First-generation protease inhibitor-triple therapy: SVR 24, safety, and predictors of response in a large single center cohort

The clinical features of our study cohort are presented in Table 1. This retrospective analysis includes all 131 consecutively recruited patients, who were set on a PI-triple therapy including PR and TVR (102 patients), or PR and BOC (29 patients) between July 2011 and May 2012 at our center. From July 2011 until beginning of September 2011 all patients were treated with BOC-triple therapy, since TVR was not approved then. However, after approval of TVR in late September 2011, almost all patients were treated with the TVR-triple therapy at our center. This special management of patients explains the disproportion in numbers of both groups of patients.

The treatment course for TVR-based triple therapy consisted of 12 weeks of triple therapy with TVR and PR, followed by a dual therapy with PR. The length of that second dual therapy period was variable, according to the approved treatment recommendations: if patients were (i) treatment naïve or relapsers to a classical dual treatment regimen consisting of PR alone, (ii) had no liver fibrosis of higher grade, and (iii) exhibited a negative HCV viral load (Cobas AmpliPrep/Cobas TaqMan HCV Test, Roche Diagnostics GmbH, Mannheim, Germany; lower limit of detection [LLOD] and lower limit of quantification [LLOQ] 15 IU/ml,) after 4 weeks of triple therapy, they were eligible for treatment shortening down to altogether 24 weeks; if not so, treatment was extended to 48 weeks. According to guidelines, those patients with non-pegylated IFN as last treatment were eligible for treatment shortening as well, even if they were classified as non-responders as last treatment response. Patients, who presented with a HCV viral load above 1000 IU/ml at treatment week (TW) 4 or at any time afterwards had to discontinue treatment prematurely; the same applied for patients, who exhibited a rise in HCV viral load of 1 log₁₀ above a previous nadir viral load.

BOC-based triple therapy was conducted in accordance with the European and German regulations (which differ from the treatment schedule in the US): following a “lead-in” phase of four weeks of sole PR treatment, BOC was added to the dual treatment regimen. Duration of triple therapy differed according to the respective fibrosis grade and previous treatment responses: if patients were treatment naïve and achieved a HCV RNA level below LLOD after TW 8, treatment duration could be shortened to 28 weeks altogether. When a patient was treatment naïve, and failed to achieve a HCV RNA level below LLOD at TW 8, but reached this goal thereafter (until TW 24) and additionally had no proven liver cirrhosis, then BOC could be abandoned after TW 36, followed by additional 12 weeks with PR alone until TW 48. If the patient had been on PR previously (resulting in virological relapse, breakthrough, or confirmed “partial response”), demonstrated a HCV viral load below LLOD after TW 24, and additionally had no proven liver cirrhosis, then BOC could be abandoned after TW 36, followed by additional 12 weeks with PR alone until TW 48. In patients with liver cirrhosis and/or previous non-response (null-response) to a PR treatment, BOC had to be given in addition to PR from TW 4 on to TW 48. Patients, who presented with a HCV viral load above 100 IU/ml at TW 12 or at any time afterwards had to discontinue treatment prematurely; the same applied for patients, who exhibited a rise in HCV viral load of one log₁₀ above a previous nadir viral load. Due to our status as a tertiary referral center (being quite used to referrals of cases exhibiting incomplete data on viral loads in previous treatment attempts performed “outside”), we a priori did not differentiate between so called “null-response” and “partial response” patients, and subsumed those patients as “non-response” patients, a procedure which was commonly accepted in the pre-PI era.

In response to anemia, we administered EPO based on individual decisions as follows: (i) if patients were symptomatic or (ii) if we assumed that RBV dose reduction could be avoided or (iii) if we assumed that we would have to reduce RBV dosing to a lesser extent when applying EPO. Granulocyte-colony stimulating factor (G-CSF) was given if patients exhibited leukocyte counts below 1000/μL.

Baseline laboratory values are shown in Table 1. 4/131 patients presented with leukopenia (leukocyte count < 2500/μl), 15/131 patients had a moderate (platelets < 100/nl) and 2/131 a severe thrombopenia (platelets < 50/nl). All of them, except 3 patients with moderate thrombopenia, were in the TVR patient group. One of those patients had immune-mediated thrombopenia. Transaminases were found to be elevated in 113/131 patients (GPT > 35 IU/ml: TVR-group 87/102; BOC-group 26/29, respectively). 3/131 patients exhibited a Quick value below 60% (one of those was on phenprocoumon for atrial fibrillation); all of them were in the TVR subset of patients. Baseline and follow-up data of albumin levels were available only on an occasional basis. Data were statistically analyzed using Microsoft Office Excel, Graph Pad Prism 6.0, and SPSS 21.

Overall viral response rates and SVR 24 rates according to previous treatment response (naïve, relapse, non-response, non-pegylated IFN pretreated) are shown in Figures 1 and 2. In the overall cohort, 73% (95/131) of patients achieved SVR 24. Four patients were lost to follow-up. Of those four patients, 2 had reached SVR 12, but then did not show up for follow-up visits. SVR 24 rates in patients, (i) who were treatment naïve, (ii) suffered from virological relapse after a PR treatment, or (iii) had a non-PEG-IFN-based previous treatment, ranged from 77-84%, irrespective of the administered type of PI-triple therapy.

For evaluation of predictors of SVR 24, see Tables 2 and 3. By univariate analysis of the overall cohort, extended rapid virological response (eRVR; p < 0.0001) and a negative viral load at PI-TW 4 (p < 0.0001) were significantly associated with an increased likelihood of SVR 24, while a baseline platelet count of < 100/nl (p < 0.0001), a viral load above LLOQ at TW 4 (p < 0.0001), and a previous non-response to PR therapy (p = 0.006) were negative predictors of SVR 24. In the TVR-group of patients, the best positive predictor for SVR 24 was an eRVR, being defined as negative HCV-viral loads (below LLOD) at both TW 4, and 12, respectively. In our TVR-cohort, 64 out of 102 (63%) patients met those criteria. When TVR-treated patients achieved eRVR, the probability of achieving a SVR 24 later on was 91% (58/64 patients, p < 0.0001, Fisher’s exact test). For BOC triple therapy, due to the different treatment schedule, eRVR is defined as negative HCV-viral loads at both TW 8, and 24, respectively. In our BOC-cohort, patients, who achieved eRVR had a probability of 81% (13/16 patients) of achieving a SVR 24.

For the overall cohort, the probability of achieving a SVR 24 was lower in patients, who had a detectable, but not quantifiable viral load (below LLOQ) at TW 4: In this subset of patients, only 69% (20/29 patients) achieved a SVR 24 (p = 0.0193, Fisher’s exact test). Reasons for treatment failure were virological failure (n = 4), premature discontinuation (n = 4), and relapse (n = 1).

At least in our cohort, for TVR-patients the following prediction could be made: if patients showed a quantifiable viral load 4 weeks after onset of PI-administration (viral load above LLOQ at TW 4), no patient (0/7 patients) achieved a SVR 24 later on, in all cases due to virological failure (p < 0.0001, Fisher’s exact test).

15 out of 131 patients (12 patients TVR-treated, 3 patients BOC-treated, respectively) had platelet counts below 100/nL at baseline thus resembling an extra-“difficult-to-treat” group of patients [16]. Of those, just two patients achieved SVR 24 after regular completion of a full treatment course (13%), and thus had a significantly lower SVR 24 rate than others (p = 0.0001, Fisher’s exact test).

Patients who previously had shown a virological non-response to PR treatment, exhibited lower SVR 24 rates than other subgroups (p = 0.0075), especially in the BOC group of patients (p = 0.0014, Fisher’s exact test).

Almost all patients with a virological failure during PI-triple therapy (13/14 patients) had a high viral load (> 800.000 IU HCV RNA/ml) at baseline. However, no statistical significance was achieved with respect to predictability of SVR 24, at least in our cohort. Older patients (> 60 years), knowingly prone to side effects, are “difficult-to-treat” patients [17]. However, in our cohort, irrespective of previous treatment status, we could achieve favorable SVR rates in this subgroup of patients (73% with TVR, and 4 out of 5 patients with BOC, respectively).

In the TVR-subset, patients with liver cirrhosis achieved SVR 24 rates of 71%, and remarkably, in the subgroup of TVR-patients, who had both liver cirrhosis and were of older age, a SVR 24 rate of 67% (8/12 patients) could be achieved. In the BOC group of patients with liver cirrhosis 2 out of 7 patients achieved SVR 24.

Multivariate analysis identified eRVR, a baseline platelet count of > 100/nl, and previous response status as independent predictors of higher SVR 24 rates (see Table 3 for odds ratios, and confidence intervals).

In the TVR-group at baseline 48 of the 102 patients had no fundamental contraindications for a shortened treatment course of 24 weeks, as e.g. liver cirrhosis or previous non-response to a PR therapy. Of those, 34 out of 48 (71%) were found to reach a HCV viral load below LLOD at TW 4, thus meeting the basic requirement for treatment shortening. In fact, treatment shortening finally was implemented in only 26 out of 48 patients (54%). 25 out of those 26 patients (96%) achieved SVR 24, whereas 1 patient was lost to follow-up. With BOC-triple therapy, 4 out of 7 patients, who had no basic contraindications for shortening of treatment at baseline, met all criteria at TW 8 for an implementation of such a procedure. Of those, two patients achieved SVR 24, while one patient suffered from virological relapse and another patient was lost to follow-up.

Altogether, 14 of the 131 patients (11%) had to stop PI-based triple therapy due to on-treatment virological failure (see Figure 1). Additionally, 4 patients (3%) suffered from virological relapse during follow-up, all of those occurred between EOT and first follow-up, which was routinely done 12 weeks after EOT. 10 of the 102 TVR-patients (10%) suffered from on-treatment virological failure (4 patients meeting the futility rule of a viral load > 1000 IU/ml, 3 patients with virological breakthrough at TW 12, 18, and 24, respectively, and 3 patients with rising viral loads after having reached a nadir). Additionally, 1 patient experienced a virological relapse. In the TVR-group of patients no virological failure occurred after TW 24. In the BOC-group of patients, altogether 4 out of 29 patients (14%) experienced an on-treatment virological failure: 3 patients suffered from virological non-response, when meeting the futility rule (HCV viral load > 100 IU/ml) at TW 12 or later. One patient suffered from virological breakthrough at TW 48. Additionally, 3 patients (10%) experienced a virological relapse between EOT and first follow-up visit.

Altogether, 33 out of 131 (25%) patients prematurely discontinued PI-triple therapy. Of those, 14 discontinued due to virological failure (see above) and 18 due to side effects. One additional patient belonging to the TVR-patient cohort was treated after having been listed for liver transplantation due to hepatocellular carcinoma. This patient qualified for an exceptional MELD score before and during therapy. At TW 36 an organ offer by EUROTRANSPLANT had been accepted for this patient followed by successful transplantation. This patient therefore discontinued treatment prematurely. Nevertheless, this patient stayed HCV negative thereafter in a sustained fashion (SVR 24). In the TVR-group of patients, 14 other patients prematurely discontinued complete treatment: 4 patients due to rash, 3 patients due to infections, 3 patients due to intolerance of the treatment, 2 patients due to hepatic decompensation, 1 patient due to intracerebral bleeding, and 1 patient due to lung cancer. Despite their premature discontinuation of treatment, three of those 14 patients achieved a SVR 24, including one patient, who had to undergo LTx due to hepatic decompensation at TW 4.

In the BOC-group of patients, 4 patients discontinued treatment prematurely: Two patients due to intolerance to the treatment, one patient due to non-treatment associated posttraumatic stress syndrome, and one patient due to pruritus. The latter one nevertheless achieved a SVR 24.

Remarkably, 14 out of the 18 discontinuations which had been due to side effects had occurred in patients of older age (≥ 60 years), or in patients with liver cirrhosis, among those 5 patients exhibiting a liver cirrhosis at age ≥ 60 years. This again emphasizes the “difficult-to-treat” status of these subgroups.

In 4 of the 102 TVR-treated patients TVR alone was withdrawn prematurely due to development of a severe rash. Additionally, one of the 29 BOC-treated patients discontinued taking BOC due to diarrhea. Subsequently, PR treatment was continued in all patients. All five patients achieved SVR 24 later on.

Side effects were frequent during PI-triple therapy, and so were therapeutic interventions. However, no patient was lost during therapy.

Hematological toxicity is a specific feature of PI-triple therapy, irrespective of which PI has been used (see Table 4 for details). Thus, therapeutic interventions were frequent and had to be performed in 36% (TVR), and 28% (BOC), respectively. The difference between the high rate of anemia below 10 g/dL (45%), and the lower rate of RBV reduction is most probably due to our policy, not to reduce RBV dosage in asymptomatic or oligosymptomatic patients with anemia with hemoglobin values above 8.5 g/dL during the first phase of PI-triple therapy. When at the EASL 2012 preliminary results from the study of Poordad et al. [18] were published, showing no difference in outcome if the RBV dosage is significantly reduced, we changed our policy, and reduced RBV dosage at an earlier stage, contrary to the treatment with PEG-IFN and RBV alone, where RBV reduction if ever possible should be avoided. Nevertheless, no patient had to discontinue treatment due to anemia.

Severe rash was a rare event in former dual treatment PR regime, and mostly was attributed to RBV. In PI-triple therapy, pruritus and rash seem to constitute a PI class effect, but in frequency and intensity much more associated with TVR than with BOC: 11 of the TVR-treated patients developed a severe rash, resulting in the necessity of a complete discontinuation of treatment in 4 patients, and premature discontinuation of TVR alone in additional 4 patients.

During PI-triple therapy, hospital admissions were required in 23 patients (21 TVR-treated, 2 BOC-treated) with some of those patients being hospitalized more than once. Hospitalizations were due to infections (6 patients), rash (5 patients), hepatic decompensation (4 patients), and anemia (2 patients), respectively. Further hospital admissions occurred due to a non-ST elevated myocardial infarction (hemoglobin level in this patient at that time point 12.9 g/dl), severe headaches, exsiccosis/ diarrhea, deep venous thrombosis, diagnosis and treatment of lung cancer, and performance of LTx, respectively.

During follow-up, four patients developed hepatocellular carcinoma. All of those patients had discontinued the antiviral treatment prematurely due to side effects (n = 3) or virological failure (n = 1) before the diagnosis of HCC was made. Of those, one patient underwent LTx later on. One patient was diagnosed with a malignant brain tumor one year after successful conclusion of BOC-triple therapy, and one patient died from an accident during follow-up.