Induction therapy with trizivir plus efavirenz or lopinavir/ritonavir followed by trizivir alone in naive HIV-1-infected adults

Background: Induction-maintenance strategies were associated with a low response rate. We compared the virological response with two different induction regimens with trizivir plus efavirenz or lopinavir/ritonavir.

Methods: A randomized, multicentre, open-label clinical trial with 209 antiretroviral-naive HIV-infected patients assigned to trizivir plus either efavirenz or lopinavir/ritonavir during 24-36 weeks. Patients reaching undetectable plasma viral loads during induction entered a 48-week maintenance on trizivir alone. The primary endpoint was the proportion of patients without treatment failure at 72 weeks using an intent to treat (ITT) analysis (switching equals failure).

Results: Patients were randomly assigned (efavirenz 104; lopinavir/ritonavir 105), and 114 (55%) entered the maintenance phase (efavirenz 54; lopinavir/ritonavir 60). Baseline characteristics were balanced between groups. The response rate at 72 weeks was 31 and 43% (ITT analysis, P = 0.076) and 63 and 75% (on-treatment analysis, P = 0.172) in the efavirenz and lopinavir/ritonavir arms, respectively. Virological failure occurred in 27 patients: six during induction (efavirenz, three; lopinavir/ritonavir, three; P = 1.0) and 21 during maintenance (efavirenz, 14; lopinavir/ritonavir, seven; P = 0.057). Thirty-four patients in the efavirenz arm switched treatment because of adverse events compared with 25 in the lopinavir/ritonavir arm (P = 0.17).

Conclusion: Trizivir plus either efavirenz or lopinavir/ritonavir followed by maintenance with trizivir achieved a low but similar response at 72 weeks, with a high incidence of adverse events leading to drug discontinuation during the induction phase in both arms. The study showed a trend towards an increased virological failure rate in the efavirenz arm during the maintenance phase.

Introduction

The introduction of HAART has resulted in dramatic reductions in HIV-1-associated mortality and morbidity ^[1,2]. Current recommendations support initiating antiretroviral therapy (ART) with two nucleoside reverse transcriptase inhibitors plus a non-nucleoside reverse transcriptase inhibitor or a ritonavir-boosted protease inhibitor. Although highly effective, HAART has been associated with significant short and long-term toxicities, including hyperlipidemia, insulin resistance, and fat redistribution syndromes, which may be associated with accelerated atherosclerotic cardiovascular disease ^[3-6]. Given these and other toxicities, interest has arisen in reducing prolonged exposure to antiretroviral agents. The discontinuation of ART in patients successfully treated for prolonged periods has, however, been associated with rapid virological rebound and decreasing CD4 cell counts ^[7,8], and CD4-driven intermittent ART has been associated with an increased incidence of opportunistic and non-opportunistic adverse clinical events ^[9].

Earlier studies have tested the concept of an induction-maintenance strategy with a suboptimal treatment response ^[10-12]. In those trials, patients were treated with triple (two studies) or quadruple drug therapy for 12-26 weeks, followed by simplification to monotherapy or dual combination therapy. Rebound rates ranged from 22 to 31%, and were related to poor adherence, poor pharmacokinetic profiles, and baseline viral resistance. Despite these initial failures, interest in the induction-maintenance concept has persisted. Markowitz et al. ^[13] assessed this concept, using an induction phase with abacavir/lamivudine/zidovudine (trizivir) plus efavirenz for 48 weeks and an additional 48 weeks of maintenance therapy with trizivir alone in ART-naive patients, showing that the maintenance phase improved adherence, reduced fasting lipids and ART-associated adverse events while maintaining virological control and the immunological response. More recently, other induction-maintenance approaches have been reported that showed encouraging results ^[14-18].

Here, we describe an induction-maintenance strategy in which patients were randomly assigned to an induction regimen that included a fixed-dose combination of abacavir, lamivudine and zidovudine (trizivir) plus either efavirenz or lopinavir boosted with ritonavir (lopinavir/ritonavir) for 24-36 weeks, followed by maintenance with abacavir/lamivudine/zidovudine in those who reached an undetectable plasma viral load. We hypothesized that after 24-36 weeks of treatment, the removal of one agent would result in reduced toxicity while maintaining the antiviral and immunological benefits. Trizivir should be enough for maintenance therapy in suppressed patients without previous archive mutations in the retrotranscriptase gene ^[19,20]. Moreover, we hypothesized that the utilization of efavirenz or lopinavir/ritonavir in the induction phase would be similar.

Methods

Study population

HIV-1-infected adults were eligible if they were antiretroviral naive and had an HIV-1-RNA level above 10 000 copies/ml. Exclusion criteria included the current use of experimental therapies, active opportunistic disease requiring parenteral treatment, and any contraindication to study drugs. Female patients were excluded if they were pregnant or breastfeeding.

Study design

Trizefal was a randomized, open-label, 72-week study conducted at 18 clinical sites in Spain. The duration of the induction phase of the study was 24-36 weeks followed by a maintenance phase of 48 weeks for those reaching undetectable viral loads at the end of the induction phase. Enrollment took place between April 2003 and July 2004.

Patients were centrally randomly selected (1: 1 ratio) to abacavir/lamivudine/zidovudine with either efavirenz or lopinavir/ritonavir. Randomization was stratified based on an HIV-1-RNA level at study entry (baseline) of less than 100 000 copies/ml or 100 000 copies/ml or greater. During the 24-36-week induction period, all patients were treated with the fixed-dose combination of abacavir (300 mg)/lamivudine (150 mg)/zidovudine (300 mg) (trizivir; GlaxoSmithKline, Greenford, Middlesex, UK) twice a day and with either efavirenz (600 mg, Sustiva; Bristol-Myers Squibb, New York, USA) once a day or lopinavir/ritonavir (Kaletra, Abbott, Abbott Park, Illinois, USA) three capsules × 133/33 mg twice a day. Patients with HIV-1-RNA levels less than 50 copies/ml at week 24 (or at week 36 if not reached at week 24) were eligible for simplification into the 48-week maintenance phase with only abacavir/lamivudine/zidovudine.

Screening evaluation included a clinical assessment and laboratory evaluations: plasma HIV-1 RNA, T-lymphocyte subsets, haematology, clinical chemistries, and the collection of a plasma sample later analysed for viral resistance mutations. On-study evaluation included clinical visits at baseline and weeks 4, 12, 24, 36, 48, 64 and 72. HIV-1 RNA was assessed at each visit using the local HIV-1 ultrasensitive assay. Clinical chemistries, haematology profiles, and T-cell subsets were performed at all visits. Fasting lipid profiles were also measured. Samples for possible viral resistance testing were collected at baseline and at confirmation of virological failure. Safety was assessed through the reporting of adverse clinical events and abnormal laboratory values. The severity of adverse events was assessed by the investigator using the Division of AIDS toxicity grading scale ^[21]. No substitutions were allowed for abacavir/lamivudine/zidovudine.

Adherence to therapy was not measured. Viral resistance testing at baseline and at the time of virological failure was attempted for all patients who met virological failure criteria.

The protocol was reviewed and approved by the ethics committees of each of the participating sites and the Spanish Medicines Agency. All patients provided written informed consent before entering the study.

Analyses

The main study endpoint was the proportion of patients without treatment failure at 72 weeks using an intent to treat (ITT) analysis (switches and missing equals failure). Treatment failure was defined as virological failure, loss to follow-up, discontinuation of study medication because of toxicity or any other reason, clinical disease progression (new Centers for Disease Control and Prevention grade C events), or death. Virological failure was defined as failure to suppress the HIV-1-RNA level to less than 50 copies/ml by week 24 or 36 or a confirmed HIV-1-RNA level greater than 200 copies/ml after suppression to less than 50 copies/ml at any timepoint. Time to treatment failure or virological failure and proportions of patients with virological failure at week 72 while on study medication (on-treatment analysis) were secondary endpoints. Other secondary objectives included immunology, safety and tolerability, and the development of viral resistance.

A sample size of 220 patients provided 80% power to detect the superiority of any of the study arms based on proportions of patients with an HIV-1-RNA level less than 50 copies/ml at week 72 with a type 1 error rate of 0.05, assuming a success rate of 80% in the most effective treatment group and a minimum detectable difference of 20 percentage points.

Qualitative variables were described using frequency and percentage and compared between groups by means of Fisher's exact test. The proportion of patients with virological failure and treatment failure was expressed by percentage and 95% confidence intervals (CI). Quantitative characteristics were summarized using median value and interquartile ranges (IQR), and the comparison between groups was realized using the Wilcoxon rank sum test.

Time to failure was estimated using the Kaplan-Meier method and the Log-rank test was used to compare survivor functions between groups. All tests were two-sided with a confidence level of 95%. Statistical analyses were performed using STATA (release 9; Stata Corporation, College Station, Texas, USA).

Results

Patient disposition

A total of 220 patients were randomly assigned, but only 209 were considered eligible and included in the analysis. The remaining 11 patients were excluded because 10 withdrew consent and never started treatment and one was randomly assigned twice.

Patients were predominantly male, with a median age of 39 years (range 22-74), a median baseline HIV-1-RNA level of 159 500 copies/ml (IQR 42 600; 440 000) and a median CD4 cell count of 202 cells/μl (IQR 104; 297). Table 1 shows the demographic characteristics of the patients. Forty-five per cent (95/209 patients) did not reach the maintenance phase, primarily because of adverse events (n = 58, 28%), and loss to follow-up (n = 14, 7%) or because they did not achieve virological suppression (n = 6, 3%).

Table 1

A total of 114 patients (54 in the abacavir/lamivudine/zidovudine plus efavirenz arm and 60 in the abacavir/lamivudine/zidovudine plus lopinavir/ritonavir arm) continued to maintenance therapy. Most patients (81, 69%) entering the maintenance phase completed the 48-week follow-up. More subjects in the efavirenz group, compared with the lopinavir/ritonavir group, discontinued the study during this phase because of virological failure (14 versus seven). Six patients in the efavirenz arm were lost to follow-up during the maintenance phase versus five in the lopinavir/ritonavir arm and one withdrew study medication in the efavirenz arm because of an adverse event (see Fig. 1).

Fig. 1

Virological and immunological response

At the end of the induction phase (24-36 weeks) the proportion of patients with virological failure and treatment failure were 3 versus 3% and 48 versus 42% in the efavirenz and lopinavir/ritonavir arms, respectively. At the end of the study and using an ITT approach the proportion of patients with virological failure and treatment failure (main study endpoint) were: 16% (95% CI 10-25%) versus 10% (95% CI 5-17%) and 69% (95% CI 60-78%) versus 57% (95% CI 47-67%) in the efavirenz and lopinavir/ritonavir arms, respectively (P = 0.076). The on-treatment analyses showed 63% (95% CI 49-75%) and 75% (95% CI 60-85%) of patients remaining in the study without treatment failure in the efavirenz and lopinavir/ritonavir arms, respectively (P = 0172; see Fig. 2). Time to treatment failure analysis showed a trend towards a better response in the lopinavir/ritonavir arm compared with the efavirenz arm but without statistically significant differences (see Fig. 3). There was a non-statistically significant difference in virological failures among patients with more or less than 100 000 HIV-1 viral load at baseline.

Fig. 2

Fig. 3

Patients remaining on therapy during the first 48 weeks experienced a median increase of 169 (IQR 98, 271) and 252 (IQR 136, 369) CD4 T cells/μl (P = 0.034) and the median increase for those remaining on therapy for the whole of the study experienced a median increase of 247 (IQR 119, 357) and 274 (IQR 122, 372) CD4 T cells/μl, respectively, in the efavirenz and lopinavir/ritonavir groups (P = 0.7; see Fig. 4).

Fig. 4

Adverse events

Table 2 shows the incidence of side effects in each arm and those leading to discontinuation of the study drug.

Table 2

Adverse events were very common during the study period, especially during the induction phase. In the efavirenz arm 34 patients discontinued study drug medication as a result of side effects and all but one withdrew from therapy during the induction phase compared with 25 patients in the lopinavir/ritonavir arm, all of them in the induction phase. In the efavirenz arm dermatological side effects (rash and hypersensitivity reaction) and neuropsychiatric side effects (sleep disturbances and vivid dreams) were significantly more frequent than in the lopinavir/ritonavir arm. Conversely, in the lopinavir/ritonavir arm the incidence of gastrointestinal disorders (nausea and diarrhoea) were more frequent than in the efavirenz arm.

At the end of the induction phase, the median increase in total cholesterol, high-density lipoprotein cholesterol and triglyceride levels in the efavirenz and lopinavir/ritonavir group in mg/dl were: 20 and 30.5, 9 and 8, 1 and 65, respectively, with a statistically significant difference in triglycerides (P < 0.001). At week 72, median changes from week 24 in total cholesterol, high-density lipoprotein cholesterol and triglyceride levels in mg/dl were -26 and -25, -7 and -2.7, -11 and -62 in efavirenz and lopinavir/ritonavir, respectively, with a statistically significant difference in triglycerides (P < 0.001).

Treatment-emergent viral resistance

At baseline no primary mutations were detected in 10 out of the 27 patients who subsequently developed virological failure and in 18 matched controls who did not develop virological failure. Baseline samples were not available for the 17 remaining patients who developed virological failure. Table 3 shows the treatment-emergent viral resistance.

Table 3

Discussion

The results of this study question again the convenience of this induction-maintenance approach. Quadruple therapy provided good virological and immunological responses in patients treated for 24-36 weeks. Only a fraction of those randomly assigned entered the maintenance period, however, because of an increased frequency of adverse events mainly in the induction phase. Also, a high proportion of those successfully reaching simplification had a virological failure.

We chose a 24-36-week induction phase for several reasons. Given the expectation that patients with high baseline HIV-1-RNA levels would be enrolled, and in fact it was in 60% of patients, we anticipated that no shorter induction period would be sufficient. The induction-maintenance concept is partly predicated on reducing the pool of randomly generated preexisting resistant variants during initial treatment. Given that the half-life of the second phase of HIV-1-RNA decay after HAART initiation is, on average, 2-4 weeks ^[22,23], we hypothesized that treatment of approximately six half-lives, or approximately 24 weeks, would result in a substantially reduced residual viral burden at the time of treatment simplification. In addition, we believe that virological success in the induction of this trial was achieved by using an effective quadruple combination therapy. In these terms of virological success, however, the induction phase of 24 weeks is probably not enough particularly in the efavirenz arm in which virologial failure was high (14 out of 54 patients; 26%) compared with the lopinavir/ritonavir arm in which seven out of 60 patients (11%) developed virological failure.

The large number of treatment discontinuations during the induction phase is problematical but not new. In the trial by Markowitz et al. ^[13] 37% of patients withdrew from therapy during the 48 weeks induction phase of zidovudine/lamivudine/abacavir plus efavirenz treatment, also mainly as a result of adverse events. In another trial comparing efavirenz-based HAART with triple-nucleoside therapy for the initial treatment of HIV-1, only 7% of patients discontinued therapy by week 32 ^[24], despite one of the treatment groups being abacavir/lamivudine/abacavir plus efavirenz. More recently, the Forte study ^[14] showed that starting ART with four drugs followed by a maintenance therapy with three agents improved virological outcomes compared with a three-drug regimen without significantly increasing toxicity. Also, it has been proved that after a long period of undetectability, antiretroviral strategies including nucleosides plus boosted protease inhibitors such as lopinavir/ritonavir can be simplified to lopinavir/ritonavir monotherapy with encouraging results in terms of efficacy and tolerance ^[15-18]. In all three studies, however, the rebound levels seen in HIV viral load require further investigation. Overall, simplifying therapy and making it easier in terms of adherence is a motivation for continuing to study boosted protease inhibitor monotherapy, as well as limiting the potential for side effects, containing costs and preserving therapy choices.

Our high discontinuation rate is difficult to explain. The specific four-drug regimens in this trial failed to improve the initial response rate to antiviral therapy, in contrast to established three-drug regimens ^[24-26]. Response rates in all strata of baseline HIV-1-RNA levels can be considered high and demonstrate the potency of both four-drug regimens with a low rate of virological failure during induction. This good virological response in the induction phase, however, does not accord with a good treatment success rate as shown in the ITT analyses as a result of a high treatment drug discontinuation related to the development of adverse events. This may not be the case with future better tolerated agents such as raltegravir or maraviroc.

In assessing the results of this study, we must ask what benefits treatment simplification provided: in the short term, a reduction in treatment-related adverse events and also an improvement in fasting lipid profiles. Furthermore, in the face of the emergence of drug-resistant virus, treatment options in the three classes (nucleoside reverse transcriptase inhibitors, non-nucleoside reverse transcriptase inhibitors, protease inhibitors) remained available for use in most of the patients who experienced virological failure. It is important to note that during maintenance virological failure and treatment-emergence drug resistance was more common in the efavirenz group than in the lopinavir/ritonavir therapy group, although the difference did not reach statistical significance.

This induction-maintenance approach is associated with a high rate of antiretroviral discontinuation as a result of adverse events. Also, the rate of virological failure during the maintenance phase can be considered high, especially in the efavirenz arm. Therefore, we believe that currently it can not be considered as a valid alternative for the management of treatment initiation in HIV-infected patients.

Acknowledgement

Other staff at the Clinical Trials Unit (Hospital clinic) who conducted and analysed the study include: A. Cruceta, H. Agell, E. Aragon, S. Varea.

The other TRIFEZAL investigators were: Agathe León, Joan A. Arnaiz, José L. Blanco, Esteban Martínez, José M. Miró, Elisa De Lazzari, Cristina Gil, Tomás Pumarola, Hospital Clínic IDIBAPS, University of Barcelona, Barcelona; Melcior Riera, Hospital Son Dureta, Palma de Mallorca, Mallorca; María Saumoy, Sergi Veloso, Consuelo Viladés, Hospital Joan XXIII-Universitat Rovira i Virgili, Tarragona; Mar Masiá, Sergio Padilla, Hospital Universitario de Elche, Alicante; David Dalmau, Mutua de Terrassa, Barcelona; Esteve Ribera, Hospital Vall Hebrón, Barcelona; Antonio Ocampo, Hospital Xeral-Cíes, Vigo; Miguel A. Muniaín, Hospital Virgen Macarena, Sevilla; Carlos Alonso-Villaverde, Hospital Sant Joan de Reus, Tarragona; Vicente Estrada, Hospital Clínico, Madrid; José R. Blanco, Hospital General de la Rioja, Logroño; Josep Cucurull, Hospital de Figueres, Girona; Josep M. Llibre, Hospital de Calella, Barcelona; Jorge Carmena, Hospital Peset, Valencia; and María J. Galindo, Hospital Clínico, Valencia.

The authors are indebted to the participants in the study.

Presented in part at the 8th International Congress on Drug Therapy in HIV Infection, Glasgow, November 2006.

Clinical trial registry number: ISRCTN41611080.

Sponsorship: This study was partly supported by an unrestricted grant from GlaxoSmithKline and the Red Temática Cooperativa de Grupos de Investigación en Sida del Fondo de Investigación Sanitaria (FIS), and grant no. ISCIII-RETIC RD06/006 from the Instituto de Salud Carlos III, Madrid (Spain). J.M.M. was the recipient of a research grant from the Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS) and the Conselleria de Salut de la Generalitat de Catalunya, Barcelona (Spain).

Conflicts of interest: None.