Predictors of sustained virological response to a 48-week course of pegylated interferon alfa-2a and ribavirin in patients infected with hepatitis C virus genotype 4

BACKGROUND AND OBJECTIVES: Knowledge of the predictors of sustained viral response (SVR) to pegylated interferon (PEG-INF) alfa-2a and ribavirin (RBV) therapy in patients with hepatitis C genotype-4 (HCV-4) is crucial for selecting patients who would benefit most from therapy. We assessed the predictors of SVR to this combination therapy in Saudi patients with chronic HCV-4 infection. PATIENTS AND METHODS: This retrospective study included 148 patients with HCV-4 infection who underwent clinical, biochemical and virological assessments before treatment and at 12, 24, 48 and 72 weeks post-treatment. RESULTS: Of the 148 patients, 90 (60.8%) were males. Mean (SD) for age was 48.5 (12.7) years and BMI was 27.9 (7.5) kg/m2. Seventy-nine of 148 (60.1%) patients were treatment naïve and 110 (74.3%) underwent pre-treatment liver biopsy. Eighteen (12.2%) patients did not complete therapy because of side effects or they were lost to follow up. Early virological response was achieved in 84 of 91 (92.3%) patients. In the 130 (87.8%) patients who completed therapy, 34 (26.2%) were non-responders and 96 (63.8%) achieved end-of-treatment virological response (ETVR). SVR and virological relapse (24 weeks after ETVR) occurred in 66/130 (50.7%) and 30/130 (31.2%) patients, respectively. Compared to relapsers, sustained responders were significantly younger (P=.005), non-diabetic (P=.005), had higher serum albumin (P=.028), lower alpha-fetoprotein level (P=.026), lower aspartate aminotransferase (AST) (P=.04) levels, and were treatment-naivve (P=.008). In a multivariate regression analysis, the independent predictors of SVR were younger age (P=.016), lower serum AST (P=.012), and being treatment naivve (P=.021). CONCLUSION: Approximately half of HCV-4 patients who complete the course of combination therapy achieve an SVR, especially if they are young, treatment naivve and have lower AST levels.

(RBV) therapy given for 48 weeks is now established as the standard therapy for patients with chronic HCV infection with genotypes 1 and 4. 3 This treatment has yielded overall sustained virological response (SVR) rates of 54% to 69% in randomized controlled phase III clinical trials. 4p6 However, response to treatment is not uniform across all populations 7 and is dependent on various viral and host factors. Most of the studies conducted worldwide have included patients infected with HCV genotypes 1, 2 and 3. 4p6,8p10 According to these studies, factors independently associated with higher SVR to combination therapy include serum HCVpRNA levels below 2 million copies/mL, body weight <75 kg, age younger than 40 years, an absence of preptreatment bridging fibrosis or cirrhosis, being treatment naïve, inp p fection with HCV genotype 2 or 3, and favorable initial virological response. 4p6,8p10 HCVp4 is known to be endemic in Central Africa and in the Middle East. 11,12 However, several recent studies carried out in Europe have indicated changes in genotype distribution and have underlined the increasing prevap p lence of HCVp4. 13p15 The prevalence of HCV antibody positivity in Saudi Arabia ranges from 1% to 3%, 16,17 with genotype 4 representing 60% to 70% of these infecp p tions. 18p21 There are limited reports on the treatment of chronic HCVp4 patients from the Middle East (mainly from Saudi Arabia, Egypt, Kuwait, and Qatar), 22p31 or elsewhere. 32p35 All these studies were heterogeneous and were weakened by small numbers of patients, the use of conventional interferon with or without RBV, different durations of therapy, the inclusion of patients copinfected with humanpimmunodeficiency virus (HIV), the lack of liver histopathology data and by the absence of data assessing the predictors of SVR. A summary of these studies is shown in Table 1. Also, the only available metap analysis that has assessed PEGpINF therapy in HCVp4 patients included only 6 studies, 4 of which were in abp p stract form and the other two included only 65 patients. 36 Therefore, the primary objectives of this retrospective study were to evaluate the overall efficacy and safety of 48 weeks course of PEGpINF alfap2a and RBV combip p nation therapy in 148 consecutive Saudi patients with chronic HCVp4 infection and to assess the independent predictors of SVR in these patients.  37 Patients were then treated with PEGpINF alfap2a (40 KD; Pegasys, F. HoffmannpLa Roche, Basel, Switzerland, 180 microgram weekly) plus RBV (Copegus, F. HoffmannpLa Roche, Basel, Switzerland, 1000p1200 mg daily) for 48 weeks. Clinical, biochemical and viral pap p rameters were collected both preptreatment and at weeks 12, 24, 48 and 72 of followpup.

PATIENTS AND METHODS
Serum HCV RNA was extracted using an automatp p ed extraction system. HCV detection and quantification were performed using an Abbott RealpTime M2000rt PCR assay, which utilized two sets of primers and probes and targeted a conserved region of the 5' untranslated region of the genome and an internal control. This asp p say detects and quantifies HCV genotypes (1p6) with a detection limit that ranges from 30 to 100 000 000 IU/ mL, where 1 IU/mL=4 copies/mL. Prior to treatment, HCV genotype was performed in all patients (n=148; 100%) using INNOpLiPA HCV II (Innogenetics NV, Ghent, Belgium). 38 Realptime PCR has been available in our institution since January 2006, but the lower detecp p tion limit and the unit used have changed in the last 1 to 2 years.
Before 2006, viral load testing was performed usp p ing the Bayer Quantiplex bDNA System (Bayer Corp, Tarrytown, NY, USA). The lower quantification detecp p tion limit was 3200 copies/mL. The highest detection limit was 40 000 000 copies/mL. The average period bep p tween the quantitative PCR test and the start of therapy was 2.7 months. The National Institute of Health guidep p lines state a drop of ≥2 log10 in serum HCV viral load is indicative of response. An early viral response (EVR) was defined as ≥2 log10 drop in serum HCV viral load at 12 weeks after start of treatment. An endpofptreatment vip p rological response (ETVR) was defined as an undetectp p able serum HCV RNA at 48 weeks. A sustained viral response (SVR) was defined as a persistently undetectp p able HCV RNA at 72 weeks (6 months after the end of course of treatment). Nonresponse (NR) was defined as a persistent positive HCV (PCR) after 48 weeks of treatment.
Data were collected initially in a specialized data colp p lection form, then introduced into a Microsoft Excel worksheet and finally transferred to the Statistical Package for Social Sciences (SPSS) version 15.0 for Windows (SPSS 15.0, SPSS Inc., Chicago, IL, USA) for analysis. Means of continuous variables were comp p pared using t tests or nonpparametric tests (Wilcoxon and Mann Whitney), as appropriate. The chipsquare or Fisher exact tests were used to compare frequencies and proportions. Multivariate stepwise logistic regression analysis was performed to determine the independent predictors of sustained response. An intentionptoptreat analysis was used. Patients who discontinued treatment and those who did not complete their course of treatp p ment either due to adverse effects or loss to followpup were not included in the analysis for ETVR or SVR.  (Figure 1). PCR at 12 weeks postptreatment was done in 131 (91.9%) patients from the whole cohort (n=148) and in 91 (94.8%) of those who completed the full course of treatment (n=96). EVR was achieved in 96 patients (64.9%) in the whole cohort (n=148) and in 84 (92.3%) of those who achieved ETVR. EVR was significantly more common in patients who received treatment for the first time compared to those who previously received treatment (61 of 80, 76.3% versus 35 of 56, 62.5%, P=.016) ( Table 3). However, there was no statistically significant difference between the patients who achieved SVR and those who relapsed after ETVR in the rate of EVR (P=.157).

RESULTS
Eighteen patients (12.2%) failed to complete the 48p week therapy due to side effects or loss to follow up. The full course of therapy was given to 130 (87.2%) patients; 96 (72.9%) achieved ETVR and 34 (26.2%) were nonp responders. SVR was achieved in 66 of 130 (50.7%), while the remaining 30 (31.2%) developed virological relapse after ETVR ( Figure 2). Virological responses in treatmentpnaïve patients and in those who were previp p ously treated with interferonpbased therapy are shown in Figure 3. By univariate analysis, treatmentpnaïve patients had a significantly higher SVR (P=.008) and lower rep p lapse rate after ETVR (P<.008) compared to those who were previously treated with interferonpbased therapy. However, the difference in EVR rate between treatmentp naïve patients and those who were previously treated by interferonpbased therapies did not reach statistical significance (P=.083) ( Table 3). Because the study was retrospective and most patients who gave a history of previous INFpbased therapy received their initial treatp p ment in other institutions or outside Saudi Arabia, the data on whether the previously treated patients were nonresponders or relapsers was grossly inadequate and difficult to analyze.
By univariate analysis, patients with SVR were sigp p nificantly younger (P=.005), had a lower rate of diabep p tes mellitus (P=.005), higher serum albumin (P=.028), lower preptreatment serum aspartate aminotransferase (AST) levels (P=.04), lower serum alphapfetoprotein (AFP) levels (P=.026), and were more treatmentpnaïve (P=.008) than patients who developed virological rep p lapse after ETVR (Table 3, 4). Both groups were similar in preptreatment inflammation grade, fibrosis stage, viral load, alanine aminotransferase (ALT) levels, interferon dose, RBV dose, organ transplant status, presence of overlap syndrome, copinfection with HBV or HIV, and BMI. By stepwise multivariate logistic regression analysis (using the variables that were significant in the univariate analysis), only being younger (age as a continuous varip p able), having lower AST levels and being treatmentpnaïve were independent predictors of SVR (P=.016, P=.012 and P=.021 respectively) ( Table 5).
A total of 158 side effects were encountered in 66 pap p tients (44.6%) during followpup ( Table 6). Effects that occurred in ≥5% of patients included fatigue, body aches, weight loss, skin rash, anemia, leucopenia and thrombop p cytopenia. Subcutaneous injections of erythropoietin and granulocytepcolony stimulating factor (GpCSF) were used in 7 patients (19.4% of patients who developed anep p mia) and 10 patients (25.6% of patients who developed leukopenia), respectively. Dose reduction due to side efp p fects occurred in 62 patients (41.9%). Therapy had to be stopped temporarily 64 times in 43 patients (29.1%) and permanently in only 5 patients (3.4%). A total of 13 (8.8%) patients did not complete treatment due to loss to follow up.

DISCUSSION
The present study involved the largest cohort of patients infected with HCVp4 to be reported in the literature afp p ter treatment with the combination of PEGpINF alfap 2a and RBV for 48 weeks, and shows that this group of patients can no longer be considered "difficult to treat". Indeed, with the use of this regimen, SVR was achieved in 44.6% of the whole cohort, in 50.8% of those who completed treatment and in 68.8% of those who achieved ETVR. These results are similar to the responsp p es achieved in previous studies that involved cohorts with predominantly genotype 1 and are less than the responses in patients infected with genotype 2 or 3. 36,39,40   Data are expressed as n (%). NS; not significant, INF; interferon. EVR; early virological response. SVR; sustained virological response. a Means detection of ≥1autoantibody in serum. b Liver biopsy done in 72 patients of the 9 who completed the treatment, in the SVR group and 2 of those who had virological relapse after ETVR. c PCR at 12 weeks post-treatment was done in 91 patients of the 9 who completed the treatment, 1 in the SVR group and all the 30 who had virological relapse after ETVR.
Only 18 (12.2%) did not complete their course due to either side effects (n=5) or loss to follow up (n=13) and a total of 34 (26.2%) patients were classified as nonprep p sponders. This rate of nonpresponse can be accepted if we put into consideration the tertiary nature of our hospital and the inclusion of many complicated cases such as those who failed previous interferon therapy, cases with organ transplantation, and cases copinfected with HIV and/or HBV. The impression that patients infected with HCVp4 respond poorly to interferonpbased therapy and are genp p erally "difficult to treat" came from many earlier studies   where conventional interferonpalfa was used alone or in combination with RBV (Table 1). 24,28,30p34 However, the use of PEGpINF alfap2 and RBV for 48 weeks lead to a substantial improvement in the rate of SVR as evidenced by other studies who used PEGpINF alfap2b, 23,25,29 and PEGpINF alfap2a, 35 as was the case in our study. Neither the fibrosis stage nor the inflammation grade in the preptreatment liver biopsy was found to be statisp p tically different between sustained responders and those who developed virological relapse after ETVR. This is contrary to what was previously reported by other studies in patients infected with genotype 1 4p6 and genotype 4. 29 It should be noted that only 72 of the 96 patients who achieved ETVR in the present study underwent a prep treatment liver biopsy. Also, only 23 patients in our cop p hort had fibrosis stage ≥3, and only 3 patients had fibrop p sis stage 4 (cirrhosis). In addition, liver biopsies had not been performed immediately before the onset of therapy. Moreover, in the study by Hassan et al, SVR was less in patients with an advanced fibrosis score, but this was only in a univariate analysis, and no multivariate analysis was performed. 29 The study of Kamal et al, however, showed that only age of >40 years and preptreatment viral load of >2 million copies/mL, can independently predict SVR, and not the preptreatment liver pathology. 25 We believe that the effect of preptreatment fibrosis on the SVR to therapy becomes more obvious if comparisons between cohorts with predominantly stage 3p4 are compared with those with predominantly stage 1p2 are made. Contrary to other reports, 25 preptreatment HCV vip p ral load was not found to be a predictor of SVR in our study. It is wellpknown that viral load fluctuates and a single reading of HCV quantification may not reflect the actual viral load at the time of treatment, especially if we know that viral load was assessed at varying inp p tervals from the onset of treatment. It has also been rep p ported that the differences in interferon response could be secondary to either a difference in the viral virulence and/or replication rate among different HCV genotypes and not the absolute viral load. 40 The safety profile of the combination therapy of PEGpINF alfap2a and RBV used in the present study is comparable to what was previously described in the literature. 25,29,41 Indeed, only 18 (12.2%) patients did not complete their course of treatment in our study due to the development of side effects, loss to followpup and/or transfer to liver transplantation or development of dep p compensated cirrhosis or hepatocellular carcinoma.
The significantly lower SVR in our patients who prep p viously received interferon therapy (28.8%) compared to those who were treatmentpnaïve (55.1%) is consistent with the results of many studies in both genotype 1 and genotype 4. 27,29,41p43 Our results were better than those of Shiffman et al 42   of 12% to 16% in previously treated patients and 24% to 28% in treatment naïve patients. The mechanism(s) underlying this lower response is not known. However, it may be related to the development of an intrinsic or immunological resistance to the direct antipviral effect of interferon. Interestingly, interferonpinducible protein 10 kDa (IPp10), which is a chemokine produced by hepatop p cytes that targets Tplymphocytes, natural killer cells and monocytes was recently identified. 44,45 Elevated serum levels of IPp10 before initiation of therapeutic intervenp p tion for HCV infection were reported in patients not achieving SVR. 46,47 A recent study confirmed that prep treatment IPp10 levels predict SVR in patients infected with HCV genotype 1, even in those with higher BMI and viral load. 48 Thus, assessment of preptreatment IPp 10 may help in identifying patients for whom current therapy is beneficial. This needs to be tested in patients infected with HCVp4. Better identification of the preptreatment host or vip p ral factors that can identify which patients respond betp p ter to therapy is currently attracting more attention. For instance, HCVp4 subtyping has been proposed to affect the response to PEGpINF alfap2a plus RBV combinap p tion therapy. 49p50 Other predictors under investigation include increased baseline insulin resistance and AFP levels. 51,52 We have shown that diabetes mellitus and AFP levels were less in sustained responders by univarip p ate analysis. However, neither were found to be indep p pendent predictors by multivariate regression analysis. These factors need to be assessed in HCVp4 patients in wellpdesigned prospective studies.
This study demonstrates for the first time that lower baseline serum AST and not ALT is an independent predictor of SVR to PEGpINF alfap2a and RBV in pap p tients with chronic HCVp4. We believe that these lower AST levels reflect less severe histological parameters in the sustained responders. A study by Zechini et al showed a statistically significant positive correlation of baseline aminotransferase values with the hepatitis acp p tivity index and fibrosis score. 53 In support of our results, a study by Assy et al reported a significant positive corp p relation between AST values and the extent of hepatic fibrosis. 54 We compared our patients who had fibrosis stage 0p2 (n=86) and those who scored 3p4 (n=24) for all baseline parameters and found that only younger age and lower AST levels are independent predictors of fip p brosis stage 0p2.
Unlike the situation in most randomized controlled trials, the patients included in this study were heterop p geneous: 10.1% had previous organ transplantation, 8.1% were positive for liver autoantibodies (classified as overlap syndrome), 19.6% were positive for HBV or HIV serology, 39.9% were nonpresponders to previous interferonpbased therapy and 6.1% had renal impairp p ment. These factors are known to affect the natural hisp p tory of chronic HCV infection. However, this sample of patients represents what we usually face in real life. If we exclude cases with all the above copmorbidities, our results will only be applicable to patients with isop p lated HCV infection. As shown in Table 3, sustained responders and relapsers after ETVR were similar rep p garding transplantation status, BMI, autoantibody stap p tus, renal function and HBV or HIV status. The potential limitations of the current study include the fact that postptreatment biopsy was not done as our main objective was to assess preptreatment predictors of SVR. There is a solid evidence that SVR is associated with improved outcomes, 55 stabilization, and/or regresp p sion in hepatic fibrosis stage in response to treatment, especially if it associated with viral clearance. 56p58 In adp p dition, assessing the impact of therapy on liver histop p pathology was beyond the scope of this study. Due to the retrospective nature of the present study, baseline liver biopsy was performed in the majority, but not all cases. Obviously, a better assessment of the predictive role of these two parameters can be done if they were performed in all patients. However, the similarity of the SVR achieved in the present study to what has already been reported by others makes it less likely to substanp p tially affect the results. Another limitation is that pap p tients were followed up for 24 weeks after completion of therapy and thus longer term clinical outcomes could not be determined. Indeed, it has been reported that late relapse may occur after 4 years of completion of interp p feron therapy. 59 However, by definition SVR is the perp p sistence of the ETVR for 24 weeks, which was assessed in this study. In addition, early viral kinetics at week 4, which is called rapid virological response (RVR) was not assessed in this study. Indeed, a recently published work from Egypt showed that the duration of combinap p tion therapy with PEGpINF alfap2b and RBV can be shorter treatment for patients who have attained RVR or EVR. 60 However, the concept of RVR was not enterp p tained at the beginning of our study. This needs to be confirmed in another prospective trial. We have recently published a study involving 335 patients with HCV inp p fection of all genotypes. 61 It showed that the response of genotype 4 patients to combination therapy is more or less similar to genotype 1. Both showed a worse response compared to those infected with genotypes 2 or 3.
In conclusion, combination therapy with PEGpINF alfap2a and RBV, if tolerated and completed, is effective in treating chronic HCVp4 patients especially if they are younger than 40 years of age, have no previous interp p feron therapy and have lower preptreatment AST levels. Attempts to improve adherence to therapy and the early detection together with treatment of complications are needed to achieve better response to therapy. Further studies addressing other potential predictors of SVR in chronic HCVp4 patients such as the IP10, insulin resisp p tance, HCVp4 subtype heterogeneity are warranted.