Is an À la Carte Combination Interferon Alfa-2b Plus Ribavirin Regimen Possible for the First Line Treatment in Patients With Chronic Hepatitis C?

Transcription

1 Is an À la Carte Combination Interferon Alfa-2b Plus Ribavirin Regimen Possible for the First Line Treatment in Patients With Chronic Hepatitis C? THIERRY POYNARD, 1 JOHN MCHUTCHISON, 2 ZACHARY GOODMAN, 3 MEI-HSIU LING, 4 AND JANICE ALBRECHT 4 FOR THE ALGOVIRC PROJECT GROUP Randomized trials have shown the enhancement of efficacy with interferon alfa-2b and ribavirin (IFN-R) in comparison with interferon monotherapy (IFN) as first line treatment of chronic hepatitis C. Further definition of response based on disease, patient, and treatment characteristics is needed to determine the degree of benefit for the various patient subgroups. The aim of this study was to answer this question by analyzing the data from 1,744 naive patients included in trials that compared 24- or 48-week IFN-R treatment. Response factors were identified by logistic regression and receiver operating characteristics curves. Five independent characteristics were associated with a sustained loss of hepatitis C virus (HCV) RNA (F100 copies/ml) 24 weeks after the end of treatment: genotype 2 or 3, baseline viral load less than 3.5 million copies/ml, no or portal fibrosis, female gender, and age younger than 40 years. There was a significant advantage for IFN-R in comparison with IFN alone whatever the combination of factors. The most efficient strategy is to treat all patients for 24 weeks. If the 24-week polymerase chain reaction (PCR) is positive, treatment can be stopped. If the 24-week PCR is negative, patients with fewer than 4 favorable factors should be treated for an additional 24 weeks. Conclusion: The combination of IFN-R is better as first line treatment than IFN monotherapy. For patients who are PCR negative after 24 weeks of treatment, genotyping and baseline viral load, fibrosis stage, gender, and age are useful predictive factors in determining whether to continue an additional 24 weeks of treatment. (HEPATOLOGY 2000;31: ) Abbreviations: IFN-R, interferon alfa-2b plus ribavirin; IFN, interferon monotherapy; HCV, hepatitis C virus; ALT, alanine transaminase; PCR, polymerase chain reaction; ROC, receiver operating characteristics. From the 1 Service d Hépato-Gastroentérologie Groupe Hospitalier Pitié-Salpêtrière, Université Paris VI, Paris, France; 2 Scripps Clinic and Research Foundation, Division of Gastroenterology/Hepatology, La Jolla, CA; 3 Department of Hepatic and Gastrointestinal Pathology, Armed Forces Institute of Pathology, Washington, DC; and 4 ScheringPlough Research Institute, Kenilworth, NJ. Received August 2, 1999; accepted October 14, T.P. received a grant from Schering Plough Research Institute, Association Pour la Recherche Contre le Cancer, and from Direction Recherche Clinique Assistance Publique Hôpitaux de Paris. Address reprint requests to: Thierry Poynard, Service d Hépato-Gastroentérologie Groupe Hospitalier Pitié-Salpêtrière, Boulevard de l Hôpital Paris Cedex 13, France. tpoynard@teaser.fr; fax: (33) Copyright 2000 by the American Association for the Study of Liver Diseases /00/ $3.00/0 Two recent multicenter randomized trials 1,2 have shown that the combination of interferon alfa-2b plus ribavirin (IFN-R) is effective in the treatment of previously untreated (naive) patients with chronic hepatitis C. In both studies, IFN-R was more effective than interferon alfa-2b (IFN) alone. Efficacy in decreasing order was: IFN-R for 48 weeks (IFN 3 millions unit [MU] 3 times a week and R 1,000-1,200 mg/d), IFN-R for 24 weeks, IFN (3 MU 3 times a week) for 48 weeks, and IFN for 24 weeks. The initial conclusions from these studies were that the majority of patients will benefit from combination therapy and that patients with unfavorable risk factors benefit the most from 48 weeks of treatment. However, several pragmatic questions could not be resolved by looking at these studies individually; some were addressed in a recent international consensus conference. 3 The aim of this study was to combine the data to increase the power of the analysis to respond more precisely to the following questions: (1) What factors are associated with a favorable viral response? In both studies, 4 favorable factors were associated with a sustained virological response: 2 viral characteristics (genotypes 2 and 3 versus genotypes 1, 4, and 5; low vs. high viral load), sex (female vs. male gender), and liver fibrosis stage (no fibrosis or limited to the portal tract vs. extensive fibrosis). In one study, age was associated with response (younger vs. older age). Factors associated with relapse and to the histological response according to viral responses were also evaluated. (2) When is it useful to test hepatitis C virus (HCV) RNA, qualitatively or quantitatively, to make decisions to continue or stop treatment: 4, 12, or 24 weeks? We applied an after starting approach, which was to evaluate the efficacy of the 48- versus 24-week regimen, not only according to the patient s baseline characteristics, but also according to the evolution of response markers during these first 24 weeks of combination treatment (viral load measured in both trials at 4, 12, and 24 weeks after starting the treatment). (3) Is there a subgroup of patients with several favorable response factors who still could be treated by interferon monotherapy? The results of the two separate trials strongly suggest that first line treatment should be combination interferon alfa-2b plus ribavirin whatever the patient characteristics. However, using the power of the combined data for the two trials and a new identification of response factors, it had to be verified that there is no subgroup of patients with several favorable response factors who still could be treated by interferon monotherapy. 211

2 212 POYNARD ET AL. HEPATOLOGY January 2000 (4) Do the therapeutic recommendations of the recent international consensus need to be revisited after using this database? In fact we have found that two recommendations were not correct: (1) to treat patients with genotype 2 or 3 for only 6 months, regardless of the other factors, and (2) to treat patients with genotype 1 for only 6 months if the level of viremia is low. Finally, according to these analyses, we suggest a decision algorithm for a treatment taking into account the response factors. PATIENTS AND METHODS The individual data from two randomized trials described in detail elsewhere 1,2 have been gathered after the acceptance of the principal investigators. All patients were prospectively screened and included using the same protocol. Each study was approved by the institutional review board at each of the centers, and all patients provided written informed consent. A database combining data from both studies was created containing the following information: gender; age at first biopsy; age at infection; body weight; presumed mode of infection (parenteral [usually intravenous drug use], transfusion, other, or unknown); type of treatment (IFN-R, IFN); duration of treatment (24 or 48 weeks); METAVIR fibrosis stage and activity grade at first and second biopsy; time elapsed between the two biopsies in months; serum alanine transaminase (ALT) activity (expressed as times upper normal limit ) before treatment, at 4 weeks, at 12 weeks, and at the end of treatment; ALT response at the end of follow-up (24 weeks after the end of treatment); quantification of virus before treatment, at 4 weeks, at 12 weeks, at the end of treatment, and at the end of follow-up (24 weeks after the end of treatment); and genotype (1, 2, 3, 4, 5, and 6). Liver biopsy specimens were fixed, paraffin-embedded, and stained. For each liver biopsy specimen, a stage of fibrosis and a grade of activity was established according to the following criteria. Fibrosis was staged on a scale of 0 to 4: F0 no fibrosis, F1 portal fibrosis without septa, F2 few septa, F3 numerous septa without cirrhosis, F4 cirrhosis. Reproducibility of results between pathologists using this method has been established. 4 The grading of activity and the intensity of the necroinflammation were scored as follows: A0 no histologic activity, A1 mild activity, A2 moderate activity, A3 severe activity 5 and by the Knodell histologic activity index. 6 The same pathologist (Z.G.) reviewed the biopsy specimens without any information concerning the timing of the biopsy or the clinical, biological, or treatment characteristics. Biopsy specimens of insufficient size to reliably assess the extent of fibrosis were excluded. Serum HCV RNA levels were performed in a single central laboratory (NGI, Los Angeles, CA) using a quantitative reverse transcription multicycle polymerase chain reaction (PCR) with a lower limit of detection of 100 copies/ml 7 and HCV genotyping using INNO-LiPA HCV (Innogenetics, Zwijnaarde, Belgium) secondgeneration assay. 8 Statistical Analysis. The definition of the primary endpoint was the sustained virological response defined as no detectable quantity in the serum, i.e., less than 100 copies/ml, 24 weeks after the end of the treatment. Baseline characteristics before starting treatment have been already identified, but the predictive values of quantitative factors (age, viral load, ALT) have not been systematically analyzed (by receiver operating characteristics [ROC] curves) to find the more accurate thresholds. In patients who respond at 24 weeks of treatment (PCR negative), the main goal of continuing treatment for an additional 24 weeks is to reduce the relapse. We first compared the characteristics of relapsers to those of nonresponders among patients treated with 48 weeks of the combination. We then compared the relapsers treated for 48 weeks with relapsers treated for 24 weeks to find predictive factors. In the 24-week nonresponder group, the goal of treating patients for an additional 24 weeks is to recruit late responders (patients who first become negative after 24 weeks) or to reduce progression of the liver histological lesions in comparison with no further treatment. For the identification of favorable and unfavorable response factors among the baseline characteristics, we compared sustained responders and nonresponders by Fisher s exact test, Student s t test, Mantel Haenszel test, or Mann Whitney nonparametric test. For the quantitative factors, we looked for the best threshold by ROC curves. For response factors identified in the univariate comparisons, a logistic regression analysis was performed to select the independent prognostic factors. RESULTS For these studies, the database included 1,744 treatmentnaive patients. At the end of treatment, the percentage of patients with undetectable HCV RNA was significantly higher in the combination groups, 51% (260 of 505) in the IFN-R 48-week group, 55% (278 of 505) in the IFN-R 24-week group, 29% (147 of 503) in the IFN 48-week group, and 29% in the IFN 24-week group (66 of 231). At the end of the follow-up, the percentage of patients with sustained undetectable HCV RNA was also higher in the combination groups 41% (205 of 505), 33% (166 of 505), 16% (82 of 503), and 6% (13 of 231), respectively, with significant differences between all these groups. Baseline Characteristics Associated With Virological Response. The baseline characteristics identified as being associated with sustained virological response in each treatment group are given in Table 1. Logistic regression analysis (Table 2) showed that favorable factors significantly and independently associated with response were treatment by IFN-R, duration of treatment, genotype 2 or 3, baseline viral load less than 3.5 million copies/ml, no or only portal fibrosis, female gender, and age younger than 40 years. Weight was highly associated with gender and the association to sustained response with weight lower than 75 kg, observed in univariate analysis, disappeared when female gender was taken into account. Relative Benefit of 24- Versus 48-Week Treatment Regimens. From the efficacy results (Table 1) it was obvious that IFN alone for 24 weeks was useless, whatever the combination of response factors. According to these response factors, the comparative benefit of regimens among risk groups showed also that there was a significant advantage for the combination of IFN-R in comparison with IFN alone for 48 weeks, whatever the combination of response factors (data not shown). Accuracy of Quantitative Characteristics to Predict Sustained Response to Combination Therapy. Based on the relative benefit assessments the regimen of IFN alone for 48 weeks was useless whatever the combination of response factors. Therefore, subsequent analyses have been performed only in patients receiving the combination regimen. From the baseline characteristics curves, the following simple thresholds seem reasonable: 40 years for age (not shown) and 3.5 million copies for viral load (Fig. 1). From the repeated measures, it seems that transaminase determinations (not shown), either absolute or relative to baseline levels, have no advantage over viral load determination with lower area under the curves and no peak with 100% negative or positive predictive values. With regard to the viral load, the highest area under the curve was obtained at the 4-week determination in patients treated for 24 weeks (Fig. 1). Patients with more than 400,000 copies at 4 weeks have no chance (negative predictive value 100%) to achieve a sustained response. For the 48-week combination regimen, the negative predictive value of more than

3 HEPATOLOGY Vol. 31, No. 1, 2000 POYNARD ET AL. 213 TABLE 1. Sustained Virological Response to Different Regimens According to Baseline Characteristics Baseline Characteristic IFN-R 48 wk % (N/total) IFN-R 24 wk IFN-Placebo 48 wk IFN-Placebo 24 wk Genotype 2 or 3 65% (102/157) 67% (105/157) 32% (51/161) 15% (9/62) 1, 4, 5, or 6 30% (103/348) 18% (61/348) 9% (31/342) 2% (4/169) Mean HCV RNA copies/ml 44% (112/257) 40% (101/250) 26% (59/228) 10% (9/95) copies/ml 38% (93/248) 26% (65/255) 8% (23/265) 2% (4/236) Age 40 y 48% (115/242) 40% (95/236) 23% (52/227) 7% (5/72) 40 y 34% (90/263) 26% (71/279) 11% (30/276) 5% (8/159) Fibrosis stage No or portal fibrosis 43% (158/368) 36% (129/362) 18% (63/351) 5% (7/154) Septal fibrosis or more 36% (36/101) 23% (27/118) 12% (14/119) 5% (3/65) Gender Female 46% (80/175) 39% (66/169) 21% (38/179) 9% (7/77) Male 38% (125/330) 30% (100/336) 14% (44/324) 4% (6/154) Body weight 75 kg 41% (96/240) 37% (87/233) 22% (49/225) 1% (1/71) 75 kg 40% (109/265) 29% (79/272) 12% (33/278) 8% (12/160) Combination of virological factors Genotype 2, % (57/88) 71% (57/80) 40% (33/83) 26% (6/23) Genotype 2, % (45/69) 62% (48/67) 23% (18/78) 8% (3/39) Genotype 1, 4, 5, % (55/169) 26% (44/170) 18% (26/145) 4% (3/72) Genotype 1, 4, 5, % (48/179) 10% (17/178) 3% (5/197) 1% (1/97) Combination of non virological factors Women, 40 y, no or portal fibrosis 57% (43/75) 56% (36/64) 36% (21/59) 9% (2/22) Men, 40 y, septal fibrosis or more 34% (15/44) 25% (13/53) 10% (6/58) 5% (2/38) Extreme favorable population Women, 40 y, no or portal fibrosis, Genotype 2, % (15/19) 69% (11/16) 58% (11/19) 0% (0/2) Extreme unfavorable population Men, 40 y, septal fibrosis or more, Genotype 1, 4, 5, % (2/23) 8% (2/26) 0% (0/29) 0% (0/18) 400,000 copies at week 4 and 12 for the 48-week regimen were 96.7% and 100%, respectively. A positive predictive value close to 100% was not achieved for viral load for either the 24- or 48-week combination regimens. The relative variation according to baseline value does not add significant predictive value (Fig. 1C and D). Distribution of Patient Response After Treatment by Combination Regimen According to the Dynamic of Their Virological Response During Treatment. All scheduled during and after treatment HCV-RNA PCR determinations were available for 739 patients, 425 in the 24-week combination group and 314 in the 48-week combination group. The comparison of HCV dynamics between the 24-week regimen and the 48-week regimen shows that there were no differences in the percentage of totally nonresponders defined as all PCR determinations being positive, 37% versus TABLE 2. Independent Factors Associated With Sustained Response (logistic regression analysis) Response Factor Regression Coefficient Standard Error Significance Odds Ratio (95% CI) Ribavirin combination ( ) 48-Week duration ( ) Genotype 2 or ( ) Baseline viral load lower than 3.5 mol/l ( ) No or portal fibrosis ( ) Female gender ( ) Age 40 y ( ) 35%, respectively, and in the percentage of escapers (patients with at least one PCR negative) per 24-week period (3% respectively; details not shown). The main difference was in a larger percentage of relapsers, 25% for the 24-week regimen versus 9% in the 48-week regimen. When the sustained virological response was analyzed by time to first negative PCR, there was in the 24-week regimen group, 4 weeks 85% (80 of 94) (94 patients had their first negative PCR at 4 weeks and among them 80 will be sustained responders), 12 weeks 47% (64 of 136), 24 weeks 21% (7 of 33) and in the 48-week regimen, 94% (65 of 69), 80% (74 of 92), and 50% (17 of 34), respectively. In the 48-week regimen there were 5 patients who became PCR negative for the first time at 36 weeks of therapy and 1 at 48 weeks. None of them were sustained responders. Prediction of Relapse Among Patients With Undetectable HCV- RNA After 24 Weeks of Treatment. Baseline characteristics, ALT at 4 and 12 weeks, and HCV-RNA response among relapsers and sustained responders are shown in Table 3. By regression analysis, the only independent predictors of sustained response in patients treated for 48 weeks were PCR negative at 4 and 12 weeks. In patients treated for 24 weeks, genotype 2 or 3, low viral load, and negative PCR 4 weeks and 12 weeks were predictive. Algorithms for the Choice Between 48 or 24 Weeks of Combination Regimen. Figure 2 depicts the distribution of patients according to HCV-RNA response after 24 weeks of treatment (A) or 12 weeks (B), according to the combination regimen (24 vs. 48 weeks) and to the sustained response at the end of

4 214 POYNARD ET AL. HEPATOLOGY January 2000 FIG. 1. ROC curves of baseline and 4, 12, and 24 weeks of serum HCV RNA load (in Log) for sustained response prediction in patients treated by IFN-R combination for (A) 24 weeks and (B) 48 weeks. For the 24-week regimen, the best likelihood ratio was 9.25 obtained for a 4-week HCV RNA of 2.3 Log or less ( 200 copies/ml) (sensitivity [Se] 0.66; specificity [Sp] 0.93; positive predictive value [PPV] 0.83; negative predictive value [NPV] 0.84). A sensitive peak (1.58) was obtained for 5.6 Log at 4 weeks ( 400,000 copies/ml) (Se 1.00; Sp 0.37; PPV 0.45; NPV 1.00). This 100% NPV is useful because a patient above this threshold has no chance to be a sustained responder. For the 48-week regimen, the best likelihood ratio was 5.64 obtained for a 4-week HCV RNA of 2.3 Log or less ( 200 copies/ml) (Se 0.49; Sp 0.91; PPV 0.80; NPV 0.72). ROC curves of 4-, 12-, and 24-week serum HCV RNA load difference versus baseline (in Log) for sustained response prediction in patients treated by IFN-R combination for (C) 24 weeks and (D) 48 weeks. For the 24-week regimen, the best likelihood ratio was obtained for a 4-week HCV RNA of 4 Log difference or more (Se 0.41; Sp 0.97; PPV 0.86; NPV 0.76). For the 48-week regimen, the best likelihood ratio was 6.57 obtained for a 4 Log difference between baseline and 4-week viral load (Se 0.28; Sp 0.96; PPV 0.82; NPV 0.65). TABLE 3. Baseline Characteristics and First 24 Weeks ALT and Viral Response in Relapsers and Sustained Responders Among Patients With Undetectable HCV RNA After 24 Weeks of Treatment IFN-R 48-Week Regimen PCR Negative 24 Weeks IFN-R 24-Week Regimen PCR Negative 24 Weeks Relapsers N 70 (%) Sustained Responders N 268 (%) Statistical Significance (Univariate Multivariate) Relapsers N 107 (%) Sustained Responders N 151 (%) Statistical Significance (Univariate Multivariate) Baseline characteristic Genotype 2 or 3 18 (26%) 97 (49%) P (21%) 97 (64%) P.0001 P.08 P.0001 Mean HCV RNA copies/ml 32 (46%) 107 (54%) P (39%) 90 (60%) P.001 P.72 P.005 No or portal fibrosis 49 (75%) 154 (82%) P (76%) 118 (83%) P.17 P.39 P.06 Female gender 25 (36%) 77 (39%) P (32%) 60 (40%) P.19 P.94 P.34 Age 40 years 33 (47%) 111 (56%) P (47%) 88 (58%) P.07 P.63 P.36 Body weight 75 kg 34 (49%) 91 (46%) P (52%) 79 (52%) P.99 P.20 P.18 First week markers Normal ALT 4 weeks 59 (88%) 188 (95%) P (87%) 141 (94%) P.83 P.17 P.63 Normal ALT 12 weeks 65 (91%) 191 (97%) P (96%) 144 (95%) P.94 P.91 P.53 Negative PCR 4 weeks 10 (14%) 91 (46%) P (10%) 83 (55%) P.0001 P.01 P.0001 Negative PCR 12 weeks 45 (64%) 175 (88%) P (67%) 141 (93%) P.0001 P.005 P.03

5 HEPATOLOGY Vol. 31, No. 1, 2000 POYNARD ET AL. 215 regimen, 9% of sustained responders had a fibrosis stage worsening versus 24% in nonresponders (P.001). Among the 24-week regimen, 6% of sustained responders had a fibrosis worsening versus 20% of nonresponders (P.001). For relapsers, the response was intermediate between the sustained responders and the nonresponders. Figure 4 depicts the distribution of patients according to two recent international consensus recommendations. The recommendation to treat patients with genotype 1 for only 24 weeks if the level of viremia is low, is inadequate, because we observed in this population 53% of sustained response versus 71% when treated for 48 weeks. The greater efficacy of 48 weeks in genotype 1 persisted after stratification by viral load (Mantel Haenszel odds ratio 3.3; 95% CI ; P.001) in favor of 48 weeks. The recommendation to treat patients with genotype 2 or 3 for only 24 weeks, regardless of the other factors, could also be inadequate, because we observed among patients with extensive fibrosis 65% of sustained response versus 80% when treated for 48 weeks (odds ratio 2.1; 95% CI ; P.30). Figure 5 depicts a proposed algorithm of treatment with the observed repartition of patients. FIG. 2. Algorithm showing the distribution of patients according to the HCV-RNA response after (A) 24 weeks or (B) 12 weeks of treatment to the combination regimen and to sustained response at the end of follow-up. follow-up. In patients who had detectable HCV-RNA after 24 weeks of treatment, there was no advantage in continuing the treatment for an additional 24 weeks in terms of virological response. The probability of achieving a sustained virological response (undetectable HCV RNA 24 weeks after the end of the 48-week treatment) was only 2% versus 3% in patients who do not continue treatment. In patients who had detectable HCV RNA after 12 weeks of treatment (B), there was still 4% of patients among the 24-week regimen and 10% among the 48-week regimen who became sustained responders. Figure 3 depicts the distribution of patients according to the virological and histological responses. There was a very significant improvement of histological activity scores and less fibrosis stage worsening among sustained responders versus nonresponders. Among the 48-week combination DISCUSSION This study has allowed us to improve the monitoring of patients and the choice of treatment regimen according to the response factors. In comparison with the separate publications of the two trials, a better à la carte regimen can be proposed, with a better view of factors associated with sustained response, relapse, and histological improvement. We have excluded that the kinetics of viral load at 4 and 12 weeks permit taking very early therapeutic decision. We have also excluded that in some subgroups of patients with many favorable factors a treatment with IFN only could be sufficient. The independent prognostic values of 5 baseline characteristics have been confirmed. HCV genotypes 2 and 3 are associated with better response to the combination than other genotypes. However, this conclusion is validated only versus genotypes 1a and 1b, because too few patients with genotypes 4, 5, or 6 were included in these trials. For viral load, the ROC curves showed that there is no threshold that had either a positive or negative predictive value. Therefore, the simplest way to classify viral load into high or low is to take the median, which was 3.5 million copies. For age, which was the only continuous quantitative variable, the threshold of 40 years seems to have the best accuracy. The nonidentification of age as a response factor in the United States trial was probably because of a lower power. Because the multivariate analysis showed that these 5 factors can only explain 20% of the variability of the sustained response, we need to identify other independent factors. Very few studies so far have looked to factors associated with relapse among responders at the end of the treatment. When we compare factors associated with relapse between the two treatment durations, they were different (Table 3). For the 48-week duration, genotype and PCR at 4 weeks and 12 weeks were the only predictive factors. This suggests that in patients with genotype 1, and/or with late negative PCR, the treatment should be continued for more than 48 weeks (18 or 24 months?) to achieve a sustained response. However, this should be proved by randomized comparisons. The results show that there is no place for interferon

6 216 POYNARD ET AL. HEPATOLOGY January 2000 FIG. 3. Virological and histological responses. Algorithm showing the distribution of patients according to the HCV-RNA response after 24 weeks of treatment to the sustained response at the end of follow-up to the combination regimen and to the histological responses. monotherapy at a dose of 3 million units 3 times a week for either 24 or 48 weeks, even in the most favorable patient. Among patients with genotype 2 or 3 and low viral load, the sustained response rate was much greater with the 24-week combination regimen (71%) than with 48 weeks of IFN monotherapy (40%; P.001). IFN monotherapy should be recommended only if combination IFN-R therapy is contraindicated. The choice of 24 or 48 weeks for combination therapy has been clarified. The crucial time to make this decision is at 24 weeks based on the results of HCV-PCR testing. In patients who are PCR negative at 24 weeks (59% of the patients in these studies), the goal is to reduce the relapse rate. There was an overall highly significant improvement with 48 weeks of treatment (74% sustained responders) versus 24 weeks (59% sustained responders). Because patients with many favorable response factors benefit less from 48 weeks of treatment, consideration can be given to stopping at 24 weeks in these FIG. 4. Algorithm showing pitfalls of recent international consensus recommendations. Recommendations are bold in the algorithm. The recommendation to treat only 24-week patients with genotype 1 if the level of viremia is low is inadequate because we observed in this population 53% of sustained response versus 71% when treated for 48 weeks. The greater efficacy of 48 weeks in genotype 1 persisted after stratification by viral load (Mantel Haenszel odds ratio 3.3; 95% CI ; P.001) in favor of 48 weeks. The recommendation to treat 24-week only patients with genotype 2 or 3, regardless of the other factors, could also be inadequate because we observed among patients with extensive fibrosis 65% of sustained response versus 80% when treated for 48 weeks (odds ratio 2.1; 95% CI ; P.30).

7 HEPATOLOGY Vol. 31, No. 1, 2000 POYNARD ET AL. 217 FIG. 5. Proposed algorithm of treatment. The distribution of patients observed in the database is given according to the 24-week PCR, the number of baseline favorable factors (4 or 5 vs. less than 4), and the combination regimen duration. The suggested recommendation is bold. For 27 patients, the baseline fibrosis stage was unknown. patients. A simple strategy could be to consider only the HCV genotype and stop treatment at week 24 in genotype 2 and 3 responders, because the sustained response was 82% in patients treated for 24 weeks versus 84% in patients treated for 48 weeks. However, from our results it seems hazardous to recommend a strategy based only on virological characteristics. There are in fact 5 independent response factors, and to take into account only 1 factor among these 5 is an over-simplification that could lead to errors in different populations or subgroups. 3,9 For example, we have identified that patients with genotype 2 or 3 who are PCR negative at 24 weeks and who have extensive fibrosis will have a better sustained response with 48 weeks of treatment, 80%, compared with 65% in patients whose treatment is stopped at 24 weeks (Fig. 4). The sample size is small and the 2.1 odds ratio not significant. However, is a significant and independent factor and the population included in these two trials had a low prevalence of patients with extensive fibrosis and of older patients. Therefore, for a population of older men with extensive fibrosis, the choice of 48 weeks duration in responders should not be based only on genotype and viral load. The recommendation of the international consensus conference to treat patients with genotype 2 or 3 for only 24 weeks, regardless of the other factors, seems inappropriate. 3 Furthermore, from a clinical point of view there is a risk to over-simplify a decision according to a specific threshold, i.e., choice of 48 weeks of treatment if the viral load is 3.75 million and 24 weeks if 3.25 million or 39 versus 41 years of age. This argues for taking the decision on both the number of independent factors and the tolerance to the combination. Similarly, the recommendation of the international consensus conference 3 to treat patients with genotype 1 for only 6 months if the level of viremia is low, is not correct according to our results. This recommendation would lead to a reduction of 18% of the sustained response rate obtained by the 48-week regimen (Fig. 4). Finally, our recommendation is to treat naive patients with interferon ribavirin combination for 24 weeks and to test the HCV PCR at this point. If HCV RNA is undetectable (59% of the included population), the decision to continue the combination for 24 weeks or not should be taken according to the number of favorable factors (Fig. 5). Because 5 factors represent 32 different combinations, it will be impossible to test by randomized prospective trials all these scenarios. It seems reasonable to stop treatment in case of the presence of almost all the favorable factors, i.e., 4 or 5 factors. This represents 21% of the 24-week negative PCR that is 13% of the population. For patients with less than 4 factors, who represented 46% of the population, its seems useful to continue the treatment for a total of 48 weeks. For patients who remain PCR positive at 24 weeks (which represented 41% of the population), the choice of whether to treat for 24 or 48 weeks has not been fully resolved. From the perspective of HCV eradication, the combination can be stopped at 24 weeks because the probability of obtaining a sustained virological response is 2% (Fig. 2). The remaining question concerns the usefulness of continuing the treatment to reduce histological damage, because interferon and ribavirin have not only antiviral, but also antifibrotic and immunomodulatory effects In PCR-positive patients at 24 weeks, we observed that 8% more patients had an improvement among patients treated for 48 weeks (42%) versus 24 weeks (34%) with no difference for fibrosis. Studies are needed to assess whether patients who fail to respond to combination therapy will benefit from either long-term interferon monotherapy or combination therapy. A remaining question is whether treatment can be stopped at 12 weeks in some subgroups because of a high probability of nonresponse. There was no consensus at the recent international conference. 3 From our data this approach cannot be recommended because in the 48-week regimen, among the patients who had a positive PCR at 12 weeks, we observed a sustained response in 10% of patients. Even the 24-week regimen induced a sustained response in 4% of these patients (Fig. 2). Our analysis has also permitted a clarification of the usefulness of genotype and viral load determinations. For patients who remain PCR positive after 24 weeks of IFN-R, there is no need to determine the genotype or perform any quantitative measurement before or during treatment. The combination can be stopped whatever the response factors. In contrast, for patients who are PCR negative at 24 weeks, clinicians need to know the genotype and the baseline viral load to decide whether to continue treatment for an additional 24 weeks. ALT, PCR, or viral load measurements at 4 or 12 weeks are not useful. Therefore, in the population of patients who are PCR negative at 24 weeks, genotype and baseline viral load could be performed retrospectively on stored samples. However, for the very high risk group (men, older than 40 years and with at least septal fibrosis, F2) therapy should be continued whatever their viral characteristics. The predictive values of initial decline in viral load were finally disappointing for the 48-week regimen. For the 24-week combination regimen we found as Zeuzem et al. 16 found in patients treated by interferon alone, an absence of decline to less than 400,000 copies at 4 weeks was highly predictive of an absence of sustained response. Indeed for the 24-week regimen, when the ROC curves are compared (Fig. 1A and D), the 4-week viral load had a higher area under the curve than 12- or 24-week assessments. In contrast this

8 218 POYNARD ET AL. HEPATOLOGY January 2000 advantage disappeared for the 48-week regimen because of the late induction of sustained response among patients with more than 400,000 copies at 4 weeks (Fig. 1B and C). In conclusion, combination IFN-R is the new first line treatment for chronic hepatitis C. Based on the data from the currently available studies and the relative risks and response factor combinations, it seems reasonable to recommend 48 weeks of treatment only for patients who are PCR negative at 24 weeks and who do not have 4 or 5 favorable response factors. REFERENCES 1. Poynard T, Marcellin P, Lee SS, Niederau C, Minuk GS, Ideo G, Bain V, et al. and the International Hepatitis Interventional Therapy Group (IHIT). Randomised trial of interferon alfa-2b plus ribavirin for 48 weeks or for 24 weeks versus interferon alfa-2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. Lancet 1998;352: McHutchison J, Gordon S, Schiff E, Shiffman M, Lee W, Rustgi V, Goodman Z, et al. for the Hepatitis Interventional Therapy Group (HIT). Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. N Engl J Med 1998;339: Consensus Statement. EASL International Consensus Conference on Hepatitis C. J Hepatol 1999;30: Bedossa P. The French METAVIR cooperative group. Inter- and intraobserver variation in the assessment of liver biopsy of chronic hepatitis C. HEPATOLOGY 1994;20: Bedossa P, Poynard T, for the French METAVIR group. An algorithm for the grading of activity in chronic hepatitis C. HEPATOLOGY 1996;24: Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N, Kiernan TW, et al. Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis. HEPATOLOGY 1981;1: Tong MJ, Hwang SJ, Lefkowitz, Lee SD, Co RL, Conrad A, Schmid P, et al. Correlation of serum HCV RNA and alanine aminotransferase levels in chronic hepatitis C patients during treatment with ribavirin. J Gastroenterol Hepatol 1994;9: Stuyver L, Rossau R, Wyseur A, Duhamel M, Vanderborght B, Van Heuverswyn H, Maertens G. Typing of hepatitis C virus isolates and characterisation of new subtypes using a line probe assay. J Gen Virol 1993;74: Reichard O, Norkrans G, Fyden A, Braconier JH, Sönnerborg A, Weiland O, for the Swedish Study Group. Randomised, double-blind, placebocontrolled trial of interferon alfa-2b with and without ribavirin for chronic hepatitis C. Lancet 1998;351: Sobesky R, Mathurin P, Charlotte F, Moussalli J, Olivi M, Vidaud M, Ratziu V, et al. for the MULTIVIRC group. Modeling the impact of interferon alfa treatment on liver fibrosis progression in chronic hepatitis C: a dynamic view. Gastroenterology 1999;116: Manabe N, Chevalier M, Chossegros P, Causse X, Guerret S, Trepo C, Grimaud J-A. Interferon- 2b therapy reduces liver fibrosis in chronic non-a-nonb hepatitis: a quantitative histological evaluation. HEPATOL- OGY 1993;18: Castilla A, Prieto J, Fausto N. Transforming growth factor 1 and in chronic liver disease: effect of interferon alpha therapy. N Engl J Med 1991;324: Shiffman M, Hoffman C, Thompson E, Feirreira-Gonzales A, Contos M, Koshy A, Luketic V, et al. Relationship between biochemical, virological, and histological response during interferon treatment of chronic hepatitis C. HEPATOLOGY 1997;26: Poynard T, Bedossa P, Chevalier M, Mathurin P, Lemonier C, Trepo C, Couzigou P, et al. and the multicentre study group. A comparison of three interferon alpha-2b regimens for the long-term treatment of chronic non A, non B hepatitis. N Engl J Med 1995;332: Ning Q, Brown D, Parodo J, Cattral M, Gorczynski R, Cole E, Fung L, et al. Ribavirin inhibits viral-induced macrophage production of TNF, IL-1, the procoagulant fg12 prothrombinase and preserves Th1 cytokine production but inhibits Th2 cytokine response. J Immunol 1998;160: Zeuzem S, Lee JH, Franke A, Ruster B, Prummer O, Herrmann G, Roth WK. Quantification of the initial decline of serum hepatitis C virus RNA and response to interferon alfa. HEPATOLOGY 1998;27: