The management of patients with virologic failure and multiple drug resistance remains a challenge. With the introduction of new drug classes such as integrase inhibitors and CCR5 antagonists, treatment of multiple-class resistant HIV infection has become easier and more successful, and today's goal even for a salvage therapy is a complete virologic suppression.1 However, many regimens for multiresistant HIV infection are still complex, expensive, and associated with toxicity and drug interactions.2-4 Different strategies have therefore been pursued as alternatives to the switch to a completely new, fully suppressive salvage regimen. Treatment interruptions in failing patients with multidrug resistance have shown conflicting results in comparison with an uninterrupted, immediate start of a new salvage regimen, with some patients rapidly progressing to a severe immune deficiency or new AIDS-defining clinical events.5,6 Also, structured treatment interruptions in patients with virologic suppression have yielded negative results, and this strategy has since been largely abandoned.7 Today, a complete treatment interruption must be considered as disadvantageous for virtually all patients, with or without virologic failure. Partial treatment interruptions, however, may still be an option for selected patients who present with virologic failure and arguments against a new, fully suppressive regimen.
A number of resistance mutations have been shown to impair the replication capacity (RC) of HIV in vitro. This is particularly true for the reverse transcriptase (RT) M184V mutation, causing resistance to lamivudine (3TC) and emtricitabine, and for various mutations in the protease gene.8-11 Consequently, treatment with 3TC monotherapy in patients failing virologically and harboring the M184V mutation was proposed to maintain the selection pressure on M184V mutants, thus keeping a virus population with reduced fitness. In a randomized study in such patients comparing 3TC monotherapy with a complete treatment interruption, 3TC monotherapy was associated with a slower immunologic and clinical progression and with a delayed recovery of the RC of the virus.12 The goal of this prospective, pilot study was to further characterize the time course of different laboratory and clinical markers in patients using the strategy of 3TC monotherapy as a partial treatment interruption.
Patients and Study Procedures
This was a prospective one-arm, 48-week pilot trial performed within the Swiss HIV Cohort Study and two additional German centers. Inclusion criteria were documented HIV-1 infection with clade B virus, virologic failure on a stable (at least 3 mo duration) antiretroviral regimen, defined as two consecutive HIV RNA values greater than 400 copies/mL with the screening HIV RNA between 1,000 and 100,000 copies/mL, CD4 greater than 300 cells/μL, no active opportunistic infection, and written informed consent. The HIV genotypic resistance testing had to show presence of the 184V or 184I RT mutation, or phenotypic 3TC resistance, and additional resistance mutations, defined as either one or more major nucleoside reverse transcriptase inhibitor (NRTI) mutation (K65R; 69 insertion; T69D,N; L74I,V; V75A,T; Q151L,M; T215F,Y) and one or more major protease inhibitor (PI) mutation (D30N; G48V; I50N; V82A,F,S,T; I84V,A; L90M), or phenotypic resistance to at least one additional NRTI and one PI. Exclusion criteria were documented prior intolerance of 3TC, participation in another clinical trial, treatment with immune modulators, or being on a structured treatment interruption.
All participants had study visits at weeks -2 (screening), 0, 6, 12, 24, 36, and 48. CD4 lymphocytes and HIV RNA were determined at weeks -2 (screening), 6, 12, 24, 36, and 48. RC was measured at weeks -2 (screening), 12, 24, and 48. At week 0, patients switched to monotherapy with 3TC 300 mg qd. The study was approved by each center's institutional review board, and all patients gave voluntary written informed consent.
The course of HIV infection was followed using standard laboratory procedures. HIV-1 RNA was measured using ultrasensitive polymerase chain reaction (PCR) with a detection limit of 50 copies/mL (Amplicor, Roche Diagnostics, Rotkreuz, Switzerland).
Genotypic resistance was determined using the ViroSeq HIV-1 Genotyping kit (Abbott Molecular, Des Plaines, IL, USA); PI and RT sequences were amplified with the Platinum PCR kit (Invitrogen, Paisley, UK). Resistance-associated mutations were identified using version 4.2.6 of the STANFORD algorithm (http://hivdb.stanford.edu). From the viral genotype, a genotypic sensitivity score (GSS, indicating the number of drugs in study regimen to which the patient has genotypic sensitivity) was determined using the Rega Institute algorithm v.7.1.1 (http://hivdb6.stanford.edu/asi/deployed/hiv_central.pl?program=hivalg&action=showMutationForm, accessed on June 19, 2008).13 Phenotypic resistance was determined as reported previously.14-16
RC was assessed using a biological assay using a viral infection system with three to four rounds of HIV replication.14-16 In brief, PI and RT were amplified directly from patients' plasma and site-specifically placed into one of two HIV-1 provirus cassettes that carries a deletions of the respective gene. This system had no vector background because a replication competent virus is reconstituted only upon insertion of the patient-derived sequence. We deliberately inserted protease (PR) or RT genes separately. This strategy allowed us to verify that any change in the replicative capacity can be assigned to the respective protein. Aside from the protease or RT sequences, derived from the patient's virus, all other HIV genes in the construct remained strictly isogenic, and a position-precise insertion by way of restriction sites eliminates variability of the insertion site. The RC was reported as the absolute percentage of replication against the wild-type virus performance, analyzed in parallel in the same experiment. Numerous publications demonstrate that regions outside protease can influence viral function and fitness. It has been shown that particularly the p6 protein, which is responsible for interaction with other viral proteins,17,18 is sensitive to alterations, leading to significant effects on viral replication. In a pilot study (Louvel, manuscript in preparation), we have shown that exchanges between protease and p6 of different donor viruses can reduce or enhance infectivity in vitro. Because no clear trend was visible given that the precise exchange of the protease gene simplified testing and analyses, we did not include the p6 region in the transferred patients' viral gene sequence.
Sample Preparation and Amplification
Virus was pelleted by centrifugation at 50,000g for 80 minutes at 4°C, prepared from 1 mL of human EDTA-plasma. Pellets were redissolved in 600 μL of guanidinium isothiocyanate lysis buffer and RNA extracted according to the Cobas protocol (Roche Molecular Diagnostics, Indianapolis, IN, USA). Reverse transcription was performed using the ViroSeq HIV-1 Genotyping kit (Abbott Molecular, Des Plaines, IA); PR and RT sequences were amplified with the Platinum PCR kit (Invitrogen, Inc., Paisley, UK). Therein, the forward primer spans the existing restriction site ApaI, and the reverse primer contained a restriction site such as PinAI site. The amplification product spans the p7-p1-p6 protease cleavage sites in gag to the end of the coding region of pr and, for the second cassette, including the start of RT up to codon 335.
Construction of Recombinant Virus
Two retroviral vector cassettes designed to assess antiretroviral drug susceptibility were constructed based on NL4-3. They carried either a deletion between the ApaI (RE1) and a newly introduced, neutral-restriction site SmaI (RE2) precisely between protease and RT gene (“PR-cassette”) or alternatively between the SmaI site (RE2) and a site for PinAI (RE3) to exchange the relevant region of the RT gene. Recombinant viruses were prepared from human plasma without clonal selection (pools) to capture and preserve to the extent possible the heterogeneity of PR and RT sequences of a patient's virus population. Primers for amplifying the corresponding regions were designed to contain the same restriction sites. For PR analysis, amplicons were digested with RE1 and RE2, purified by agarose gel electrophoresis, and ligated to RE1- and RE2-digested vector DNA. The identical procedure was followed for RT. Natural occurrence of internal RE1, RE2, and RE3 sites was infrequent (between 0.1 and 2%) in the Swiss collection (PhenoBase, 2006). Ligation reactions were used to transform competent HB101/lambda (Promega GmbH, Mannheim, Germany). RV plasmid DNA was purified using the NucleoSpin Plasmid kit (Macherey-Nagel AG, Oensingen, Switzerland).
Drug Susceptibility Assay
Human epitheloid cells were transfected with either recombinant virus-DNA preparations or a wild-type reference and dispensed into 96-well plates containing serial dilutions of various PI and RTI. Cells were cultivated at 37°C in a 7% CO2 atmosphere for virus propagation. For amplification, progeny virus was then used to initiate an infection of the human lymphocyte line CEM-SS in 96-well plates in the presence of antiretroviral drugs. Under infection-enhancing conditions, virus from these cells readily transmitted to a third reporter cell containing an LTR-driven lacZ gene. For determination of infection and RC, a colorimetric assay readout (at 405 nm) was normalized for time of reporter development after cell fixation and expressed as “percent viral inhibition.” Positive and negative control wells were included in each 96-well plate and for curve fitted a suitable software was used (XLfit v 4.0.1, IDbusiness solutions, Guilford, UK). The fold change in drug susceptibility was determined by comparing for each sample the IC50 value for the virus with the IC50 for the drug-sensitive reference virus NL4-3. Replicative fitness of patient-derived virus was expressed as percent of the reference (NL4-3) in the same experiment.
Because this was a pilot project exploring a new treatment strategy, no formal sample size calculation was made. However, we assumed that within the 48 weeks of the study treatment with 3TC monotherapy, not more than 10 of 30 (33%) patients would reach the study endpoint (CD4 decrease by 30% or below 200 cells/μL), necessitating the restart of combination highly active antiretroviral therapy (HAART). In case of a failure rate of 30% (10 of 30), the corresponding upper 95% confidence interval is 50%, thus ascertaining that at least 50% of patients could safely undergo this treatment strategy during 48 weeks. An interim analysis was scheduled to be performed after 10 patients had available 3-month data. The study would have been prematurely terminated if more than four patients had CD4 counts less than 200/μL or a drop of the absolute CD4 number exceeding 50%.
The protocol-defined primary study endpoints were time to CD4 decrease by 30% or to below 200 cells/μL or the need to restart HAART within 12 months. Because only those patients restarted HAART who reached a CD4 endpoint, the criterion of restarting HAART was not separately analyzed. Therefore, the duration of follow-up is primarily reported as the time until the CD4 endpoint was reached or until the withdrawal of consent or until the end of observation at week 48, whichever came first. Because some patients elected to continue 3TC monotherapy even after they reached the CD4 endpoint, we also report a secondary analysis of the entire follow-up on 3TC monotherapy up to the maximum at week 48. For the secondary analyses comparing baseline and the end of study, the last observation carried forward principle was applied.
The statistical comparisons were two sided at the 5% level. The analysis and graphics were performed using Stata 9.2 (StataCorp, College Station, TX, USA) and Prism 5.01 (GraphPad Software, San Diego, CA, USA).
Overall, 28 patients were recruited in the study. Two were not evaluable because of violation of the inclusion/exclusion criteria or because of loss of follow-up. The demographic characteristics of the 26 evaluable patients are shown in Table 1. All patients had virologic failure with a median HIV-1 RNA of 3866 copies/mL (interquartile range [IQR]: 1,112-22,000) and CD4 cell count of 432/μL (IQR: 378-540). The median number of prior antiretroviral regimens with virologic failure was four, the highest historical HIV-1 RNA 148,592 copies/mL, and the CD4 nadir 150/μL. Eight (31%) had prior AIDS-defining diseases. The antiretroviral therapy at baseline was PI based in 16 patients (mainly lopinavir, nelfinavir, and atazanavir), non-nucleoside based in 7 (nevirapine and efavirenz), and nucleoside analogue based in 3 patients. The nucleoside analogue backbone included 3TC in 23, abacavir without 3TC in 2 patients, and didanosine and zidovudine in 1 patient.
At baseline, viral isolates from 24 patients contained the M184V mutation, and 2 patients had prior documented M184V mutation, 1 of whom showing phenotypic resistance to 3TC at baseline. The median GSS for the regimen at baseline was 1.2. RC, available for 25 patients, and was 46% (IQR: 36-77) for RT and 61% (IQR: 29-81) for PI.
During the follow-up, 17 (65%) patients experienced a CD4 drop by greater than 30% or to less than 200/μL before or at week 48. Two additional patients withdrew their consent at weeks 12 and 24, respectively, without reaching this CD4 endpoint. For all 26 patients, the probability to reach the predefined primary endpoint of CD4 drop by greater than 30% or to less than 200/μL CD4 endpoint was 15% at the study visit of 12, 36% at 24, 57% at 36, and 70% at 48 weeks, respectively (Kaplan-Meier estimates). The time to reaching the CD4 endpoint overall was a median of 6.0 months, but it was significantly shorter in patients who were on a PI-containing HAART versus those without a PI at baseline (Fig. 1). In a multivariable Cox regression analysis, the predictors for reaching the CD4 endpoint were HIV RNA (hazard ratio 1.148 for each increase by 10,000 copies/mL, 95% confidence interval 0.945-1.394) and PI-containing HAART (hazard ratio 3.45, 95% confidence interval 1.09-10.9) at baseline. All other demographic and laboratory predictors were not significant predictors, in particular nadir and baseline CD4 count, and the RC. For all 26 patients, the CD4 count decreased by 147 cells/μL (IQR: −195 to −88; wk 48 or CD4 endpoint or study discontinuation), and the HIV RNA increased by log 0.78 (IQR: 0.38-1.12).
Extended Follow-up and Secondary Analyses
Despite reaching an earlier CD4 endpoint, two patients elected to continue 3TC monotherapy until week 36, and another eight continued until week 48. In total, 18 patients stayed on 3TC monotherapy during the entire 48 weeks, whereas 8 changed from 3TC to combination antiretroviral therapy before reaching week 48. The overall follow-up on 3TC monotherapy was 21.7 patient-years. At week 48, CD4 fell by 147 cells/μL (IQR: −63 to −192), and HIV RNA increased by a median of 0.93 log (IQR: 0.49-1.17) in the 18 patients followed until week 48. None of the patients suffered any new or relapsing CDC B or C diseases, and no clinical symptoms suggestive of an acute retroviral syndrome and no 3TC-related adverse events were reported.
Because patients who switched from a PI-containing HAART to 3TC monotherapy reached the primary endpoint faster and more frequently, we examined the influence of the type of HAART regimen at baseline in more detail (Table 2). At the time of switch, patients on non-PI-containing and those on PI-containing HAART did not differ regarding their laboratory markers. Also, they had similar disease severity (4 with AIDS in each group) and duration of prior antiretroviral therapy (8.7 vs. 9.3 yr). The number of patients who reached the CD4 endpoint, however, differed significantly, with 4 of 10 (40%) versus 13 of 16 (81%), and the median time to the CD4 endpoint was 10.3 versus 6.0 months (P < 0.02), respectively. Also, the decrease in CD4 lymphocytes was significantly greater in patients who were on a PI-containing regimen at the time of switch to 3TC monotherapy (Fig. 2). We therefore evaluated whether these differences might be explained by the RC of the viral isolates. Indeed, the PI-containing group showed a trend toward lower RC of the protease at baseline, which increased during the study (from mean 51% to 72%; P = 0.07). The increase was less pronounced in patients on non-PI-containing HAART (RC of the protease changing from 70% to 82%; P = 0.32). These results suggest that the fitness of the protease gene at baseline was higher in the absence of the selection pressure exercised by a PI and that the withdrawal of the PI in the PI group allowed for a higher recovery of the protease gene fitness (Fig. 3). The RC of the RT increased significantly from mean 53% to 73% (P < 0.01), illustrating that 3TC monotherapy was not able to keep the fitness of the RT gene as much suppressed as the prior HAART. There were no differences between the non-PI versus PI group regarding the RC of the RT; both groups had similar values at baseline (53 vs. 52%) and similar rises.
The genotypic resistance testing yielded a large spectrum of mutations in both the RT and the protease gene (Table 3). During the course of the study, the frequency of numerous mutations decreased in favor of the wild type, except for the M184V mutation, which was maintained by 3TC in all patients. At baseline, the M184V mutation was not detected in two patients, but the mutation was known from prior genotypic resistance determinations. We determined the GSS (Rega Institute algorithm) for the combination regimen at baseline for every patient throughout the study. At baseline, the mean GSS was 1.20 (range: 0-2.25), and by the end of the study, the GSS rose nonsignificantly to 1.55 (range: 0-3.5; P = 0.13), in concordance with the regression of resistance mutations. In the PI-containing group, the GSS at baseline was 1.33 (composed of 0.53 for RT inhibitors and 0.80 for PIs) and rose to 1.70 (composed of 0.84 for RT inhibitors and 0.95 for PIs) at the end of study.
In this prospective pilot study including patients with virologic failure and 3TC resistance, continuation of 3TC alone led to a substantial decline in CD4 cells and to an increase in HIV-1 RNA of 0.93 log by week 48. Unreported before, we saw a differential effect based on presence or absence of a PI in the failing regimen at baseline. The number of patients who reached the predefined CD4 endpoint were significantly higher in the PI-containing group, and the magnitude of CD4 decrease was also significantly higher in this group. Our data indicate that withdrawal of the selection pressure exercised by the PI led to the reversal of some of the PI-induced resistance mutations with a concurrent recovery of the protease gene fitness, thus affecting the CD4 cells more strongly than in the group without a PI.
The only randomized study on 3TC monotherapy that compared complete treatment interruption with 3TC monotherapy demonstrated a significantly higher HIV RNA rebound (approximately +1.1 vs. +0.5 log copies/mL) and a trend toward larger decline in CD4 cells (approximately −125 vs. −75 cells/μL at week 12) with complete treatment interruption.12 In our study, the mean change at week 12 was +0.7 log HIV RNA copies/mL and −56 CD4 cells/μL, showing comparable values with the 3TC monotherapy arm of Castagna et al.12 In other trials of structured (complete) treatment interruption in patients with virologic failure, the rebound of HIV RNA ranged from +0.3 to +0.8 log HIV RNA copies/mL, and the CD4 declined from −60 to −75 cells/μL.5,19,20 In contrast, patients on successful therapy with virologic suppression who stop treatment completely deteriorate much faster, with HIV RNA increases of greater than 2 log copies/mL and CD4 drops of approximately 200 cells/μL within 12 weeks.7,21 These differences support the hypothesis that viruses selected by failing antiretrovirals have an impaired fitness, and withdrawal of the failing treatment will have a smaller effect on the laboratory parameters than withdrawal of a fully suppressive therapy. Further analyses of the above-mentioned randomized trial of 3TC monotherapy show that the RC rose both in patients who switched to 3TC monotherapy and in those with a complete treatment interruption, but the recovery of the viral fitness was greater in patients with complete treatment interruption.22 This is consistent with our observation that 3TC monotherapy is only partially effective in keeping the viral fitness suppressed.
The limitations of this study are the lack of a comparative group, the selective inclusion criteria, and the small sample size. The strengths, on the other hand, are a detailed virologic characterization of the isolates. In particular, the differential measurement of the fitness of the RT and of the PI gene allowed us to distinguish between the effects of RT inhibitors and PIs that were contained in the baseline regimen.
3TC monotherapy had an excellent tolerability, and no HIV-related or other complications occurred in our patients. Our study, however, was not powered to detect infrequent adverse events or deleterious clinical effects. CD4-guided treatment interruptions bear the risk of faster clinical progression in comparison with treatment continuation,7,23 and this risk may exist with 3TC monotherapy, too. Since the start of this study, the options for patients with virologic failure have improved because of the availability of new drugs such as raltegravir, maraviroc, or etravirine. These drugs make the option of a temporary “low-fitness” holding strategy partly obsolete. Therefore, 3TC monotherapy should only be considered as an alternative in patients who fail a regimen consisting of NRTI and non-NRTI and have either adverse events or drug interaction issues that raise caution against the immediate introduction of a fully suppressive new combination therapy. Even in such a case, 3TC monotherapy will remain a temporary solution in most cases, allowing a patient to gain several months with a simple and well-tolerated therapy.
Maintenance of the M184V mutation can be achieved in vitro also by emtricitabine, abacavir, or by lamivudine at subtherapeutic doses.24 Therefore, a daily dose of 150 mg of lamivudine may be sufficient to achieve this goal, but this dose has not been prospectively studied. Continuation of abacavir in the presence of viral replication, however, bears the risk of selecting additional resistance mutations, and abacavir monotherapy as a substitute for 3TC should not be used.
Keeping the M184V mutation may not be beneficial within a combination therapy in which other drugs are effective enough to suppress viral replication. In a randomized clinical trial comparing the continuation versus the discontinuation of 3TC in patients failing a 3TC-containing regimen and starting a new therapy, no virologic or immunologic benefit of continuing 3TC could be demonstrated.25 In the context of full viral suppression, the M184V-related impairment of viral fitness will not translate into any measurable benefit.
In some studies, reasonable maintenance of the immune status could be achieved in virologically failing patients by continuing two or more NRTI while discontinuing the PIs.26-29 It remains unclear whether such a strategy is really superior to 3TC monotherapy because no randomized comparisons have been performed. The same holds true for another small study using a maintenance treatment with 3TC and boosted low-dose indinavir.30 Overall, these approaches would require larger and randomized studies to evaluate the best strategy in patients for whom a fully suppressive regimen is not an option.
In conclusion, 3TC monotherapy as a partial treatment interruption in this particular group of patients did not prevent immunologic deterioration in the majority of them. It may be considered as a temporary maintenance strategy in selected patients who are failing under RT inhibitors only. The withdrawal of the residual activity of a PI within the baseline regimen led to a faster decline in CD4 cells, possibly because of an additional increase in the fitness of the protease gene.
The authors thank Christine Schneider, RN, for her expert help with patient management and data entry.
1. Hammer SM, Eron JJ Jr, Reiss P, et al. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society-USA panel. JAMA
2. Hill AM, Smith C. Analysis of treatment costs for HIV RNA reductions and CD4 increases for darunavir versus other antiretrovirals in treatment-experienced, HIV-infected patients. HIV Clin Trials
3. Keiser O, Fellay J, Opravil M, et al. Adverse events to antiretrovirals in the Swiss HIV Cohort Study: effect on mortality and treatment modification. Antivir Ther
4. Tozzi V, Bellagamba R, Castiglione F, et al. Plasma HIV RNA decline and emergence of drug resistance
mutations among patients with multiple virologic failures receiving resistance
testing-guided HAART. AIDS Res Hum Retrovir
5. Benson CA, Vaida F, Havlir DV, et al. A randomized trial of treatment interruption before optimized antiretroviral therapy for persons with drug-resistant HIV: 48-week virologic results of ACTG A5086. J Infect Dis
6. Katlama C, Dominguez S, Gourlain K, et al. Benefit of treatment interruption in HIV-infected patients with multiple therapeutic failures: a randomized controlled trial (ANRS 097). AIDS
7. El-Sadr WM, Lundgren JD, Neaton JD, et al. CD4+
count-guided interruption of antiretroviral treatment. N Engl J Med
8. Barbour JD, Wrin T, Grant RM, et al. Evolution of phenotypic drug susceptibility and viral replication capacity
during long-term virologic failure of protease inhibitor therapy in human immunodeficiency virus-infected adults. J Virol
9. Bleiber G, Munoz M, Ciuffi A, et al. Individual contributions of mutant protease and reverse transcriptase to viral infectivity, replication, and protein maturation of antiretroviral drug-resistant human immunodeficiency virus type 1. J Virol
10. Deval J, White KL, Miller MD, et al. Mechanistic basis for reduced viral and enzymatic fitness of HIV-1 reverse transcriptase containing both K65R and M184V mutations. J Biol Chem
11. Zennou V, Mammano F, Paulous S, et al. Loss of viral fitness associated with multiple Gag and Gag-Pol processing defects in human immunodeficiency virus type 1 variants selected for resistance
to protease inhibitors in vivo. J Virol
12. Castagna A, Danise A, Menzo S, et al. Lamivudine monotherapy
in HIV-1-infected patients harbouring a lamivudine
-resistant virus: a randomized pilot study (E-184V study). AIDS
13. Van Laethem K, De Luca A, Antinori A, et al. A genotypic drug resistance
interpretation algorithm that significantly predicts therapy response in HIV-1-infected patients. Antivir Ther
14. Sune C, Brennan L, Stover DR, et al. Effect of polymorphisms on the replicative capacity of protease inhibitor-resistant HIV-1 variants under drug pressure. Clin Microbiol Infect
15. Holguin A, Sune C, Hamy F, et al. Natural polymorphisms in the protease gene modulate the replicative capacity of non-B HIV-1 variants in the absence of drug pressure. J Clin Virol
16. Louvel S, Battegay M, Vernazza P, et al. Detection of drug-resistant HIV minorities in clinical specimens and therapy failure. HIV Med
17. Gottlinger HG, Dorfman T, Sodroski JG, et al. Effect of mutations affecting the p6 gag protein on human immunodeficiency virus particle release. Proc Natl Acad Sci U S A
18. Yu XF, Dawson L, Tian CJ, et al. Mutations of the human immunodeficiency virus type 1 p6Gag domain result in reduced retention of Pol proteins during virus assembly. J Virol
19. Lawrence J, Mayers DL, Hullsiek KH, et al. Structured treatment interruption in patients with multidrug-resistant human immunodeficiency virus. N Engl J Med
20. Walmsley SL, Thorne A, Loutfy MR, et al. A prospective randomized controlled trial of structured treatment interruption in HIV-infected patients failing highly active antiretroviral therapy (Canadian HIV Trials Network Study 164). J Acquir Immune Defic Syndr
21. Skiest DJ, Su Z, Havlir DV, et al. Interruption of antiretroviral treatment in HIV-infected patients with preserved immune function is associated with a low rate of clinical progression: a prospective study by AIDS Clinical Trials Group 5170. J Infect Dis
22. Gianotti N, Tiberi S, Menzo S, et al. HIV-1 replication capacity
and genotype changes in patients undergoing treatment interruption or lamivudine monotherapy
. J Med Virol
23. Lundgren JD, Babiker A, El-Sadr W, et al. Inferior clinical outcome of the CD4+
cell count-guided antiretroviral treatment interruption strategy in the SMART study: role of CD4+
Cell counts and HIV RNA levels during follow-up. J Infect Dis
24. Petrella M, Oliveira M, Moisi D, et al. Differential maintenance of the M184V substitution in the reverse transcriptase of human immunodeficiency virus type 1 by various nucleoside antiretroviral agents in tissue culture. Antimicrob Agents Chemother
25. Fox Z, Dragsted UB, Gerstoft J, et al. A randomized trial to evaluate continuation versus discontinuation of lamivudine
in individuals failing a lamivudine
-containing regimen: the COLATE trial. Antivir Ther
26. Deeks SG, Hoh R, Neilands TB, et al. Interruption of treatment with individual therapeutic drug classes in adults with multidrug-resistant HIV-1 infection. J Infect Dis
27. Sturmer M, Staszewski S, Doerr HW. Quadruple nucleoside therapy with zidovudine, lamivudine
, abacavir and tenofovir in the treatment of HIV. Antivir Ther
28. Abadi J, Sprecher E, Rosenberg MG, et al. Partial treatment interruption of protease inhibitor-based highly active antiretroviral therapy regimens in HIV-infected children. J Acquir Immune Defic Syndr
29. Llibre JM, Bonjoch A, Iribarren J, et al. Targeting only reverse transcriptase with zidovudine/lamivudine
/abacavir plus tenofovir in HIV-1-infected patients with multidrug-resistant virus: a multicentre pilot study. HIV Med
30. Launay O, Duval X, Dalban C, et al. Lamivudine
and indinavir/ritonavir maintenance therapy in highly pretreated HIV-infected patients (Vista ANRS 109). Antivir Ther