TRIZAL study: switching from successful HAART to TrizivirTM (abacavir-lamivudine-zidovudine combination tablet): 48 weeks efficacy, safety and adherence results
HIV Medicine. Volume 4 Issue 2 Page 79, April 2003. C Katlama et al. Hopital Pitie-Salpetriere, Paris, France.
Summary: Switching to Trizivir from HAART led to sustained virologic suppression and maintained immunological status: the proportion of subjects with a plasma viral load of <50 copies/mL was also similar in both groups: 75% (79 of 106) and 69% (71 of 103) in the simplified and continued arms, respectively, by intent-to-treat (missing equals failure) analysis; Cd4 increases were similar. At week 48, the proportion of treatment failures in Trizivir arm (23/106, 22%) was noninferior to that observed in continued arm (23/103, 22%) with a treatment difference stratified by prior ART of 1.2%[-10.1; 12.5]. Twenty and 21% of subjects discontinued randomized treatment before week 48 in the Trizivir and continued therapy groups, respectively. The Kaplan Meier estimate of the time to treatment failure was similar between the simplified and continued arms. In the simplified arm, the most frequent drug-related adverse events were malaise and fatigue (10%), hypersensitivity reactions (10%), nausea and vomiting (6%), and diarrhoea (5%). In the continued arm, the most frequent drug-related adverse events were clinical manifestations of lipodystrophy syndrome (9%) and diarrhoea (7%). The most common treatment-limiting adverse event leading to permanent discontinuation of at least one
study drug was possible ABC hypersensitivity reaction in the simplified arm, and possible manifestations of lipodystrophy syndrome and gastrointestinal events in the continued arm. The incidence of premature discontinuation of study drug due to adverse events was similar in both treatment arms (15% and 16%, respectively) with an early onset for those that occurred in the Trizivir group.
Observational cohort studies have reported that following initial declines in plasma HIV-1 RNA, virologic failure has been observed in a substantial proportion of subjects with approximately 25% of subjects discontinuing their initial protease-inhibitor (PI)-containing regimens within 8-12 months due to virologic failure, toxicity, or adherence issues. These concerns, plus the desire for more conveniently dosed and better-tolerated regimens, have necessitated a search for simplified therapies.
The purpose of this study, AZL30002, was to evaluate the antiviral efficacy, safety, and adherence in subjects who switched to Trizivir following long-term HIV-1 RNA suppression on triple-combination antiretroviral therapy. Unlike previous maintenance studies, the subjects in this study received the fixed-dose combination of Trizivir as a single tablet containing lamivudine (3TC) 150 mg, zidovudine (ZDV) 300 mg and ABC 300 mg twice daily.
This study was designed as a randomized open-label, multicentre, 48-week comparative study of the safety and efficacy of Trizivir vs continued triple-combination antiretroviral therapy.
Subjects were recruited from 47 centres in nine European countries between November 1999 and March 2001. Adults who were HIV-1 seropositive, 18 years of age, with CD4 lymphocyte counts 100 cells/L, and had been receiving two nucleoside reverse transcriptases (NRTIs) plus a protease inhibitor (PI) or three NRTIs, or two NRTIs plus a nonnucleoside reverse transcriptase inhibitor (NNRTI) for at least 6 months, with a plasma viral load of < 400 HIV-1 RNA copies/mL since initiation of therapy (subjects previously treated by monotherapy were not included) and a plasma viral load of < 50 copies/mL within 14 days of study entry were eligible to enroll.
The primary outcome measure was treatment failure, defined as either virological failure (viral load of 0.5400 copies/mL on two consecutive occasions), or premature discontinuation of randomized study treatment. Secondary outcomes included change in CD4 cell count from baseline, proportion of patients with a plasma viral load of < 50 copies/mL, development of adverse events, changes in metabolic parameters, clinical lipodystrophy
manifestations and treatment adherence.
Adherence to the randomized treatments was assessed using the Patient Medication Adherence Questionnaire (PMAQ7) instrument which was administered at each visit. The PMAQ7 is a self-administered questionnaire that
provides a quantitative assessment of adherence. The quantitative portion evaluates the frequency of the prescribed daily dose and the actual drug intake by the subject over the last 7 days, leading to an index of adherence. A patient was said to be completely adherent if he/she reported a 100% intake of antiretrovirals over the last 7 days. The instrument also provides a qualitative evaluation of the extent to which patient, disease and treatment regimen characteristics serve as barriers to or motivators for adherence. Patients answer 24 questions using a 5-point Likert scale from 1 ('definitely true') to 5 ('definitely false'). Responses to questions were re-coded, summed, and transformed to create a total score ranging from 0 to 100, where a high score reflects few adherence problems.
All analyses were performed for the 'intent-to-treat' (ITT) population, including randomized subjects who received at least one dose of study drug. In the ITT analysis, patients were considered as treatment failures if they experienced either a virological failure (viral load of 400 copies/mL on two consecutive occasions) or a premature discontinuation of randomized study treatment. Any missing data were also considered a failure in this analysis. Furthermore, an 'as treated' (AT) analysis was performed on patients without major protocol violations and including only data from patients continuing randomized treatment. This study sought to establish that the combination tablet regimen was not inferior to the continuation regimen with respect to proportion of treatment failures. Based on an expected 20% treatment failure rate, 100 subjects per treatment group was deemed necessary to provide 85% power to conclude noninferiority if the lower limit of 95% CI around the difference in treatment failure rate between the two arms was greater than 15% (i.e. Trizivir is no worse than continuation regimen). The CI was calculated using the Mantel-Haenszel statistics controlling for the previous antiretroviral regimen.
The time to treatment failure was determined using the Kaplan-Meier product-limit survival method. Time to failure was defined by the number of days from the first dose of study drug until the first time the plasma viral load was 400 copies/mL, confirmed twice at least 1 week apart or the first time of premature discontinuation of randomized treatment. The time to treatment failure was compared between treatment groups using the log-rank test. The hazard ratio between the two treatment groups was given, with its associated 95% CI. The proportions of patients with viral loads of < 400 and < 50 copies/mL was also given by visit throughout the study.
The analysis of adherence was performed on an ITT population. Quantitative, qualitative scores and subscore (see below) were computed. Since patients had virological suppression at inclusion and had not changed their antiretroviral regimen over the last 6 months, patients in both arms were anticipated to be highly adherent to therapy both at inclusion and at 48 weeks of treatment. For this reason, more attention was focused on the qualitative part of the PMAQ7. Total scores across all 24 items were determined at baseline, week 24 and at week 48. In addition, a panel of 14 French, expert HIV physicians were asked to determine which of the 24 items would be predominantly influenced by a prescribed drug regimen rather than patient-specific or disease-specific factors. The panel identified eight items that met this criteria. Scores for patient responses to these eight items were summed and transformed to create a subscale of the PMAQ7 having a possible score range of 0-100. Sub-scale scores were calculated for both treatment groups at baseline and week 48.
A total of 250 subjects were screened and 219 enrolled in the study; 111 subjects were randomized to receive Trizivir and 108 to remain on their current regimen. Ten randomized subjects, five in each treatment group
received no study drug and were excluded from the ITT population. In the AT population 10 subjects were excluded for protocol deviations (five in each treatment group); the main reason concerned subjects with monotherapy in first line or subjects having received initial triple therapy with regimens other than those specified (two NRTIs plus a PI, two NRTIs plus a NNRTI or three NRTIs). It was planned in the protocol that subjects continue their scheduled
assessments even if they change their randomized group; thus, 97 (92%) in the Trizivir group completed 48 weeks with 85(80%) on randomized treatment and 12 (11%) on nonrandomized treatment. On the same line, in the continued arm, 100 (97%) subjects completed 48 weeks with 81 (79%) subjects on randomized treatment and 19 (18%) on nonrandomized treatment. Baseline characteristics were well balanced between the two treatment groups. Pre-study antiretroviral exposure was similar in both treatment groups. The most frequent prestudy regimen was two NRTIs plus a PI (131/209, 63%) with combivir plus indinavir (IDV) as the most common combination (28/131, 21%). The duration of the last antiretroviral regimen before study entry was slightly higher in the simplified group with a median duration of exposure of 27 months (range, 8-129) compared to 24 months (range, 6-44) in the continuation group. 62% in the simplified and Continuated arms received in their last ART regimen two NRTIs plus one PI; 19% in both arms received two NRTIs plus one NNRTI; interestingly, 17% received 3 NRTIs; 0-2% received one NRTI plus 1 NNRTI plus one PI. Some study patients had prior experience with double NRTI therapy prior to receiving their last regimen of HAART or triple NRTIs.
Twenty and 21% of subjects discontinued randomized treatment before week 48 in the Trizivir and continued therapy groups, respectively. The most common reason for premature discontinuation of study drug was the occurrence of adverse events in similar proportions (16 [15%] and 16 [16%] of subjects in Trizivir and continued arm, respectively).
At week 48, the proportion of subjects with treatment failure (defined viral load of 400 copies/mL or premature discontinuation of study drug) was: 22% (23/106) and 22% (23/103) in the simplified and continued arms, respectively (treatment difference: 1.2%; adjusted 95% CI: 10.1% to 12.5%), supporting the hypothesis of noninferiority of Trizivir. The proportion of subjects with a plasma viral load of <50 copies/mL was also similar in both groups: 75% (79 of 106) and 69% (71 of 103) in the simplified and continued arms, respectively, by
intent-to-treat (missing equals failure) analysis. In the AT population, 94% (74/79) and 90% (66/73) had a viral load of <50 copies/mL, respectively.
A total of six subjects met protocol-defined virological failure (five and one subject(s) in the simplified and continued arms, respectively). Of the five subjects in the simplified arm, two remained on the same triple therapy
with Trizivir alone, one added efavirenz (EFV) to the Trizivir, and two switched to a new regimen. Four of five subjects subsequently experienced a reduction in plasma viral load to 400 copies/mL, with three out of five achieving a viral load of 50 copies/mL. Three of the five subjects experiencing virological failure in the simplified group had received dual NRTI therapy prior to initiating triple therapy. One of the subjects randomized to simplified arm experienced virological failure the day of Trizivir initiation and was, therefore, considered as a virological failure, only one subject had detectable RT genotypic mutations at virological rebound partially associated with prior history of dual ART with ZDV/didanosine (ddI). RT mutations (M41L, M184V, L210W, T215Y) were detected in RT DNA at the time of virological rebound, treatment was switched to the previous triple therapy. This patient returned to consecutive viral load measurements of < 50 copies/mL by week 48. The one subject experiencing a virological failure in the continuation group remained on his current therapy with a stable plasma viral load of between 77 and 1200 copies/mL throughout the study.
The Kaplan Meier estimate of the time to treatment failure was similar between the simplified and continued arms. The associated hazard ratio was 0.97 [95% CI: 0.54-1.72]. Most failures occurred prior to week 4 in the simplified arm, whereas they generally occurred after week 12 in the continued arm. These early failures can be explained by the occurrence of the hypersensitivity reaction in the first 2 weeks in the Trizivir arm.
M edian CD4 cell counts increased slightly from baseline through week 48 in both treatment groups. At week 48, the median increase in CD4 cell count was not significantly different between groups: 26 cells/L in both groups in the ITT population.
Mean self-reported adherence at baseline was > 98% in both treatment arms and remained this way through week 48. The total score for the 24 qualitative items did not show any statistical difference between the Trizivir and continued treatment arms at baseline (mean: 77.8 vs 76.3) or at week 48 (mean: 82.0 vs 80.6). Baseline scores between treatment groups were similar for the subscale composed of the eight items related to treatment regimen
(P = 0.58). However, at week 48, the subscale score was higher for the Trizivir arm than for the continued treatment arm (87.9 vs 81.3, P = 0.001).
Most of the adverse events (95%) were transient, mild to moderate in severity. A total of 37 subjects (35%) and 30 subjects (29%) experienced at least one drug-related adverse event in the simplified and continued arms, respectively. In the simplified arm, the most frequent drug-related adverse events were malaise and fatigue (10%), hypersensitivity reactions (10%), nausea and vomiting (6%), and diarrhoea (5%). In the continued arm, the most frequent drug-related adverse events were clinical manifestations of lipodystrophy syndrome (9%) and diarrhoea (7%). The most common treatment-limiting adverse event leading to permanent discontinuation of at least one
study drug was possible ABC hypersensitivity reaction in the simplified arm, and possible manifestations of lipodystrophy syndrome and gastrointestinal events in the continued arm. The incidence of premature discontinuation of study drug due to adverse events was similar in both treatment arms (15% and 16%,
respectively) with an early onset for those that occurred in the Trizivir group.
In the simplified arm, 11 subjects (10%) experienced a possible ABC hypersensitivity reaction. Time of onset after initiation of study medication was between 1 and 10 days (10 subjects) and 6 weeks (one subject). None of the cases were life threatening or required hospitalization. ABC was discontinued in all but one case within 8 days after the first sign/symptom onset. All cases resolved rapidly after treatment discontinuation. Other treatment-limiting adverse events in the simplified arm were haematologic (two), gastrointestinal (one), malaise/fatigue (one) and hyperamylasemia (one). In the continued arm, treatment-limiting adverse events were possible symptoms of lipodystrophy (seven), gastrointestinal (three), neuropathies (two), renal colic on IDV (one), grade 4 neutropenia (one), depression (one), and miscarriage (one). There were no major differences between treatment groups in the reporting or the appearance of symptoms that may be associated with mitochondrial toxicity. This was confirmed by measurements of lactates and calculation of the anion gap, which did not differ between groups. However, two cases of drug-related neuropathies were reported in subjects who received ddI-stavudine (d4T)-EFV and d4T-3TC-EFV in the continued arm.
The incidence of grade 3 or 4 emergent clinical chemistry abnormalities was similar (12% and 14% in simplified and continued arms, respectively) in both treatment groups. The incidence of grade 3 or 4 emergent haematological
abnormalities was similar (four subjects in each arm, i.e. 4%). At week 48, the median decrease from baseline in cholesterol total was greater in the simplified regimen (-0.80 mmol/L) than in the continued group (-0.44 mmol/L) (P < 0.001) (Fig. 3). For triglycerides, the median change from baseline was 0.17 mmol/L and 0.01 mmol/L in the simplified and continuation groups, respectively (P = 0.006).
At baseline, 42 (40%) subjects in the simplified group and 51 (50%) subjects in the continuation group presented at least one clinical lipodystrophy manifestation (not significantly different). By week 48, only 27 subjects (28%) in the simplified group had at least one clinical lipodystrophy manifestation compared with 42 subjects (42%) in the continued arm (P < 0.03). Furthermore, at week 48, the median number of lipodystrophy symptoms per subject
was higher in the continued arm (n = 4) compared to Trizivir arm (n = 2) (P = 0.016). No significant difference in weight change from baseline was noted between the two treatment groups at week 48.
One subject in the continued arm progressed from CDC grade B to C.
Discussion by authors
Switching to Trizivir from HAART led to sustained virologic suppression and maintained immunological status. These immuno-virological responses were accompanied by significantly greater reductions in total cholesterol and triglyceride levels in the simplified regimen than in the continued arm. The incidence of treatment-limiting adverse events was similar in both groups and
was the major factor leading to discontinuation of randomized treatment; this could be one of the major drawbacks of an open-label design of a large comparative trial in contrast to a double-blind scheme. The adherence findings from this study are consistent with those previously reported in studies of ABC-containing nucleoside regimens and other simplified regimens of similar study
design. They also reflect the fact that the subject population of this trial was highly compliant at inclusion and remained so throughout the 48 weeks of the trial. However, in terms of convenience, dosage and taste, a fixed-combined tritherapy, such as TrizivirTM, conveys a better adherence profile for HIV patients.
This study selected for highly adherent patients, as evidenced by the fact that both treatment arms were highly adherent at inclusion and throughout the 48 weeks of the trial. Unfortunately, this provided little opportunity to observe any potential adherence improvement in patients who switched to TrizivirTM. The study design did, however, allow a reasonable opportunity to assess whether or not TrizivirTM was more convenient to take, a factor that has been shown to be associated with adherence. The qualitative portion of the PMAQ7 asks patients about many factors that may serve as barriers to their adherence. Some of these factors, such as healthcare-provider relationship and social
support, have little to do with treatment regimens and would only serve to mask potential dosing-convenience differences between treatments if they were included in a summary score. It was, therefore, not surprising that total scores across all 24 PMAQ7 items did not differ at week 48 for the two treatment groups. However, when we evaluated the subscore for the eight items judged by physicians to be regimen related, we found that Trizivir had higher scores than continued treatment, suggesting that fixed-dose combination therapy with Trizivir may facilitate adherence more than the other therapies evaluated in this study. This finding is consistent with a previous study that showed simpler regimens were viewed by patients to facilitate adherence.
The triple nucleoside regimen was associated with significant reductions from baseline in cholesterol and triglyceride levels, consistent with previous results from similar studies. Although there was no objective, radiologic measurement of lipodystrophy and the study was not blinded. The overall estimation of clinical manifestation of lipodystrophy suggests a slight improvement in the
Trizivir arm, based on improvement of metabolic disorders and reductions of lipodystrophic symptoms. Longer follow-up may be needed to confirm the demonstrated benefits in lipodystrophic symptoms of switching to Trizivir.
The incidence of possible ABC-related hypersensitivity reactions in this study (10%) is higher than that reported in most other studies (3%5%). This may be due to particular diligence on the part of the investigators in reporting presumed cases resulting from much communication about the possible signs and symptoms of hypersensitivity at the time the trial was conducted. Of note, all cases resolved following withdrawal of medication and none required hospitalization or were life threatening.
In conclusion, this study showed that a maintenance regimen with a triple nucleoside combination tablet, with Trizivir, is an alternative treatment option for physicians and subjects, concerned with adherence challenges and toxicities observed with PI or NNRTI containing regimens. The Trizivir triple combination tablet offers a number of clinically important advantages, including sustained virological control with a convenient, dosing regimen of one tablet twice daily, low frequency of lipid abnormalities, and the potential to spare other drug classes for future therapy.