Low-level HIV-1 viraemia in patients on HAART: risk factors and management in clinical practice

Transcription

1 J Antimicrob Chemother 2015; 70: doi: /jac/dkv099 Advance Access publication 28 April 2015 Low-level HIV-1 viraemia in patients on HAART: risk factors and management in clinical practice Marc Wirden 1 3 *, Eve Todesco 1 3, Marc-Antoine Valantin 1,2,4, Sidonie Lambert-Niclot 1 3, Anne Simon 5, Ruxandra Calin 1,2,4, Roland Tubiana 1,2,4, Gilles Peytavin 6, Christine Katlama 1,2,4, Vincent Calvez 1 3 and Anne-Genevieve Marcelin INSERM, UMR S_1136, Institut Pierre Louis d Epidémiologie et de Santé Publique, F Paris, France; 2 Sorbonne Universités, UPMC Univ Paris 06, UMR_S 1136, Institut Pierre Louis d Epidémiologie et de Santé Publique, F Paris, France; 3 AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Virologie, Paris, France; 4 AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Maladies Infectieuses, Paris, France; 5 AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Médecine Interne, Paris, France; 6 AP-HP, Groupe Hospitalier X Bichat-C Bernard, laboratoire de toxicologie, Paris, France *Corresponding author. Laboratoire de Virologie, Bât. CERVI, Hôpital Pitié-Salpêtrière, 83 bvd de l Hopital, Paris, France. Tel: ; Fax: ; marc.wirden@psl.aphp.fr Received 13 January 2015; returned 26 February 2015; revised 17 March 2015; accepted 22 March 2015 Objectives: Characterization of the conditions favouring HIV-1 low-level viraemia (LLV) during treatment is required to guide strategies for prevention and cure. Methods: The characteristics and treatments of 171 patients experiencing a confirmed LLV of (PLLVs) were compared with those of 146 patients with persistently controlled viraemia. We analysed the risk factors for LLV, the parameters affecting the level of viraemia and the presence of resistance-associated mutations (RAMs). We compared outcomes for PLLVs on fully effective HAARTas a function of treatment modifications. Results: LLV was,500 in at least 90% of cases. A higher zenith viral load (VL) (5.27 versus 4.91 log 10, OR 2.23; P¼0.0003), a shorter time on continuous HAART (4.3 versus 6.8 years, OR 0.88; P¼0.0003) and previously detected RAMs (43% versus 23%, OR 2.42; P¼0.0033) were independent predictors of LLV. NNRTIs were less frequently used in PLLVs and were associated with more stable treatment. The presence of any RAM during LLV was associated with a lower zenith VL and a higher LLV. In the absence of resistance, virological success was achieved in similar proportions of patients with and without treatment modification. Conclusions: Viraemia.500 should no longer be considered to be LLV. In patients with a high zenith VL, several years on continuous HAART may be required to decrease the HIV reservoir and prevent LLV. Resistance testing is useful to detect RAMs, leading if necessary to treatment modifications. In the absence of resistance, treatment changes seemed dispensable. Keywords: HIV, low-level viraemia, highly active antiretroviral therapy, resistance-associated mutations, treatment failure Introduction The quantification of HIV-1 RNA in plasma has been the mainstay of assessments of HIV replication and efficacy of ART for almost 20 years. The first assays had a limit of quantification (LOQ) of 400, but new technologies with an LOQ of 40 or 20 have greatly improved the sensitivity. 1,2 The clinical significance of this low-level viraemia (LLV) is a matter of debate. 3,4 Most guidelines indicate that treatment should aim to achieve a viral load (VL),50, but treatment failure is still defined as a confirmed value above 200 or even 400 copies/ ml. 5 8 The range of is thus a grey zone in that there is no consensus regarding patient management. Doubts about the origin of LLV partly account for this situation. An artefactual HIV-DNA quantification in improperly processed samples has been reported The activation of latently infected cells and/or the continuous expression of an HIV proviral reservoir, as well as ongoing viral replication in sanctuaries, may be potential sources, although these hypotheses have been mainly proposed for viraemia,50. 3,12,13 The consequences of LLV are also widely debated. Some authors have suggested that viraemia below 200 or even 50 may be predictive of subsequent treatment failure or associated with the development of drug resistance, whereas others refer to a large # The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oup.com 2347

2 Wirden et al. retrospective analysis reporting no difference between virological failure thresholds of 200 and ,21 Regardless of the causes and consequences of LLV, no study has yet shown how to restore these cases of LLV to levels, 50. The principal approach is therefore the avoidance of conditions favouring LLV. We aimed to identify the characteristics and treatments of patients with confirmed LLV (PLLVs) through comparisons with patients with persistent controlled viraemia below the LOQ (PPCVs). We also identified factors affecting replication levels and the presence of resistance-associated mutations (RAMs). Finally, we compared the outcome for PLLVs with no evidence of HIV resistance as a function of the presence or absence of treatment modification. Patients and methods Study population and control group We searched the laboratory database for HIV-1-infected patients receiving treatment who had LLV (a range of ) in two consecutive tests performed from 2008 to LLV1 and LLV2 were defined as the two first LLVs found consecutively for a particular individual in that period. Patients were selected consecutively and once only. In terms of the duration and composition of treatment, the conditions commonly accepted as achieving or maintaining HIV-1 viraemia at levels,50 were required. Patients were therefore included in line with the following criteria: (i) patients were receiving HAART including at least three antiretroviral drugs from at least two different classes; (ii) patients were either in rebound, i.e. had achieved virological success (two consecutive VLs,50 ) just before LLV1 on the same HAART (PLLVAs), or had already persistent LLV for at least 1 year before LLV1 despite continuous HAART (that had been stable for at least 6 months; PLLVBs). We excluded patients being treated for hepatitis and those with incomplete viro-immunological follow-up since the time of HIV diagnosis. We selected 171 PLLVs (107 PLLVAs and 64 PLLVBs) in total. The 107 PLLVAs had achieved virological success for a median of 14 months (IQR 8 23 months) before LLV1. The control group of PPCVs had to have had viraemia consistently below the LOQ for more than 24 months, i.e. beyond the last quartile observed for the PLLVAs (23 months). This period could include VL measurements after the time of selection, and patients had to be maintained on the same HAART, with at least three VL determinations per year. Period bias was avoided by selecting PPCVs retrospectively and only once from among the subjects undergoing viraemia determinations in the same week as the PLLVs with whom they were matched. HAART groups HAART was classified into four groups according to its composition. Patients receiving therapies based on three drugs, including two NRTIs plus one NNRTI, a PI or raltegravir (an integrase inhibitor), were defined as the tnnrti, tpi and tral groups, respectively. All other alternative regimens including more than three drugs and/or maraviroc and/or enfuvirtide were assigned to the AltTrt group. HIV-1 VL and genotypic resistance testing (GRT) Only historical measurements of viraemia performed after November 2006 were considered, when the first and second versions of Roche Cobas CA/CTM assays (Roche, Meylan, France) were available for quantifying viraemia down to 50, with LOQs of 40 and 20, respectively. The pre-analytical steps were optimal to prevent any artefactual HIV-DNA quantification. Whole blood had been collected in EDTA tubes, shipped to the laboratory and centrifuged within 12 h. Plasma was tested rapidly or transferred into a secondary tube before freezing. GRT was performed on LLV2 plasma samples. The list of mutations from the International AIDS Society was used to establish the presence of RAMs. For PIs, we considered only primary mutations. A RAM was considered to be new if it had not previously been reported in the patient s history or if no previous GRTdata were available, because it was considered as new information provided by the GRT for treatment management. We defined three groups, as follows: failure of GRT (NoGRT), any RAM detected (ResGRT) and no RAM detected (WtGRT). The genotypic susceptibility score (GSS) for a HAART was calculated by adding the susceptibility scores of the component drugs as follows: susceptible¼1, intermediate resistance¼0.5 and resistance¼0, in accordance with the 2013 ANRS HIV-1 algorithm ( Statistical analysis The parameters investigated in a logistic regression model were age, sex, the nadir CD4 count (the lowest value since diagnosis), the zenith VL (the highest value since diagnosis), the number of years since diagnosis, the number of years on continuous HAART treatment, the number of lines of treatment before inclusion and the presence of any RAM in the patient s history. Variables yielding a P value,0.20 in univariate analysis were selected for a stepwise inclusion procedure in the multivariate model. Depending on the number of groups and the expected values, categorical variables were compared using the x 2 or Fisher s exact tests, with Wilcoxon or Kruskal Wallis tests used to compare continuous variables. Values of P,0.05 were considered statistically significant. Results Risk factors for PLLVs The characteristics and treatments of the 171 PLLVs and 146 PPCVs are compared in Tables 1 and 2. Threefactorswere identified as independent predictors of PLLVs: a higher zenith VL (5.27 versus 4.91 log 10 ; OR 2.23; P¼0.0003), a shorter time on continuous HAART (4.3 versus 6.8 years; OR 0.88; P¼0.0003) and the presence of any RAM in the patient s history (43% versus 23%; OR 2.42; P¼0.0033). The number of lines of treatment before inclusion was also significant in univariate analysis, but excluded in the multivariate model because of collinearity with the presence of any RAM in the patient s history. In terms of treatment (Table 2), PLLVs received the same proportion of tpi regimens as PPCVs, but fewer tnnrti regimens (12.4% versus 32.2%; P,0.0001) and more AltTrt regimens (18.2% versus 3.4%; P,0.0001). Both efavirenz and nevirapine were less frequently used in PLLVs than in PPCVs, whereas the use of etravirine, darunavir and maraviroc was more frequent in PLLVs. However, associations were found between patients characteristics and their treatments (Table 3). The patients in the tral and AltTrt groups had received almost three times as many lines of treatment as those in the other groups and also had the highest frequency of previous resistance. In a subanalysis, the patients taking etravirine or darunavir also had a higher frequency of previously reported RAMs than those receiving another drug from the same class (87% versus 9%, and 64% versus 29%, respectively; P, in both cases). Thus, the most recent drugs and AltTrt regimens were used in patients with the heaviest treatment histories and resistance backgrounds. Conversely, the tnnrti group had several points in common with the PLLV group: the lowest rate of previous RAMs and the longest period 2348

3 Risk factors for HIV low-level viraemia JAC Table 1. Comparison of population characteristics and logistic analyses of factors associated with PLLVs Population characteristics Factors associated with PLLVs parameter PLLVs (n¼171) PPCVs (n¼146) P a univariate analysis P multivariate analysis P OR (95% CI) Male, % Age (years), median (IQR) 47 (41 54) 49 (42 56) Years of VL,50 before inclusion b, median (IQR) 1.2 ( ) c 3.8 ( ) d, Years since diagnosis, median (IQR) 13 (7 17) 13 (8 18) No. of lines of any ART before inclusion, median (IQR) 6 (3 11) 4 (3 8) NI e Years of continuous HAART before inclusion, median (IQR) 4.3 ( ) 6.8 ( ) ( ) Nadir CD4 cell count (cells/mm 3 ), median (IQR) 147 (52 228) 187 (90 286) 0.01 f 0.02 f NS f Zenith VL (log 10 ), median (IQR) 5.27 ( ) 4.91 ( ),0.0001, ( ) Presence of any RAM in the patient s history, % ( ) NS, not significant. a Categorical variables were compared with the x 2 test and continuous variables with the Wilcoxon test. b Not included in the model since it was linked to the selection criteria for the PPCVs. c Calculated only for PLLVAs (n¼107). d Time after the inclusion point was not take into account for this median. e NI, not included in the model because of collinearity with the presence of any RAM in the patient s history. f Per quartile increase. of continuous HAART (P¼0.0055; Table 3), particularly in comparison with the tpi group (8.5 versus 4.8 years, respectively). Interestingly, analyses of the PPCV control group only showed that the period of the unchanged continuous regimen was still longer for patients in the tnnrti group than for those in the tpi group (4.3 versus 3.3 years; P ¼ 0.003). This suggests that NNRTI-based regimens may be more stable than PI regimens, regardless of the LLV context. Level of LLV and resistance mutations in PLLVs The 75th and 90th percentiles for viraemia in PLLVs were 180 and 422 for LLV1, respectively, and 196 and 463 for LLV2, respectively. Thus, despite the use of criteria allowing the inclusion of patients with viraemia of up to 1000, at least 75% and 90% of LLV values were below 200 and 500, respectively. LLV1 and LLV2 were not statistically different, with median values of 101 and 108, respectively. The median difference (LLV2 minus LLV1) was +10 with an interval of 2.5 months between the two measurements. These results did not differ between treatment groups, but they did differ between GRT groups (Figure 1). The ResGRT group (n¼61, 36%) had significantly higher levels of viraemia than the NoGRT (n¼68, 40%) and WtGRT (n¼42, 24%) groups and was the only group to display a significant, albeit very small, difference between LLV1 and LLV2 (+0.11 log ; P¼0.024). Above the threshold of 200 (with the NoGRT group excluded), there tended to be more patients in the ResGRT group (72% versus 53%; P¼0.07), and there were about twice as many new RAMs than were seen with avl,200 (53% versus 25%; P¼0.006; Figure 2). The multivariate analysis identified two factors as independent predictors of resistance mutations at the time of LLV2: a lower Table 2. Percentage distribution of drugs in PLLVs and PPCVs PLLVs (n¼171) PPCVs (n¼146) P Treatment group tnnrti , tpi NS tral AltTrt , NRTIs abacavir+3tc NS tenofovir+ftc NS other NS no NRTI NNRTIs efavirenz nevirapine etravirine no NNRTI NS PIs a atazanavir NS atazanavir/ritonavir NS lopinavir/ritonavir NS darunavir/ritonavir no PI Other drugs raltegravir , maraviroc NS, not significant; 3TC, lamivudine; FTC, emtricitabine. a Only the most frequent PIs used in these patients are reported. 2349

4 Wirden et al. Table 3. Comparison of population characteristics depending on the treatment group for both PLLVs and PPCVs (Kruskal Wallis and x 2 test) Parameter tnnrti (n¼68) tpi (n¼188) tral (n¼25) AltTrt (n¼36) P Age (years), median (IQR) 50 (43 58) 46 (40 53) 53 (49 62) 48 (44 57) Nadir CD4 cell count (cells/mm 3 ), median (IQR) 217 (93 303) 17 (91 270) 93 (46 192) 75 (16 155), Zenith VL (log 10 ), median (IQR) 4.8 ( ) 5.1 ( ) 5.0 ( ) 5.6 ( ), Years since diagnosis, median (IQR) 12 (8 16) 12 (6 16) 17 (11 22) 15 (13 21) No. of lines of any ART before inclusion, median (IQR) 4 (3 6) 5 (3 8) 12 (5 20) 13 (5 20), Years of continuous HAART before inclusion, median (IQR) 8.5 ( ) 4.8 ( ) 6.2 ( ) 5.2 ( ) Presence of any RAM in the patient s history, % , P = NoGRT HIV-1 VL ResGRT WtGRT 90th percentile P = LLV1 P = P < P = LLV2 Figure 1. Distribution (box plot) and comparison of both LLV values at inclusion (LLV1 and LLV2) depending on the results of GRT. The median values of LLV1 and LLV2 were, respectively: 86 for both values in the NoGRT group (n¼68); 122 and 155 in the ResGRT group (n¼61); and 117 and 112 in the WtGRT group (n¼42). zenith VL (OR 0.29; 95% CI ; P¼0.0022) and a larger number of previous treatments (OR 1.18; 95% CI ; P¼0.0002) (Table 4). The proportion of patients from the ResGRT group was similar for the PLLVAs (35%) and PLLVBs (37%) and 83% of these individuals received a treatment with a GSS score,3. RAMs tended to be more frequent with tral treatment, but they were already present in most patients histories and were new information for only 43% of them (Figure 3). New RAMs were observed in 20% of the 171 PLLVs, 34% of the LLV2 GRT group and 57% of the ResGRT group. They tended to be more frequent in patients receiving tnnrtis than in patients on other regimens (53% versus 30%; P¼0.09). (a) 100% 75% 53% 50% 25% 47% 72% 28% ResGRT WtGRT P = (b) 25% 75% 53% 47% Presence of new RAM Absence of new RAM P = Treatment interventions and outcomes for PLLVs receiving effective HAART We assessed the best way to manage PLLVs receiving a treatment presumed to be fully effective by retrospectively observing, when possible, the outcomes for 111 patients from the NoGRT (n¼65) and WtGRT (n¼38) groups and for patients in the ResGRT group who were receiving HAART with a GSS that remained 3 (n¼8). 0% (n = 71) (n = 32) (n = 71) (n = 32) Figure 2. Percentage of patients for whom the GRT showed RAMs (a) and new RAMs (b) at the time of LLV2, in relation to the threshold of 200. The x 2 test was used for the comparison. 2350

5 Risk factors for HIV low-level viraemia JAC Table 4. Population characteristics and logistic analyses of factors associated with ResGRT Population characteristics Factors associated with ResGRT univariate analysis multivariate analysis parameter ResGRT (n¼61) WtGRT (n¼42) P OR (95% CI) P OR (95% CI) Male, % NS Age (years), median (IQR) 49 (44 56) 49 (44 55) NS Nadir CD4 cell count (cells/mm 3 ), median (IQR) 107 (50 222) 154 (75 217) NS (per quartile increase) Years since diagnosis, median (IQR) 15 (12 18) 10 (6 15) ( ) NS Years of continuous HAART before inclusion, 8.1 ( ) 3.1 ( ) ( ) NS median (IQR) No. of lines of any ART before inclusion, 9 (6 17) 4 (3 9) ( ) ( ) median (IQR) Zenith VL (log 10 ), median (IQR) 5.23 ( ) 5.41 ( ) ( ) ( ) NS, not significant. 90 RAMs to NRTIs RAMs to NNRTIs Percentage of genotypes with RAMs RAMs to PIs New RAMs RAMs to RAL 10 0 All groups (n = 103) tnnrti group (n = 18) tpi group (n = 50) tral group (n = 13) AltTrt group (n = 22) Figure 3. Percentage of patients for whom the GRT, performed at LLV2, showed RAMs depending on the antiretroviral class and the treatment group. RAL, raltegravir. The treatment was modified after LLV2 (in terms of the choice of drugs or their dosage) for 60 patients (the Change group) and remained unchanged in the other 51 patients (the NoChange group). The frequency of patients for whom virological success was re-established (two consecutive VLs,50 ) did not differ significantly between the Change and NoChange groups (58% versus 53%; P¼0.7) during a median follow-up of 13 and 16 months (P¼0.62), respectively. Similarly, the median change in CD4 cell count was comparable: +73 versus +54 CD4 cells/mm 3 (P¼0.92). The treatment changes were too diverse for a more detailed analysis. However, the success rates were highest (6/6) for changes involving an increase in the dose of a PI (atazanavir or darunavir). In four cases, PI concentrations were suboptimal before the dose increase. Among the 53 patients in the ResGRT group with a GSS,3, 83% (n¼44) had their treatment modified, with a virological success of 70% (31/44) versus 33% for the patients with no change in their treatments. Discussion In this study, we aimed to determine the conditions associated with LLV through comparisons with individuals in whom viraemia 2351

6 Wirden et al. was persistently maintained below the LOQ. We studied several aspects of this issue, which may involve many factors. This, together with the precise selection criteria and confirmed LLV with consensual and stable treatments, is a particular strength of this study. The first analysis showed that PLLVs had a higher zenith VL and had been on continual HAART for shorter periods. These findings are consistent with those of a previous study reporting a higher zenith VL and higher viraemia before the initiation of therapy for patients with a persistent LLV of They are also consistent with the reactivation of latently infected cells being one of the sources of LLV. Indeed, higher zenith VL values generate a larger reservoir and thus a greater capacity to release detectable numbers of HIV. 12 A longer duration of continuous HAART may therefore be required to decrease the size of this reservoir and to minimize the risk of repeated LLV. As observed in PPCVs, a median duration of 7 years of continuous treatment may be required to prevent LLV. This implies the use of a stable, sustainable antiretroviral regimen for several years to keep the viraemia below the LOQ. In this study, the frequency of tnnrti use in PLLVs was only one-third that in PPCVs. This may be due to the low barrier to genetic resistance and the high-level cross-class resistance observed for first-generation NNRTIs. These problems prevent the use of these drugs in patients with a history of NNRTI resistance or a risk of poor adherence to treatment. However, previous studies have also reported links between NNRTI regimens and an absence of LLV or a higher frequency of PI-based regimens in PLLVs. 14,22,23 Moreover, we found that the patients treated with tnnrti regimens had the longest durations of continuous HAART and the smallest number of lines of treatment (Table 3). Limiting our analysis to PPCVs to eliminate the context of LLV and suspected poor adherence, we again observed that tnnrti was associated with longer periods on the same continuous HAART regimen than were seen for patients receiving tpi. Thus, in addition to the better penetration of NNRTIs into sanctuaries that has been described by other authors, 14,24 these observations suggest that NNRTI-based treatments may form the basis of more stable regimens that can be administered without interruption over several years, such conditions being associated with a lower risk of LLV over time, as shown above. Better long-term tolerability, longer half-lives and fewer pharmacological interactions with other drugs may account for the sustainability of such treatments. 25,26 However, further studies focusing specifically on this issue are required to confirm this hypothesis. In contrast to the observations for efavirenz and nevirapine, the most recent compounds (etravirine, darunavir, raltegravir andmaraviroc)weremorefrequentlyusedinpllvsthanin PPCVs (Table 2). Indeed, during the study period, these new drugs and new classes of drugs were initially used in the patients with the heaviest treatment histories, particularly those displaying resistance to the oldest drugs (Table 3 and Figure 3). The use of these drugs is therefore more likely to reflect a history of heavy treatment rather than being the cause of LLV. After 2011, the wider use of these compounds at earlier stages of treatment may have generated different results. The potential role of drug resistance in the ongoing replication of HIV led us to our second hypothesis concerning the origin of LLV. Indeed, the presence of any RAM in a patient s history was also identified as a risk factor for belonging to the PLLV group, independently of the zenith VL and the duration of HAART (Table 1). Moreover, RAMs were also found to be associated with a higher level of viraemia and a significant increase between LLV1 and LLV2, as well as with a lower zenith VL. All these findings confirm that resistance mutations, whether already recorded in the patient s history or discovered during LLV, constitute a second factor associated with LLV, independently of a high zenith VL inducing large viral reservoirs. Overall, these data are consistent with the two hypotheses on the biological origin of LLV, both of which may apply within a single patient. The first hypothesis involves a higher zenith VL and a larger HIV reservoir, leading to the release of HIV and requiring several years of stable, continual treatment to reduce it. The second hypothesis involves ongoing replication, with RAMs either already present or emerging. Tobin et al. 13 put forward the same conclusion following a sequencing analysis of LLV. Another interesting result of this study is the level of LLV reported. In the present study, 90% and 75% of values for viraemia were below 500 and 200, respectively. Thus, confirmed or persistent LLV in the range of appears to be uncommon. Laprise et al. 18 recently reported a much higher risk of experiencing viraemia of more than 1000 for patients with LLV in this range (almost five times higher versus twice as high). Thus, the virological outcome is clearly different above 500 and, contrary to what is observed in many studies, 15,16,18,19,27 such viraemia should be considered purely as treatment failure rather than LLV. Moreover, the use of an upper cut-off point of LLV closer to the thresholds defining virological failure 500 or 200 would clarify debates on this issue. The final aim of this work was to describe the management of patients experiencing LLV with no evidence of drug resistance. Our findings indicate that treatment modification is no more effective than the status quo for achieving virological success in the following months. Previous studies have also shown that changes to the treatment regimen had no impact on residual viraemia. 28,29 These observations provide support for the hypothesis that, in the absence of RAMs, the origin of LLV is likely to be release from a reservoir, which can be resolved by treatment for longer periods, as demonstrated in the first analysis in this study. However, improvements in treatment adherence, as probably claimed by physicians, may contribute to these results. Increasing the dose of darunavir or boosting atazanavir also appeared to be an effective intervention. One of the limitations of this study is the absence of more extensive pharmacological data that could provide useful information about suboptimal drug concentrations. In conclusion, the present study provided objective data to define LLV as viraemia below 500 or even 200. LLV is associated with a high zenith VL, a shorter duration of HAARTand the presence of previously detected or new RAMs. Our findings confirm that HAART during the earliest stages of HIV infection, particularly during primary infection, must be useful for minimizing the zenith VL. Otherwise, patients may require several years of uninterrupted HAART to prevent LLV originating from the viral reservoir. To this end, the advantage of NNRTI-based regimen requires confirmation in other studies. As LLV can also originate from ongoing replication, genotyping may be useful for assessing the presence of resistance and adapting the antiretroviral regimen, even for viraemia below 200. In the absence of RAMs, virological success is achieved in similar proportions of patients with or without 2352

7 Risk factors for HIV low-level viraemia JAC changes to their treatment. However, more comprehensive prospective studies are required to determine what specific intervention,suchasanincreaseinpidose,isaneffectivewayof increasing the rate of virological success in such a context of LLV. Acknowledgements This work has been presented in part at the International Workshop on Antiretroviral Drug Resistance, Berlin, Germany, 2014 (Abstract 96). Funding This work was supported by the Agence Nationale de Recherches sur le SIDA (ANRS). Transparency declarations None to declare. References 1 Saag MS, Holodniy M, Kuritzkes DR et al. HIV viral load markers in clinical practice. Nat Med 1996; 2: Cobb BR, Vaks JE, Do T et al. Evolution in the sensitivity of quantitative HIV-1 viral load tests. J Clin Virol 2011; 52 Suppl 1: S Doyle T, Geretti AM. Low-level viraemia on HAART: significance and management. Curr Opin Infect Dis 2012; 25: Ryscavage P, Kelly S, Li JZ et al. Significance and clinical management of persistent low-level viremia and very-low-level viremia in HIV-1-infected patients. Antimicrob Agents Chemother 2014; 58: Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Department of Health and Human Services aidsinfo.nih.gov/contentfiles/adultandadolescentgl.pdf. 6 Thompson MA, Aberg JA, Hoy JF et al. Antiretroviral treatment of adult HIV infection: 2012 recommendations of the International Antiviral Society-USA panel. JAMA 2012; 308: Williams I, Churchill D, Anderson J et al. British HIV Association guidelines for the treatment of HIV-1-positive adults with antiretroviral therapy 2012 (Updated November All changed text is cast in yellow highlight). HIV Med 2014; 15 Suppl 1: Ribaudo HJ, Kuritzkes DR, Schackman BR et al. Designissuesininitial HIV-treatment trials: focus on ACTG A5095. Antivir Ther 2006; 11: Rebeiro PF, Kheshti A, Bebawy SS et al. Increased detectability of plasma HIV-1 RNA after introduction of a new assay and altered specimenprocessing procedures. Clin Infect Dis 2008; 47: Stosor V, Palella FJ Jr., Berzins B et al. Transient viremia in HIV-infected patients and use of plasma preparation tubes. Clin Infect Dis 2005; 41: Wan H, Seth A, Rainen L et al. Coamplification of HIV-1 proviral DNA andviralrnainassaysusedforquantificationofhiv-1rna.jclin Microbiol 2010; 48: Chun TW, Murray D, Justement JS et al. Relationship between residual plasma viremia and the size of HIV proviral DNA reservoirs in infected individuals receiving effective antiretroviral therapy. J Infect Dis 2011; 204: Tobin NH, Learn GH, Holte SE et al. Evidence that low-level viremias during effective highly active antiretroviral therapy result from two processes: expression of archival virus and replication of virus. JVirol2005; 79: Doyle T, Smith C, Vitiello P et al. Plasma HIV-1 RNA detection below 50 and risk of virologic rebound in patients receiving highly active antiretroviral therapy. Clin Infect Dis 2012; 54: Santoro MM, Fabeni L, Armenia D et al. Reliability and clinical relevance of the HIV-1 drug resistance test in patients with low viremia levels. Clin Infect Dis 2014; 58: TaiwoB, GallienS, AgaEet al. Antiretroviral drug resistance in HIV-1-infected patients experiencing persistent low-level viremia during first-line therapy. J Infect Dis 2011; 204: Henrich TJ, Wood BR, Kuritzkes DR. Increased risk of virologic rebound in patients on antiviral therapy with a detectable HIV load,48. PLoS One 2012; 7: e Laprise C, de Pokomandy A, Baril JG et al. Virologic failure following persistent low-level viremia in a cohort of HIV-positive patients: results from 12 years of observation. Clin Infect Dis 2013; 57: Aleman S, Soderbarg K, Visco-Comandini U et al. Drug resistance at low viraemia in HIV-1-infected patients with antiretroviral combination therapy. AIDS 2002; 16: Gandhi RT, Deeks SG. Plasma HIV-1 RNA levels during antiretroviral therapy: how low is low enough? Clin Infect Dis 2012; 54: Ribaudo H, Lennox J, Currier J et al. Virologic failure endpoint definition in clinical trials: is using HIV-1 RNA threshold,200 better then.50? An analysis of ACTG studies. In: Abstracts of the Sixteenth Conference on Retroviruses and Opportunistic Infections, Montreal, Canada, Abstract 580. Foundation for Retrovirology and Human Health, Alexandria, VA, USA. 22 Vancoillie L, Demecheleer E, Callens S et al. Markers associated with persisting low-level viraemia under antiretroviral therapy in HIV-1 infection. J Antimicrob Chemother 2014; 69: Geretti AM, Smith C, Haberl A et al. Determinants of virological failure after successful viral load suppression in first-line highly active antiretroviral therapy. Antivir Ther 2008; 13: Haim-Boukobza S, Morand-Joubert L, Flandre P et al. Higher efficacy of nevirapine than efavirenz to achieve HIV-1 plasma viral load below 1 copy/ml. AIDS 2011; 25: Ribaudo HJ, Haas DW, Tierney C et al. Pharmacogenetics of plasma efavirenz exposure after treatment discontinuation: an Adult AIDS Clinical Trials Group Study. Clin Infect Dis 2006; 42: Riddler SA, Haubrich R, DiRienzo AG et al. Class-sparing regimens for initial treatment of HIV-1 infection. N Engl J Med 2008; 358: Swenson LC, Min JE, Woods CK et al. HIV drug resistance detected during low-level viraemia is associated with subsequent virologic failure. AIDS 2014; 28: Archin NM, Cheema M, Parker D et al. Antiretroviral intensification and valproic acid lack sustained effect on residual HIV-1 viremia or resting CD4+ cell infection. PLoS One 2010; 5: e Dinoso JB, Kim SY, Wiegand AM et al. Treatment intensification does not reduce residual HIV-1 viremia in patients on highly active antiretroviral therapy. Proc Natl Acad Sci USA 2009; 106: