No benefit of a structured treatment interruption based on genotypic resistance in heavily pretreated HIV-infected patients

Objective: To evaluate the potential benefits of a tailored antiretroviral treatment interruption with duration based on the observed reversion of resistance mutations.

Methods: In this open single-arm pilot study, 23 patients with multiple treatment failure interrupted therapy and underwent longitudinal genotypic resistance testing. Salvage gigatherapy was started when resistance mutations to at least two antiretroviral drug classes reverted. The primary endpoint was a fall in viral load by > 1 log₁₀ copies/ml after 12 weeks of salvage therapy.

Results: Baseline median viral load was 5.14 log copies/ml and CD4 cell count 43 × 10⁶ cells/l. Genotypic resistance testing showed a median of six, two and nine resistance mutations to nucleoside analogue reverse transcriptase inhibitors, non-nucleoside analogue reverse transcriptase inhibitors and protease inhibitors, respectively; viral strains were susceptible to no more than one drug in 17/23 patients. The median duration of treatment interruption was 24 weeks (range, 12-37), leading to median changes from baseline of + 0.54 log₁₀ copies/ml and -30 × 10⁶ cells/l. At the end of treatment interruption, plasma HIV was susceptible to at least three drugs in 16/23 patients. After 12 weeks of salvage multitherapy, only one patient had a decrease in viral load > 1 log copies/ml. All baseline resistance mutations recurred after treatment resumption. AIDS-defining events occurred in two-thirds of patients during the study period.

Conclusion: In HIV-infected patients with multiple failures and no therapeutic options at baseline, significant reversion of resistance mutations after prolonged treatment interruption failed to restore antiviral efficacy of a salvage regimen and was clinically deleterious.

Introduction

Despite the success of highly active antiretroviral therapy (HAART) ^[1], viral resistance remains problematic ^[2-4]. Severe treatment failure, defined as a CD4 cell count < 200 × 10⁶ cells/l and plasma HIV RNA load > 30 000 copies/ml, affects 5% of patients enrolled in the French hospital HIV database ^[5]. Treatment options for such patients are often extremely limited, owing to broad intraclass drug cross-resistance ^[6] and limited access to new drugs. It has been shown that treatment interruption allowed reemergence of archived wild-type virus that out-competes the less-fit resistant strains ^[7,8]. However, although this may lead to an improved virological response on treatment resumption ^[9], the wild-type virus may have an increased capacity to replicate and, therefore, to delete peripheral blood CD4 cells ^[8]. Conflicting results have been obtained in this clinical setting ^[10-15]. We recently reported that an 8-week treatment interruption in the GigHAART study was beneficial for treatment-experienced patients with multidrug-resistant virus on resuming a salvage regimen ^[12]. As reversion of resistance mutations in plasma increases with the duration of treatment interruption ^[16,17], the present study examined the possible benefits of prolonged treatment interruption based on genotypic resistance testing (GRT), followed by optimized salvage gigatherapy in heavily pretreated HIV-infected patients with multidrug-resistant viruses.

Methods

Study design

The Reverse study was an open, single-arm study in which the length of treatment interruption was based on the observed reversion of resistance mutations; the endpoint was the virological efficacy of a salvage regimen. The failing regimen was discontinued for at least 8 weeks, and GRT was monitored. A salvage regimen was restarted when resistance mutations to at least two antiretroviral classes had reverted, as follows: loss of at least three thymidine-associated resistance mutations (TAM), loss of all mutations associated with resistance to non-nucleoside analogue reverse transcriptase inhibitors (NNRTI), loss of at least three primary mutations associated with resistance to protease inhibitors (PI). Salvage therapy was restarted at any time if an AIDS-related event occurred or after a maximum of 26 weeks even if reversion criteria were not met.

Patients

Patients were eligible if they had plasma HIV-1 RNA > 30 000 copies/ml despite treatment with three or more antiretroviral drugs. Genotypic viral resistance had to include at least three TAM, plus one NNRTI resistance-associated mutation and two primary PI resistance-associated mutations. All the patients gave their written informed consent.

Study treatment

The salvage regimen consisted of six to eight antiretroviral drugs per patient, comprising three to four NRTI, one NNRTI and two or three ritonavir-boosted PI, optimized according to GRT results and patient treatment history.

Clinical and laboratory evaluations

Patients were screened within 4 weeks of enrollment (day 0), then after 2, 4 and 8 weeks of treatment interruption (TI-W2, TI-W4, TI-W8) and every 6 weeks thereafter until resumption of salvage therapy (R-W0). They were then examined every 4 weeks for 24 weeks (to R-W24). Plasma HIV RNA was quantified using the Amplicor Monitor assay 1.5 (Roche Diagnostics, Meylan, France) with a detection limit of 400 copies/ml. GRT (population-based full sequence analysis) was performed on plasma HIV-1 RNA as previously described ^[18,19] and interpreted using the ANRS AC11 algorithm (http://www.HIVfrenchresistance.org/). A genotypic susceptibility score (GSS) ^[20] was calculated as the total number of active antiretroviral drugs among a panel of NRTI, NNRTI and PI according to the post-treatment interruption regimen prescribed.

Antiretroviral drug assays

Steady-state drug concentrations were measured at weeks 2 and 12 of salvage therapy. Minimal effective plasma trough concentrations were those used in the GigHAART study ^[12]. Concentrations were considered adequate at a given time-point if all but one drug trough concentration were above the cut-off at both visits.

Study endpoints

The primary endpoint of the study was the percentage of patients with a decrease in viral load (viral load) of ≥ 1 log₁₀ copies/ml HIV RNA at 12 weeks after treatment resumption (R-W12), relative to the baseline. Secondary endpoints were the changes in CD4 cell count, the evolution of plasma viral load after 12 and 24 weeks of salvage therapy (R-W12, R-W24) and the proportion of patients experiencing AIDS-defining events.

Statistical analysis

The number of subjects enrolled in the study (23) was sufficient to estimate a treatment success rate equal to that observed (60%) in the GigHAART study ^[12], with a precision of ± 20%. Changes in continuous variables were tested for significance using Wilcoxon's paired test.

Results

Characteristics of the patients

A total of 23 patients were enrolled in the study from June 2001 to March 2002. Patients had all received substantial treatment for HIV with a median duration of treatment of 8.5 years (range, 4.5-13.2) and with a median of 13 antiretroviral drugs (range, 11-16). Thirteen patients (57%) had a history of AIDS-defining event(s). At baseline, the median plasma viral load was 5.14 log copies/ml (range, 4.51-5.69), and the median CD4 cell count 43 × 10⁶ cells/l (range, 1-372). The baseline median number of resistance mutations to the three classes of drug was 17 (range, 9-28), with a median of six NRTI-associated mutations (range, 4-11) including four TAM (range, 3-6), two NNRTI-associated mutations (range, 1-3) and nine PI-associated mutations (range, 4-14) (Fig. 1). Overall, the mean global GSS for NRTI, NNRTI and PI was 0.8 ± 1.1. Ten patients (44%) had HIV strains susceptible to no drugs and seven patients (30%) had HIV strains susceptible to only one drug.

Fig. 1

Treatment interruption

The median duration of treatment interruption was 24 weeks (range, 12-37).

Plasma viral load increased by a median of 0.53 log copies/ml (range -0.17 to + 1.54) and the CD4 cell count fell by a median of 30 × 10⁶ cells/l (range, -252 to + 6) from baseline (Fig. 2).

Fig. 2

Evolution of resistance mutations is shown in Figure 1. At the end of treatment interruption, GRT showed a significant reversion of resistance mutations to wild-type amino acid residues. Sixteen patients (70%) met reversion criteria at salvage therapy resumption, a shift to a complete wild-type virus occurring in five patients (22%). Overall, the mean total GSS at treatment resumption was 4.9 ± 2.8, with a significant increase of 4 from baseline (P < 0.001). At the end of the treatment interruption, 16 patients (70%) had plasma HIV susceptible to at least three antiretroviral drugs.

Salvage therapy

Salvage therapy comprised a median of seven drugs (range, 4-8) with all 23 patients taking lamivudine; other drugs were abacavir and tenofovir (19), didanosine (18), efavirenz (12), nevirapine (7), delavirdine (1), ritonavir (23), lopinavir (17), indinavir (13), saquinavir (11), and amprenavir (9). Only three patients received fewer than six drugs because of multiple drug intolerance. In 70%, at least three drugs that were prescribed within the salvage regimen were classified as active after treatment interruption according to GRT. All patients had adequate plasma drug concentrations at R-W2 and R-W12 (data not shown).

By intent-to-treat analysis, only one patient achieved a 1 log copies/ml decline in viral load after 12 weeks of salvage therapy. Overall, the median change from baseline was + 0.04 log copies/ml (range, -1.06 to + 0.58) in plasma HIV RNA and -27 × 10⁶ cells/l (range, -189 to + 44) in CD4 cell count. Results were similar at 24 weeks (Fig. 2). All resistance mutations present at baseline recurred 4 weeks after resuming treatment in 13 patients and by 24 weeks in the 10 remaining patients.

Clinical outcome

An AIDS-defining event was seen in 15 (65%) of the 23 patients, including oesophageal candidiasis (eight), lymphoma (one), pulmonary Kaposi's sarcoma (one), Pneumocystis carinii pneumonia (one), multiple progressive leucoencephalopathy (one), disseminated Mycobacterium avium intracellulare infection (one), AIDS-related wasting syndrome (two) and cytomegalovirus colitis (one). Four of these patients experienced more than one AIDS-defining event. All these events occurred between R-W0 and R-W24. There was no significant difference in the CD4 cell count, at baseline (P = 0.548) or at the end of treatment interruption (P = 0.585), between the 15 patients who had AIDS-defining events (baseline: 36 × 10⁶ cells/l; end of treatment interruption: 12 × 10⁶ cells/l) and the eight who did not have AIDS-defining events (baseline: 76 × 10⁶ cells/l; end of treatment interruption: 27 × 10⁶ cells/l).

The seven patients who experienced a major AIDS-related event (all events listed above except oesophageal candidiasis) had a baseline CD4 cell count of 6 × 10⁶ cells/l, compared with 76 × 10⁶ cells/l in the other 16 patients (P = 0.065). However, in these seven patients, the CD4 cell count at the time of AIDS-related event (median 4 × 10⁶ cells/l) was similar to that at baseline (median 6 × 10⁶ cells/l), suggesting that the progression of HIV disease could not be related to a significant decrease in CD4 cell count.

Discussion

This study shows that, in patients with multiple treatment failure, prolonged treatment interruption adds no virological benefit, despite an initial reversion to a nearly wild-type virus. Indeed, these patients harboured multidrug-resistant viruses at baseline, with a GSS of < 1. Treatment interruption led to a significant decrease in the burden of resistance mutations, improving the GSS of plasma HIV from 0.9 to 4.9 active drugs at the end of treatment interruption (P < 0.001). Two-thirds of the patients received a salvage regimen containing at least three drugs to which their plasma HIV dominant variant had become susceptible after treatment interruption, and plasma drug concentrations were adequate. However, despite this improvement in HIV drug sensitivity, a median seven-drug salvage regimen was not able to achieve a 1 log copies/ml reduction in viral load except in one patient. This strongly contrasts with the results observed in the GigHAART study, in which viral load fell by 1.91 log copies/ml in the group with an 8-week treatment interruption ^[12]. It seems that the key issue in the efficacy of a salvage regimen is the number of drugs retaining potential viral activity before any intervention. Indeed, in GigHAART, baseline GSS was 1.4, meaning that, in most patients, one drug was still genotypically active, mainly lopinavir or amprenavir at that time (1998-1999). In our Reverse study, baseline GSS was < 1, with no available new active antiretroviral drugs to which the patients had not been exposed. The improvement in GSS after a median 24-week-treatment interruption in our Reverse study did not restore the antiviral efficacy of a multiple salvage regimen that comprised recycled drugs. Furthermore, the rapid reemergence after treatment resumption of the resistant variants with the same mutational pattern as at baseline suggests that those resistant variants had not disappeared ^[21,22] but had simply been out-competed by wild-type strains during treatment interruption.

We observed a high rate of HIV-related events (15/23 patients, 65%). Of note, these events occurred after the end of treatment interruption. Interestingly, the slope of CD4 cell count change did not alter in a major way in any of the 23 patients in the 6 months prior to inclusion, despite a median viral load > 5 log copies/ml. The median CD4 cell count in the seven patients who experienced major AIDS-defining events did not differ significantly between baseline (6 × 10⁶ cells/l) and the event (4 × 10⁶ cells/l). Furthermore, prolonged treatment interruption led only to a small increase (0.54 log copies/ml) in viral load from baseline. The studies CPCRA 064 ^[10] and ACTG 5086 ^[13], which both involved a 16-week treatment interruption, also showed deleterious clinical consequences despite significant reversion of resistance mutations. Cohort studies have also suggested that treatment interruption had a negative impact on the prognosis for heavily pretreated patients ^[23,24]. In Reverse, prolonged treatment interruption led neither to a significant fall in CD4 cell count in patients experiencing major AIDS-events nor to a significant increase in viral load; therefore, one can hypothesize that progression of HIV disease was most likely related to the greater fitness of the reverted strains after treatment interruption, as previously shown by Deeks et al. ^[8], and/or a different impact on the immune system of wild-type virus compared with the highly mutated virus at baseline. This is supported by the GigHAART results, where an 8-week treatment interruption in very severely immunocompromised patients induced few reversions of resistance mutations and no increase in clinical events.

In conclusion, significant reversion of resistance mutations in plasma HIV strains did not yield any virological benefit for patients with multiple resistant virus at baseline in the absence of any possibility of adding new potent drugs to the regimen. Furthermore, in our experience, prolonged treatment interruption in patients with CD4 cell counts < 200 × 10⁶ cells/l was clinically deleterious. However, whether short-term treatment interruption would restore the antiviral efficacy of a subsequent salvage gigatherapy with several potent drugs to which the patient has not yet been exposed (including drugs still under investigation in 2004, e.g., tipranavir, TMC 114) and new classes of drug with no cross-resistance to the previously failing treatment) warrants further evaluation.

Note: Presented in part at the Ninth EACS Conference, October 2003, Warsaw, Poland [abstract F7/5].