iconstar paper   Hepatitis C Articles (HCV)  
Back grey arrow rt.gif
 
 
'IL28B Predicts SVR in Coinfection, Associated With 2-Fold Higher Rate of SVR' - published pdf attached
 
 
  Download the pdf here
 
The published study in CID October was presented at this past CROI:
 
Interleukin 28 B Genotype Is a Potent Predictor of Response to ...
Juan Pineda*1, A Caruz2, A Camacho3, K Neukam1, I Salas3, A Martinez3, ... Conclusions: rs12979860 polymorphism upstream of IL28B gene has a marked impact ... www.natap.org/2010/CROI/croi_145.htm
 
--------------------------
 
MAJOR ARTICLE
 
Prediction of Response to Pegylated Interferon plus Ribavirin by IL28B Gene Variation in Patients Coinfected with HIV and Hepatitis C Virus
 
Juan A. Pineda,1 Antonio Caruz,3 Antonio Rivero,4 Karin Neukam,1 Irene Salas,3 angela Camacho,4 Jose C. Palomares,2 Jose A. Mira,1 Antonio Mart’nez,5 Carmen Roldan,1 Julian de la Torre,4 and Juan Mac’as1
 
Units of 1Infectious Diseases and 2Microbiology, Hospital Universitario de Valme, Seville, 3Immunogenetics Unit, Faculty of Sciences, Universidad de Jaen, Jaen, and 4Unit of Infectious Diseases and 5Molecular Genetics Laboratory, Unit of Clinical Analysis, Hospital Universitario Reina Sof’a, Cordoba, Spain
 
"The results of this study show that the IL28B genotype is a strong predictor of SVR to pegylated interferon plus ribavirin in HIV/HCV-coinfected patients with genotype 1 and non-genotype 1 HCV infection. Specifically, rs12979860 genotype CC is associated with a 2-fold higher rate of SVR than TC/TT genotype. IL28B gene variation predicts SVR independently from other well-defined factors that are associated with this outcome in HCV-monoinfected and HIV/HCV-coinfected patients, such as HCV genotype, baseline plasma HCV RNA burden, and sex [1, 11]. After viral genotype, IL28B is the strongest predictor of SVR, being even more potent than HCV RNA load.
 
The effect of rs12979860 variation on SVR had been proven in HCV-monoinfected patients with HCV genotype 1 [2] or in those harboring genotype 1 or 4 as a whole [5]. The present study shows that the impact of rs12979860 genotype is more potent in patients with HCV genotype 1 and in HCV genotype 4 carriers. However, the effect also appears to be exerted in patients with genotype 3.....The mechanisms whereby rs12979860 has an impact on the response to pegylated interferon plus ribavirin is not completely understood.....it is possible that changes in rs12979860 genotype are associated with abnormalities in the IFN-λ3 signal transduction pathway, although functional data are lacking."
 
"Determinations of the IL28B genotype should be incorporated into daily clinical care soon. In fact, the rs12979860 genotype allows us to select a subpopulation with a high likelihood to respond to therapy, and more importantly, used along with other predictors of SVR, it identifies patients with a very low probability of SVR. Thus, just 10% of patients with HCV genotype 1 or 4, a baseline plasma HCV RNA load >600,000 IU/mL, and rs12979860 TC/TT achieved SVR in the population included herein. According to these findings, therapy with pegylated interferon plus ribavirin in patients with the former profile could be deferred until new options are available, at least in patients without advanced fibrosis. This is important because a course of therapy with a low likelihood of success could be spared in almost one-quarter of all patients, because 25% of the participants in this study carried HCV genotype 1 or 4, an HCV RNA load >600,000 IU/mL, and rs12979860 genotype TC/TT. Moreover, the IL28B genotype, along with other baseline predictors of response or the viral kinetics in the early stages of treatment, might allow us to design models to accurately predict SVR or lack thereof, at least in some of the candidates treated with pegylated interferon plus ribavirin."
 
"Beside HCV genotype 2-3 and rs12979860 genotype CC, female sex, baseline HCV RNA level <600,000 IU/mL, an exposure to the planned dosage of HCV therapy >80%, lack of concomitant antiretroviral treatment, and lack of treatment with abacavir were associated with SVR in the univariate analysis (Table 3). The median baseline level of plasma LDL-C was 90 mg/dL (interquartile range, 74-117 mg/dL) in patients who achieved SVR and 73 mg/dL (interquartile range, 61-97 mg/dL) in those who did not (p=.033). In the multivariate analysis, HCV genotype 2-3, rs12979860 genotype CC, plasma HCV level <600,000 IU/mL, and female sex independently predicted SVR (Table 3)."
 
"In the intention-to-treat analysis, 77 patients (46%) attained SVR. Specifically, 33 (30%) of 111 patients with genotypes 1-4 and 44 (76%) of 58 patients with genotypes 2-3 (p<.001) achieved SVR.....rs12979860 genotypes were TT in 20 patients (13%), TC in 66 patients (43%), and CC in the remaining 68 patients (44%)......34 patients (41%) with HCV genotype 1, 1 (100%) with genotype 2, 28 (56%) with genotype 3, and 5 (24%) with genotype 4 harbored rs12979860 CC (p=.049).......HCV viral load among patients with rs12979860 CC was 6.11 log10 IU/mL versus 6.09 log10 IU/mL among those with genotype TC/TT (p=.467).....The frequency of the allele C was significantly higher among patients with SVR (75% vs 56%; p<.005). The rates of SVR according to the rs12979860 genotype were 50% in TT carriers, 29% in TC carriers, and 71% in CC carriers (p<.001). SVR was significantly more common among patients with genotype CC than among those with genotypes TC/TT, considered as a whole (71% vs 34%; odds ratio [OR], 4.7; 95% confidence interval [CI], 2.4-9.4; p<.001). Differences between patients with rs12979860 CC and those with TC/TT regarding SVR were mainly seen in patients with HCV genotype 1 or 4 (Figure 1). The rates of SVR in patients with HCV genotype 1-4 were 54% in CC carriers and 19% in TC/TT carriers (p<.001). The corresponding rates in patients with HCV genotype 2 or 3 were 93% for genotype CC and 77% for TC/TT (p=.104).....Nine nonresponders (18%) had the CC genotype, and 41 (82%) had the TC/TT genotype ( ). No statistically significant difference was found in terms of viral breakthroughs or relapses in relation to rs12979860 (Table 2). The rates of response at each time point of the follow-up with respect to the IL28B genotype are given in Table 2."

 
Figure1. Rate of sustained virologic response (SVR) according to rs12979860 genotype in patients with hepatitis C virus (HCV) genotype 1, 3, and 4.
 

ABSTRACT
 
Background. Variation in the IL28B gene is associated with sustained virologic response (SVR) to pegylated interferon plus ribavirin in hepatitis C virus (HCV)-monoinfected patients with genotype 1. Data on other genotypes and on patients coinfected with human immunodeficiency virus (HIV) and HCV are more limited. We aimed to assess the predictive ability of variations in the single-nucleotide polymorphism rs12979860 for SVR in HIV/HCV-coinfected patients, regardless of HCV genotype.
 
Methods. The rs12979860 genotype was determined by polymerase chain reaction in 154 patients who had received therapy against HCV with pegylated interferon plus ribavirin.
 
Results. rs12979860 genotype was TT in 20 patients (13%), TC in 66 patients (43%), and CC in 68 patients (44%). Rates of SVR in patients with genotype CC and in those with genotype TC or TT, according to HCV genotype, were, respectively, 50% and 17% (p<.001) in patients with genotype 1, 80% and 25% (p=.027) in patients with genotype 4, and 93% and 77% (p=.115) in patients with genotype 3. The median (interquartile range) low-density lipoprotein cholesterol level in patients with rs12979860 CC was 89 mg/dL (73-120 mg/dL) versus 75 mg/dL (55-91 mg/dL) (p=.001) in those with TC or TT. Independent predictors of SVR were HCV genotype 2-3 (odds ratio [OR], 13.98; 95% confidence interval [CI], 4.87-40.1; p<.001), rs12979860 CC (OR, 5.05; 95% CI, 2.04-12.5; p<.001), baseline plasma HCV RNA load of 600,000 IU/mL (OR, 1.99; 95% CI, 1.18-3.34; p=.009), and female sex (OR, 4.28; 95% CI, 1.08-16.96;p=.039 ).
 
Conclusions. IL28B gene variations independently predict SVR in HIV/HCV-coinfected patients with HCV genotype 1 and non-genotype 1 HCV infection. The association between rs12979860 and plasma low-density lipoprotein cholesterol suggests that the system low-density lipoprotein ligand/receptor might be involved in the effect of this genotype.
 
The likelihood of attaining a sustained virologic response (SVR) in patients with chronic hepatitis C virus (HCV) infection depends on viral-, disease-, and host-related factors [1]. Among host-related factors, genetic factors may play a critical role. Thus, it has been proven that polymorphisms near the IL28B gene on chromosome 19, which encodes the type III interferon (IFN-λ3), predict SVR in HCV-monoinfected patients bearing genotype 1 who are treated with pegylated interferon plus ribavirin [2-5]. Specifically, the single-nucleotide polymorphism (SNP) rs12979860, located 3 kilobases upstream of the IL28B gene, is associated with more than a 2-fold difference in the rate of SVR [2]. Likewise, this polymorphism confers a 3-fold higher ability to spontaneously clear HCV [6]. The use of these genetic markers may help us to select patients who are more or less prone to respond to pegylated interferon plus ribavirin. The information on the predictive value of variations in the IL28B gene in patients harboring HCV genotypes other than 1 is more limited, but recent studies have shown that they are also associated with response to pegylated interferon plus ribavirin in genotype 1 or 4 carriers, considered as a whole, but not in those bearing genotype 2 or 3 [5, 7].
 
Patients who are coinfected with human immunodeficiency virus (HIV) and HCV have singularities regarding predictors of SVR. Thus, the overall rate of response in HIV/HCV-coinfected patients is lower than that in HCV-monoinfected patients [8]. Moreover, certain conditions that may have a negative impact on SVR are more common among or exclusive to HIV-infected patients, such as antiretroviral drugs interfering with hepatitis C therapy, CD4+ cell depletion, insulin resistance, steatosis, or advanced fibrosis [8]. A recent study has suggested that the rs12979860 genotype also predicts SVR in HIV/HCV-coinfected patients with genotypes 1-4 considered together [7], but additional studies are required to confirm this point, to know the role of the IL28B genotype in patients with specific HCV genotypes, and to analyze the associations of the IL28B genotype with other factors that may influence SVR in HIV/HCV-coinfected patients.
 
In this study, we aimed to assess whether the polymorphism rs12979860 in the IL28B region independently predicts SVR in a cohort of HIV-infected patients with chronic hepatitis C who were treated with pegylated interferon plus ribavirin without HCV genotype restriction.
 
Discussion
 
The results of this study show that the IL28B genotype is a strong predictor of SVR to pegylated interferon plus ribavirin in HIV/HCV-coinfected patients with genotype 1 and non-genotype 1 HCV infection. Specifically, rs12979860 genotype CC is associated with a 2-fold higher rate of SVR than TC/TT genotype. IL28B gene variation predicts SVR independently from other well-defined factors that are associated with this outcome in HCV-monoinfected and HIV/HCV-coinfected patients, such as HCV genotype, baseline plasma HCV RNA burden, and sex [1, 11]. After viral genotype, IL28B is the strongest predictor of SVR, being even more potent than HCV RNA load.
 
The effect of rs12979860 variation on SVR had been proven in HCV-monoinfected patients with HCV genotype 1 [2] or in those harboring genotype 1 or 4 as a whole [5]. The present study shows that the impact of rs12979860 genotype is more potent in patients with HCV genotype 1 and in HCV genotype 4 carriers. However, the effect also appears to be exerted in patients with genotype 3. Statistically significant differences were not reached in terms of SVR between patients with HCV genotype 3 who bore rs12979860 CC and those who did not in this study, although there was a difference of 16% in the rate of SVR. Recently, an association has also been reported between rs12979860 CC in HIV/HCV-coinfected patients with HCV genotype 1-4 but not in those with genotype 3 [7]. However, both in this case and in our study, this finding might be merely a matter of statistical power.
 
The distribution of rs12979860 genotypes according to the HCV genotypes that the patients harbored was notable. Indeed, rs12979860 genotype CC was significantly more common in patients with genotype 3 than in patients with HCV genotype 1 or 4, similar to what has been reported in HCV-monoinfected patients [12]. The underlying mechanism for this finding is not clear. Whether genotype CC leads the individual to be more prone to infection with HCV genotype 3 or whether infection with HCV genotype 3 becomes chronic more often than does infection caused by other genotypes in patients with rs12979860 CC are topics to be investigated.
 
The mechanisms whereby rs12979860 has an impact on the response to pegylated interferon plus ribavirin is not completely understood. In previous studies involving HCV-monoinfected patients, allele C was unexpectedly associated with a higher baseline plasma HCV RNA level [2]. This finding has not been confirmed in this study. In any case, rs12979860 CC does not seem to negatively influence the replication of HCV, at least in untreated patients. This SNP has a strong linkage disequilibrium with a nonsynonymous coding variant in the IL28B gene (213A>G, K70R; rs81031142) [2]. Thus, it is possible that changes in rs12979860 genotype are associated with abnormalities in the IFN-λ3 signal transduction pathway, although functional data are lacking. IFN-λ1, another type III interferon, inhibits HCV replication, increases the levels of interferon-stimulated genes, and enhances the antiviral effect of interferon alfa [13]. It is conceivable that IFN-λ3, a closely related cytokine with activity against other viruses comparable to that of IFN-λ1 [14], works in a similar way against HCV [6]. However, the lack of association between rs12979860 genotype CC and lower baseline plasma HCV RNA burden argues against this hypothesis.
 
The association between rs12979860 genotype and plasma levels of LDL-C is striking [15]. In vitro studies have shown that LDL may competitively inhibit the binding of HCV to the LDL receptor, which functions as one of the cellular receptors for HCV [16, 17]. This competitive blockade would hamper the infection of hepatocytes with HCV [18]. Accordingly, higher levels of plasma LDL-C have been shown to be an independent predictor of SVR, both in HCV-monoinfected [19, 20] and in HIV/HCV-coinfected patients [21], in studies specifically designed to appraise this issue. Likewise, SNP in LDL receptor, similar to rs12979860 variations, are associated with both response to treatment and spontaneous clearance of HCV [22]. How variations in rs12979860 could determine the levels of LDL-C is unclear. Certain soluble LDL receptor isoforms are induced in response to interferon stimulation [23]. We could speculate that an rs12979860 genotype other than CC could induce soluble isoforms of the LDL receptor, which join to plasma LDL, decreasing LDL levels and allowing an easier entrance of HCV into the hepatic cell. Studies aimed to search for a genetic interaction between the IL28B locus and the LDL receptor or the LDL ligands and LDL receptor system are warranted. In the meantime, an effect on this system should be regarded as one of the putative underlying mechanisms that explain the impact of IL28B gene variations on the spontaneous and drug-induced clearance of hepatitis C virus.
 
This study has several limitations. A pretherapy liver biopsy and a determination of baseline LDL-C level were not available in all patients. Concerning preexisting advanced liver fibrosis, an assessment of this parameter could be performed in most participants using either biopsy or transient elastography. In any case, the potential impact of LDL-C level and advanced liver fibrosis on the probability of SVR was much lower than that of rs12979860 genotype. In addition, insulin resistance, a factor that has been reported to be associated with a lower rate of SVR [24], was not measured here. However, the role of insulin resistance in response to HCV therapy in HIV-coinfected patients is controversial [25, 26]. Because of all these reasons and the results of previous studies, it is extremely unlikely that the association between the IL28B genotype and SVR found in this study is the result of confounding factors. On the other hand, these results provide data on the predictive value of IL28B gene variations coming from daily clinical practice, outside randomized clinical trials. Moreover, we provide specific data on this factor in HCV genotype 3 and 4 infections. These are strengths of this study.
 
Determinations of the IL28B genotype should be incorporated into daily clinical care soon. In fact, the rs12979860 genotype allows us to select a subpopulation with a high likelihood to respond to therapy, and more importantly, used along with other predictors of SVR, it identifies patients with a very low probability of SVR. Thus, just 10% of patients with HCV genotype 1 or 4, a baseline plasma HCV RNA load >600,000 IU/mL, and rs12979860 TC/TT achieved SVR in the population included herein. According to these findings, therapy with pegylated interferon plus ribavirin in patients with the former profile could be deferred until new options are available, at least in patients without advanced fibrosis. This is important because a course of therapy with a low likelihood of success could be spared in almost one-quarter of all patients, because 25% of the participants in this study carried HCV genotype 1 or 4, an HCV RNA load >600,000 IU/mL, and rs12979860 genotype TC/TT. Moreover, the IL28B genotype, along with other baseline predictors of response or the viral kinetics in the early stages of treatment, might allow us to design models to accurately predict SVR or lack thereof, at least in some of the candidates treated with pegylated interferon plus ribavirin.
 
In summary, variation in the IL28B locus is a more potent predictor of response in HIV/HCV-coinfected patients than others currently used, such as plasma HCV RNA load. Its effect is evident not only in patients with HCV genotype 1 but also in those with HCV genotype 4 and, probably, in genotype 3 carriers. The SNP rs12979860 correlates with plasma LDL-C level, which might play a role in the mechanism of action of this polymorphism. The use of this genotype in routine clinical practice may select patients with very high or very low likelihood of therapy success.
 
Results
 
Features of the study population.

 
All patients were of European ancestry. Nine (5%) of 169 patients who started therapy discontinued it because of adverse events, and 6 (4%) voluntarily dropped out. Consequently, 154 patients constituted the on-treatment population. The main baseline characteristics of this group are given in Table 1. Baseline plasma HCV RNA load was <600,000 IU/mL in 62 patients (40%). CD4+ cell counts were <250 cells/mm3 in 17 patients (11%).
 
Response to HCV therapy.
 
In the intention-to-treat analysis, 77 patients (46%) attained SVR. Specifically, 33 (30%) of 111 patients with genotypes 1-4 and 44 (76%) of 58 patients with genotypes 2-3 (p<.001) achieved SVR. The corresponding figures in the on-treatment analysis were 50% for the overall population, 32% for genotype 1-4 carriers, and 86% for those harboring genotype 2-3 (p<.001).
 
IL28B genotype.
 
rs12979860 genotypes were TT in 20 patients (13%), TC in 66 patients (43%), and CC in the remaining 68 patients (44%). These genotypes were in the Hardy-Weinberg equilibrium (p=.97). The frequency of the allele C was significantly higher among patients with SVR (75% vs 56%; p=.005). The rates of SVR according to the rs12979860 genotype were 50% in TT carriers, 29% in TC carriers, and 71% in CC carriers (p<.001). SVR was significantly more common among patients with genotype CC than among those with genotypes TC/TT, considered as a whole (71% vs 34%; odds ratio [OR], 4.7; 95% confidence interval [CI], 2.4-9.4; p<.001) (Figure 1). Differences between patients with rs12979860 CC and those with TC/TT regarding SVR were mainly seen in patients with HCV genotype 1 or 4 (Figure 1). The rates of SVR in patients with HCV genotype 1-4 were 54% in CC carriers and 19% in TC/TT carriers (p<.001). The corresponding rates in patients with HCV genotype 2 or 3 were 93% for genotype CC and 77% for TC/TT (p=.104).
 
Nine nonresponders (18%) had the CC genotype, and 41 (82%) had the TC/TT genotype (p<.001). No statistically significant difference was found in terms of viral breakthroughs or relapses in relation to rs12979860 (Table 2). The rates of response at each time point of the follow-up with respect to the IL28B genotype are given in Table 2.
 
The distribution of rs12979860 genotypes in carriers of different HCV genotypes was not uniform. Thus, 34 patients (41%) with HCV genotype 1, 1 (100%) with genotype 2, 28 (56%) with genotype 3, and 5 (24%) with genotype 4 harbored rs12979860 CC (p=.049). Median (interquartile range) baseline HCV viral load among patients with rs12979860 CC was 6.11 log10 IU/mL versus 6.09 log10 IU/mL among those with genotype TC/TT (p=.467). There was a strong relationship between rs12979860 genotype and the baseline level of plasma low-density lipoprotein cholesterol (LDL-C) level (Figure 2). Thus, the median (interquartile range) LDL-C level in patients with rs12979860 CC was 89 mg/dL (73-120 mg/dL) versus 75 mg/dL (55-91 mg/dL) (p=.001) in those with genotype TC/TT.
 
Predictors of SVR. Beside HCV genotype 2-3 and rs12979860 genotype CC, female sex, baseline HCV RNA level <600,000 IU/mL, an exposure to the planned dosage of HCV therapy >80%, lack of concomitant antiretroviral treatment, and lack of treatment with abacavir were associated with SVR in the univariate analysis (Table 3). The median baseline level of plasma LDL-C was 90 mg/dL (interquartile range, 74-117 mg/dL) in patients who achieved SVR and 73 mg/dL (interquartile range, 61-97 mg/dL) in those who did not (p=.033). In the multivariate analysis, HCV genotype 2-3, rs12979860 genotype CC, plasma HCV level <600,000 IU/mL, and female sex independently predicted SVR (Table 3).
 
Thirty-nine patients (25%) carried HCV genotype 1-4, rs12979860 TC/TT, and a plasma HCV load 600,000 IU/mL. Only 4 (10%) of these individuals attained SVR.
 
Table 3. Predictors of Sustained Virologic Response in the Univariate and Multivariate Analyses

Methods
 
Study cohort.

 
From October 2001 through June 2008, a cohort of 169 HIV/HCV-coinfected patients, previously naive for pegylated interferon and ribavirin, consecutively started therapy for chronic HCV infection in 2 tertiary care centers in southern Spain. Patients were prospectively followed up. Visits were scheduled at least every 4 weeks during the first 24 weeks of treatment and every 8-12 weeks thereafter. In addition, all patients were assessed 24 weeks after stopping therapy. At each visit, clinical, biochemical, and hematologic assessments were performed. A whole blood sample was collected from all patients and cryopreserved at -70°C for genetic determinations.
 
Drug therapy.
 
All patients were given pegylated interferon alfa-2a at a dosage of 180 µg once per week or pegylated interferon alfa-2b at a dosage of 1.5 µg/kg once per week, both in combination with ribavirin at a daily dose of 800-1200 mg. Patients harboring HCV genotype 2 or 3 received HCV therapy for 24 weeks if they had an undetectable plasma HCV RNA load at week 4. The length of therapy was 48 weeks in the remaining patients. At weeks 12 and 24, HCV therapy was prematurely discontinued in nonresponders.
 
Definition of viral response.
 
The outcome variable in this study was SVR, defined as undetectable HCV RNA in serum 24 weeks after the completion of HCV therapy. A decrease in plasma HCV RNA level >2 log10 or below the detection threshold at week 12 was considered to be an early virologic response. An end-of-treatment response was defined as undetectable plasma HCV RNA at the completion of therapy. Patients without early virologic response, as well as those with detectable plasma HCV RNA at week 24, were considered to be nonresponders. Virologic breakthrough was defined as detectable plasma HCV RNA after week 24 of therapy in patients with a previous undetectable HCV load. Relapse was defined as a lack of SVR after having reached end-of-treatment response.
 
The plasma HCV RNA load was measured using a quantitative polymerase chain reaction assay according to the available technique (Cobas Amplicor HCV Monitor [Roche Diagnostic Systems], with a detection limit of 600 IU/mL; Cobas AmpliPrep-Cobas TaqMan [Roche Diagnostic Systems], with a detection limit of 50 IU/mL; and Cobas TaqMan [Roche Diagnostic Systems], with a detection limit of 10 IU/mL).
 
Determination of the IL28B genotype.
 
DNA was extracted using the automated MagNA Pure DNA extraction method (Roche Diagnostics). The rs129679860 SNP was genotyped using a custom TAQMAN genotyping assay (Applied Biosystems) on DNA isolated from whole blood samples. The DNA was genotyped according to the manufacturer's instructions on a MX3005 thermocycler using MXpro software (Stratagene). The researchers responsible for genotyping procedures were unaware of other data from the patients.
 
Data analysis.
 
Hardy-Weinberg equilibrium was calculated using Haploview software (http://www.broadinstitute.org/haploview/haploview) [9]. The association between SVR and rs12979860 genotype was analyzed. A dominant model (TT=TC< CC) was used. Likewise, we assessed the relationship between SVR rate and parameters that might have an impact on the response to HCV therapy. For this analysis, advanced fibrosis was defined as a stage of fibrosis of F3 or higher, according to Scheuer's scoring system [10], in patients who had undergone a pretreatment liver biopsy, or as a baseline liver stiffness of 11 kPa, as determined by transient elastography (FibroScan; Echosens), in those who had not undergone a pretherapy liver biopsy. Two sensitivity analyses were performed for estimating SVR. On the one hand, we conducted an intention-to-treat approach, considering all noncompleters or missing patients as having experienced treatment failure. On the other hand, an on-treatment analysis, excluding patients who dropped out or discontinued therapy because of adverse events, was performed. The associations between SVR and these variables were appraised on an on-treatment basis.
 
The frequencies were compared using the χ2 test or Fisher's exact test. The Student's t test was used for comparisons among continuous variables in the 2 groups if a normal distribution was followed, and the Mann-Whitney U test was used if a normal distribution was not followed. For comparing continuous variables in >2 groups, the Kruskal-Wallis test was used. The median was used as the cutoff value when continuous variables were categorized, unless otherwise specified. Variables associated with SVR in the univariate analysis with p<.20 were entered in logistic regression models, where SVR was the dependent variable. The analysis was performed using the SPSS statistical software package, version 15.0 (SPSS), and the Stata/SE 9 package (Stata).
 
Ethical aspects. The study was designed and performed according to the Helsinki Declaration and was approved by the ethics committees of both participating hospitals. All patients provided written informed consent to participate in this study.
 
 
 
 
  iconpaperstack View Older Articles   Back to Top   www.natap.org