseline. Among low viraemic patients, the therapeutic outcome was marginally better in HCV-2 than in HCV-3 patients with corresponding SVR rates of 79% and 75%. No individual trial came out with statistical significance between the two previous subgroups of patients and only after pooling data from seven clinical trials a 7.1% weighted difference (CI 1.9–12.3) emerged. Of relevance, among highly viraemic patients, SVR was attained in 75% of HCV-2 and in 58% of HCV-3 infected patient, with a weighted difference of 24.9% (CI 12.8–37.0).

These figures may affect current understanding of the category of easy-to-treat patients: low viraemic genotype 2 and 3 patients as well as high viraemic genotype 2 patients should remain in this category, as marginal differences in SVR rates detected in the present meta-analysis may not be clinically relevant. Contrarily, high viraemic genotype 3 patients respond to the recommended 24-week regimen of therapy less satisfactorily than the previously referred subgroups: the disappointing 58% rate of SVR after 24 weeks of therapy would suggest that they should no more be considered as easy-to-treat. However, from post hoc analyses of studies where these patients were given longer treatment duration (48 weeks) SVR rates varied from 52% to 78%.6, 11

Attaining viral clearance at treatment week 4 has emerged as a more robust predictor for SVR than HCV genotypes.7, 8 The question whether RVR patients with HCV-3 infection respond equally well to an abbreviated course of therapy as do HCV-2 patients remains unanswered. By inspecting the meta-analytical results (Table 4), pooled SVR rate did not differ when HCV-2 and HCV-3 patients achieving HCV-RNA negative levels at week 4 were treated for an abbreviated course of therapy. Interestingly, the outcome of the abbreviated course of therapy was not different even in high viraemic patients, a finding that is at odds with the common belief that high viraemic HCV-3 patients would be best served by the standard 24-week treatment duration. Of further relevance, RVR patients who were treated with the standard 24-week duration achieved the same percentages of SVR rates as those who received treatment for shorter duration, a finding that remained consistent even after considering separately in the analysis low and high viraemic patients. Unfortunately, in the ACCELERATE study, SVR rates are yet not available separately for low and high viraemic patients with RVR and hence, data from this trial could not be incorporated into our meta-analysis to substantiate the findings.

In patients without RVR, 24 weeks of therapy would result in an SVR rate of 62% for genotype 2 and of 46% for genotype 3 (Table 2). These SVR rates would be considered insufficient and require further studies on longer treatment duration. It is worth mentioning that also these categories of patients should no more be considered as easy-to-treat subjects.

In conclusion, the recommended 24-week duration of therapy should remain standard for HCV-2 infected patients and low viraemic HCV-3 patients, but not for high viraemic HCV-3 patients. Once the RVR status has been attained, HCV-3 patients respond to short-treatment duration (12–16 weeks) as well as HCV-2 patients, irrespective of basal viraemia. High viraemic HCV-3 patients as well HCV-2 and HCV-3 patients who do not clear the virus at treatment week 4 might achieve improved SVR rates with longer treatment duration than the currently recommended 24-week regimen.


Declaration of personal and funding interests: None.