First-wave protease inhibitors for hepatitis C genotype 1 treatment: a real-life experience in Brazilian patients

INTRODUCTION

The use of first-wave protease inhibitors (PIs) boceprevir (BOC) and telaprevir (TVR) was the first step in direct antiviral therapy against hepatitis C virus (HCV). Initial studies showed a promising increase in the sustained viral response (SVR) among patients infected with HCV genotype 1, despite high costs and increased adverse drug reactions (ADRs), which often resulted in early treatment discontinuation^1–5.

Licensed for the treatment of chronic hepatitis C in 2011 and registered in Brazil in the same year, TVR and BOC were incorporated into the Brazilian public health system in 2013. According to Brazilian Ministry of Health (MH) recommendations, patients with HCV genotype 1 infection with moderate or severe liver fibrosis and/or presenting extrahepatic manifestations (EHMs) were indicated for TVR or BOC based therapy⁶. Between 2013 and 2015 in Brazil, individuals with HCV genotype 1 infection were treated with pegylated interferon (Peg-INF) and ribavirin (RBV) in addition to one of the first-wave PIs. The experience gained by healthcare teams from reference centers authorized for the care of patients with hepatitis C led to the incorporation of new concepts of HCV treatment, including the discussion on direct-action-drug resistance induced by select mutant viral strains.

There are few published, real-life studies that have evaluated the use of first-generation PIs in Latin America, particularly in Brazil^7–9. Thus, this study was aimed at analyzing the safety and effectiveness of the above-mentioned triple regimens, in addition to sharing the experience of a referral center established at a tertiary/quaternary university hospital in Southeastern Brazil.

METHODS

RESULTS

A total of 115 patients were included, of which 78 (67.8%) were men, 96 (83.5%) were white, 58 (50.4%) had liver cirrhosis, 74 (64.3%) had been treated previously, and 103 (89.6%) used TVR. The age range was 29 to 74 years. These and other characteristics related to the clinical condition of the patients are described in Table 1.

Among patients diagnosed with EHMs, four presented with late porphyria cutanea tarda and two with mixed cryoglobulinemia without renal involvement. About treatment outcomes, considering the total number of individuals treated with triple therapy, 71 (61.7%) achieved SVR12 and nine (7.8%) were relapsers. For 65 (91.5%) of the 71 patients with SVR12, the 24-week post-treatment viral load result was available and HCV RNA remained undetectable. With regard to only those patients who used TVR, the treatment was discontinued for seven (6.8%) patients because of virological breakthrough, while the stopping rule motivated the suspension of treatment for eight (7.8%) individuals. Early discontinuation of treatment for another seven (6.8%) patients can be explained by virological breakthrough and concomitant application of the stopping rule.

Regarding patients who used TVR, the treatment of 13 (12.6%) was suspended due to ADRs. Of these individuals, for nine, therapy was discontinued because of an isolated ADR (six presented with a skin rash, one with anemia, one with decompensated liver disease, and one with psychiatric manifestations). Four other patients showed at least two adverse events that led to the discontinuation of therapy: one presented with a skin rash, anemia, neutropenia, and altered renal function; another presented with the four events associated with thrombocytopenia and gastrointestinal disorders (nausea/vomiting); the third presented with anemia, neutropenia, and altered renal function; and the fourth presented with thrombocytopenia and decompensated liver disease. Four patients who were treated for a shorter time than expected due to ADRs had SVR: two patients with a cutaneous rash were treated for eight and 10 weeks respectively, a patient with anemia whose treatment lasted 14 weeks, and a patient whose treatment suspension was motivated by six associated adverse events and treated for 11 weeks. Among these four individuals, only those who were treated for eight weeks had no cirrhosis. There was no death among the patients treated with TVR.

With regard to the patients for whom BOC was indicated, the pharmacological therapy for one (8.3%) patient was discontinued due to the stopping rule. Suspension of treatment of two other individuals can be explained by virological breakthrough and the stopping rule. Pharmacological therapy of two patients treated with BOC (16.7%) was suspended because of ADRs: one patient had anemia and the other had neutropenia. The patient who showed a significant decrease in neutrophil levels used triple therapy for only 12 weeks, showed cirrhosis of the liver, and achieved SVR. Similar to the TVR group, there was no death among patients in the BOC group. Figure 1 summarizes the treatment outcomes for each drug.

The SVR rate among individuals with cirrhosis was 51.7%. Statistical analyses showed an association between liver cirrhosis and a lower chance of achieving SVR with triple therapy. With regard to previous treatment, 76.9% of the relapsers achieved SVR. There was evidence that these patients had a greater chance of achieving SVR relative to the non-responders. On the other hand, there was no evidence of an association between SVR and the following variables: previous treatment (treated vs. untreated), type of PI used, DM, obesity, viral subgenotype, pretreatment viral load, and presence of esophageal varices (Table 2). The multivariate analysis showed that cirrhosis would be the most strongly associated variable with treatment response (Table 3).

Regarding the univariate analysis and the clinical variables associated with SVR, the TVR and BOC groups were similar in relation to subgroup frequencies, particularly with regard to the variable previous treatment response. Among the patients who used TVR, who were previously treated, and for whom previous therapy data were available, 53.2% were relapsers and 46.8% were non-responders; among those who used BOC, 54.5% were relapsers and 45.5% were non-responders. Considering the variable cirrhosis, 54.3% of patients for whom data related to hepatic fibrosis were available and received TVR treatment were classified as F4, while 63.6% of individuals who used BOC were cirrhosis patients.

Hematologic events, asthenia, cutaneous reactions, anorectal manifestations, gastrointestinal complaints (nausea and/or vomiting), neuropsychiatric disorders, loss of appetite, dry cough, dyspnea, dysgeusia, and flu-like symptoms were the main adverse reactions observed for all patients included. The frequencies of cutaneous, anorectal, and nausea and/or vomiting events were higher in the TVR group, while that of dysgeusia was higher in the BOC group (Table 4).

DISCUSSION

As observed in the major phase 3 clinical trials involving TVR and BOC^1–5, the SVR in this study were higher than the response rates associated with the use of dual therapy with Peg-INF and RBV, and in general, did not exceed 50%^14,15. On the other hand, they were lower than those found in the pre-marketing clinical trials of the above-mentioned PIs. Considering real-life data, the SVR rates were similar to those described in two European studies, one of which was a multicenter study involving 47 different centers in Italy^16,17 and higher than the SVR rates of real-life studies conducted in the USA^18,19. The non-statistical association of the SVR with the PI used can be explained by the important difference between samples that limits the robustness of the analysis and by the similarity of the groups in relation to the variables associated with the SVR.

The statistical difference between the SVR and PI used can be explained by the significant difference in sample sizes that limits the robustness of the analysis and the similarity between the groups in relation to variables associated with the response, such as the type of response to therapy.

Regarding the variables associated with SVR, evidence suggested that the diagnosis of cirrhosis is associated with a lower chance of SVR. This finding corroborates findings from several other studies, including phase 3 clinical trials and real-life studies^2,3,17–20, in addition to a meta-analysis published by Pecoraro et al²¹. The type of response to previous treatment was another variable statistically associated with SVR. Relapsers after dual therapy were more likely to achieve SVR than patients who have had another type of response since this also corroborates findings from other studies^{3,4,17,20,22,23}. A binary logistic regression analysis indicated that absence of cirrhosis would be more strongly associated with SVR. Studies suggest that cirrhosis is also associated with response to the available therapies prior to TVR and BOC approval, which may explain the outcome of the multivariate analysis^13,14.

About the variables for which no evidence of association with SVR was found, Ascione et al¹⁷ indicated, through multivariate analysis, that there is a positive association between not having diabetes and SVR. Although studies on the first-generation PIs are scarce, data from the literature involving dual therapy and analyses of the pathophysiological mechanism suggest a relationship between obesity and lack of SVR^24–26. In the present study, the analysis of these two variables was limited by the high number of patients for whom no information related to body mass index was available and the large difference between the numbers of patients with and without diabetes.

With regard to the relationship between the pretreatment HCV RNA levels and SVR, the results of published studies are contradictory and these studies used different viral load values to define high viral load and low viral load^{2,3,5,16,17,19,22,27}. In a meta-analysis, Cooper et al²⁸found no statistically significant difference between the SVR of treatment-naive patients and treatment-experienced patients, a finding similar to that observed in this study.

Another result supported by studies published in the literature is the non-association between SVR and viral subgenotype^5,16,17,27. Regarding the possible association between the type of PI used and SVR, this study also corroborates data from other studies^16,20,28. There was no evidence of an association between the diagnosis of esophageal varices and SVR. However, for this analysis, only patients with cirrhosis were considered, and the small sample size may have been a limiting factor.

Regarding safety, while phase 3 studies indicated that about 40% of patients who used TVR had anemia^1,4,5, 95.1% of the patients in this study who used the same drug had anemia. In general, all events related to the use of TVR, including other commonly observed events such as other hematological ADRs (neutropenia and thrombocytopenia), flu-like symptoms, asthenia, cutaneous events such as rash and pruritus, and anorectal discomfort were more frequent than in the previous studies^4,5,17,23. A multicenter, real-life study carried out in Spain showed frequencies of anemia and other hematological events for TVR to be similar to those in the present study²¹. The main ADRs that caused discontinuation of pharmacological therapy were rash and anemia, similar to that in phase 3 and real-life studies^1,4,5,17.

About safety related to BOC use, as observed in most previous studies, cutaneous and anorectal reactions were less frequent than in the TVR group. On the other hand, the most common ADRs were more frequently observed in the BOC group in comparison to the frequencies in previously published studies involving BOC^2,3,17,22. The frequencies of hematological events in the previously mentioned Spanish multicenter study were similar to those in the present study⁹. Neutropenia and dysgeusia were frequent adverse events associated with BOC and more commonly related to the use of BOC relative to the use of TVR; this finding is supported by those of other studies^2,21,28,29. The two BOC treatment interruptions due to ADRs were attributable to hematological reactions (anemia and neutropenia), which is explained by the high frequency of these events. It is important to emphasize that hemoglobin levels used for the diagnosis of anemia are not uniform across studies. However, the incidence of anemia in this study was higher, even if only the incidence of more severe anemia (which required the use of erythropoietin) was compared with the incidence of anemia in the other studies.

This study has limitations. It is a retrospective study based on data collection using a computerized system. Studies with this methodological design may be associated with an incomplete information retrieval due to underreporting of clinical data. Furthermore, there was no validated instrument or direct method used to measure adherence to pharmacological treatment. Considering that non-adherence is an important factor influencing the lack of SVR for antiviral drug therapies, it is a limiting factor for the interpretation of the results. Additionally, there is little information on this type of therapy in Latin American and Brazilian literature. Furthermore, although this study was carried out at a regional university reference center, which caters to 1,300,000 individuals, the sample size should be considered a potential limitation as it is relatively small and comprises patients treated at a single health center.

In conclusion, this study, which shares experience related to the indication of the first-wave PIs in triple therapy for Brazilian patients with HCV genotype 1 infection, showed lower SVR rates than those shown by pre-marketing studies, but comparable to those found in real-life studies performed in other populations. The frequency of ADRs, particularly hematological ADRs, was high and resulted in discontinuation of treatment in a considerable number of cases.