Virological and immunological impact of non-nucleoside reverse transcriptase inhibitor withdrawal in HIV-infected patients with multiple treatment failures
Aboulker a and the ANRS 107 Puzzle 2 Study Group No significant changes in viral load and CD4 cell count were observed 2-4 weeks after the withdrawal of non-nucleoside reverse transcriptase inhibitors (NNRTI) from the current therapy of patients exhibiting resistance mutations to this class of drugs. The data suggest that in the presence of specific resistance mutations NNRTI exert no residual antiretroviral activity and could be withdrawn without viral rebound. Non-nucleoside reverse
... e reverse transcriptase inhibitors (NNRTI) are presently recommended as part of an initial combined antiretroviral therapy , but are also used as a switch therapy in patients failing to tolerate protease inhibitor (PI)-containing regimens  as a replacement for PI or in an attempt at treatment simplification. They are also used as an adjunct to salvage multi-drug regimens in patients with previous failure of highly active antiretroviral therapy based on PI and other combined compounds [3,4]. The number of mutations required for the emergence of drug resistance is far greater for PI compared with NNRTI. Usually, a single key mutation leading to a single amino acid change in the NNRTI-specific pocket site of the reverse transcriptase or surrounding domains is associated with a high level of phenotypic resistance to all available drugs in that class . Therefore, the recommendation is to avoid the use of any of the available NNRTI in the presence of specific mutations to this class of drug, ruling out a residual invivo effect of NNRTI. However, in patients with limited therapeutic options and incomplete viral suppression, NNRTI are often conservatively maintained in the antiretroviral regimen, even in the presence of NNRTI-specific mutations. The aim of this study was to assess the potential residual antiretroviral effect of NNRTI in patients exhibiting viruses with NNRTI resistance mutations in the plasma, and to describe the evolution of NNRTI resistance patterns 6 months after NNRTI withdrawal. The ANRS 107 study was a randomized open-label, multiple dose study in HIV-infected patients having failed previous antiretroviral therapies including PI and NNRTI. Patients were randomly assigned to receive for the first 2 weeks either unchanged PI and nucleoside reverse transcriptase inhibitors (NRTI) (group 1) or unchanged NRTI combined with atazanavir (300 mg once a day) and ritonavir (100 mg once a day) in substitution for the failing PI (group 2). From week 3 (day 15) to week 26, patients from either group switched to ritonavir-boosted atazanavir, plus tenofovir DF 300 mg (once a day) and NRTI selected according to the reverse transcriptase baseline genotype. The protocol prescribed that ongoing NNRTI be stopped 2 weeks before the week 0 randomization visit in all recipients in order to avoid pharmacological interaction with atazanavir. The plasma HIV-RNA level and CD4 cell count were measured at week À4, week 0, week 2 and every 4-8 weeks thereafter until week 26. Genotypic mutation patterns were assessed at week À4, week 2 and week 26. In order to validate the NNRTI withdrawal, NNRTI plasma trough levels were measured at week À4, week 0 and week 2 in all patients receiving drugs from this class at the screening visit. The NNRTI withdrawal effect was assessed by the changes in the plasma HIV-RNA level and CD4 cell count from week À4 to week 0, e.g. 2 weeks after NNRTI withdrawal in all patients who stopped NNRTI, and additionally to week 2 in those from group 1 allocated to maintain the current regimen during step 1. The viral load and CD4 cell count change from week À4 to weeks 0 and 2 were summarized using means, standard deviations, 95% confidence limits, medians and ranges as appropriate. The one sample t-test and Wilcoxon's signed rank test (2-signed) were used to test the hypothesis of a null change. Fifty-three patients were enrolled in the study. Twenty-seven patients were randomly assigned to group 1 and 26 patients to group 2. Nineteen out of the 53 patients randomly selected in the trial were currently receiving a NNRTI at enrolment. The median CD4 cell counts of the patients receiving an NNRTI at enrolment and those who did not were 237 3 10 6 /l (ranged between 12 and 560) and 191 3 10 6 /l (3-639), respectively. The median plasma HIV-RNA level of the patients receiving an NNRTI at enrolment and those who did not were 4.9 log copies/ml (4-5.8) and 5.1 log copies/ml (4.1-6.2), respectively. The previously taken NNRTI were efavirenz in 35 patients (66%), nevirapine in 38 patients (72%) and delavirdine in four patients (8%). Ongoing NNRTI were efavirenz in 13 patients (68%) and nevirapine in six patients (32%). One or several NNRTI resistance mutations were found in all 19 NNRTI recipients. NNRTI mutations were not found in only nine out of the 34 remaining patients (26%) who had stopped NNRTI for virological failure for a median time of 21 months (range 2.2-42.5). Plasma trough concentrations were within the therapeutic range (1000-5000 ng/ml and 3000-8000 ng/ml for efavirenz and nevirapine, respectively) at week À4 in the 19 NNRTI recipients, and subsequently dropped below the limit of detection at weeks 0 and 2 in all of them. Two weeks after NNRTI withdrawal, the mean changes in these patients were 0.0 log 10 [95% confidence interval (CI) À0.17, 0.15] in the plasma HIV-RNA level and À13 (95% CI À29, 4) cells per mm 3 in the CD4 cell count. Mean changes assessed 4 weeks after the withdrawal in the six patients allocated to group 1 out of the 19 NNRTI recipients were 0.1 log 10 for the HIV-RNA level and À41 cells per 10 6 /l for the CD4 cell count. By comparison, mean changes were 0.0 log (95% CI À0.09, 0.09) in the plasma HIV-RNA level and À2 (95% CI À19, 16) cells in the CD4 count in the 34 patients not receiving NNRTI at entry into the study. The changes over time in NNRTI resistance mutations are presented in Table 1 . At week 26, a reversion to the wild-type NNRTI resistance mutation pattern was observed in only three of the 19 patients, whereas 15 patients did not exhibit any change in the pattern of their NNRTI mutations during the study. These data showed that none or only insignificant changes in the plasma HIV-RNA level and CD4 cell count occurred 4 weeks after NNRTI withdrawal in patients harbouring NNRTI resistance mutations. Our results are consistent with those observed in a retrospective study in 11 patients , which reported that the interruption of NNRTI was not followed by changes in the plasma HIV-RNA levels and CD4 cell counts, and that reversion to wild type occurred in only one of 11 patients 60 days after NNRTI withdrawal. In our study, three of the 19 NNRTI recipients at inclusion exhibited a reversion to wild type at week 26, and nine of the 33 patients who had stopped NNRTI for a median of 21 months before inclusion in the trial exhibited no NNRTI resistance mutation at inclusion in the study. These results suggested that the delay in observing NNRTI mutation disappearance is relatively long and that this phenomenon progresses slowly. One previous study is in accordance with this observation .