Impact of baseline HIV-1 RNA levels on initial highly active antiretroviral therapy outcome: a meta-analysis of 12,370 patients in 21 clinical trials

Authors


  • This study was presented in part at the World AIDS Conference in Washington DC, USA, July 2012 (Abstract TUPE085).

Correspondence: Dr Christoph Stephan, Johann Wolfgang Goethe University Hospital, Medical Department no.2/Infectious Diseases Unit, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany. Tel: +49 69 6301 7680; fax: +49 69 6301 5712; e-mail: c.stephan@em.uni-frankfurt.de

Abstract

Background

Individual randomized trials of first-line antiretroviral treatment do not consistently show an association between higher baseline HIV-1 RNA and lower efficacy.

Methods

A MEDLINE search identified 21 HIV clinical trials with published analyses of antiretroviral efficacy by baseline HIV-1 RNA, using a standardized efficacy endpoint of HIV-1 RNA suppression <50 copies/mL at week 48.

Results

Among 21 clinical trials identified, eight evaluated only nonnucleoside reverse transcriptase inhibitor (NNRTI)-based combinations, eight evaluated only protease inhibitor-based regimens and five compared different treatment classes. Ten of the trials included tenofovir (TDF)/emtricitabine (FTC) as only nucleoside reverse transcriptase inhibitor (NRTI) backbone, in addition but not restricted to abacavir (ABC)/lamivudine (3TC) (n = 7), zidovudine (ZDV)/3TC (n = 4) and stavudine (d4T)/3TC (n = 1). Across trials, the mean percentage of patients achieving HIV-1 RNA < 50 copies/mL at week 48 was 81.5% (5322 of 6814) for patients with baseline HIV-1 RNA < 100 000, vs. 72.6% (3949 of 5556) for patients with HIV-1 RNA > 100 000 copies/mL. In the meta-analysis, the absolute difference in efficacy between low and high HIV-1 RNA subgroups was 7.4% [95% confidence interval (CI) 5.9–8.9%; P < 0.001]. This difference was consistent in trials of NNRTI-based treatments (difference = 6.9%; 95% CI 4.3–9.6%), protease inhibitor-based treatments (difference = 8.4%; 95% CI 6.0–10.8%) and integrase or chemokine (C-C motif) receptor 5 (CCR5)-based treatments (difference = 6.0%; 95% CI 2.1–9.9%) and for trials using TDF/FTC (difference = 8.4%; 95% CI 6.0–10.8%); there was no evidence for heterogeneity of this difference between trials (Cochran's Q test; not significant).

Conclusions

In this meta-analysis of 21 first-line clinical trials, rates of HIV-1 RNA suppression at week 48 were significantly lower for patients w ith baseline HIV-1 RNA > 100 000 copies/mL (P < 0.001). This difference in efficacy was consistent across trials of different treatment classes and NRTI backbones.

Introduction

There are several factors associated with the success of first-line antiretroviral treatment, including baseline CD4 count, pre-existing drug resistance, hepatitis C virus (HCV) coinfection and adherence [1, 2]. In multivariate analyses of cohort studies, the baseline level of HIV-1 RNA has been shown to be an independent predictor of full HIV RNA suppression [1, 3]. However, inconsistent results have been obtained in randomized clinical trials. Some trials of first-line antiretroviral treatment have shown lower response rates for patients with baseline HIV-1 RNA ≥ 100 000 HIV-1 RNA copies/mL [4-8], whereas other trials have shown no effects of baseline HIV-1 RNA on efficacy [9-11].

Many randomized clinical trials of first-line antiretroviral treatment are stratified for baseline HIV-1 RNA, and the analyses of efficacy by baseline HIV-1 RNA are normally pre-planned. Individual clinical trials are normally powered to detect differences in HIV-1 RNA suppression for the entire trial population (or noninferiority) [12], but may not be large enough to detect differences in efficacy by baseline HIV-1 RNA level. A meta-analytic approach should maximize the power to analyse the effects of baseline HIV-1 RNA on the efficacy of antiretroviral treatment.

Patients with HIV-1 RNA ≥ 100 000 copies/mL may take longer to achieve full HIV RNA suppression, which might increase the chance of treatment-emergent drug resistance on certain antiretrovirals [13]. In addition, high baseline viral load may be a marker of advanced HIV disease, including low CD4 counts. The rate of decline in HIV-1 RNA during initial treatment can differ markedly between treatment classes [14], but the clinical significance of these differences is unknown [15].

The purpose of this systematic review was to analyse the effects of baseline HIV-1 RNA levels on the 48-week efficacy of different types of antiretroviral treatment in first-line clinical trials. We used a standardized efficacy endpoint – the percentage of patients with HIV-1 RNA levels <50 copies/mL at week 48 calculated using the time to loss of virological response (TLOVR) algorithm of the Food and Drug Administration (FDA) [16] or the nearest equivalent, the FDA ‘Snapshot’ algorithm [17]. The purpose of the main statistical analysis was to determine whether the association between baseline HIV-1 RNA and treatment efficacy differed between the antiretroviral treatment classes and/or NRTI backbones used.

Methods

Study selection

A systematic MEDLINE search was conducted for prospective clinical trials of antiretroviral regimens in antiretroviral-naïve HIV-1-infected individuals published between 1 January 2000 and 1 March 2012. This search used the generic names of each antiretroviral, followed by ‘clinical trial’ and ‘naïve’.

The search was further extended by a review of the proceedings and abstract books of the following international scientific conferences, organized during the above-mentioned index period: the Conference on Retroviruses and Opportunistic Infections (CROI), the Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), the European Conference on Clinical Aspects and Treatment of HIV Infection (EACS), the International AIDS Conference (also known as the ‘World AIDS Conference’), the International AIDS Society (IAS) Conference on HIV Pathogenesis and Treatment, the International Conference on Drug Therapy in HIV Infection (ICDT) and the Annual Meeting of the Infectious Diseases Society of America (IDSA).

Finally, the latest FDA-approved package inserts for each antiretroviral currently licensed for the treatment of HIV-1 infection in treatment-naïve, HIV-infected adults were examined and listed trials were reviewed.

Trials derived from this systematic review of public domain data and conference presentations were included in this analysis if they met all of the following eligibility criteria.

  • They had to include at least 50 chronically infected treatment-naïve, HIV-1-infected individuals aged 16 years or above at any stage of HIV infection.
  • The minimum duration of follow-up reported for these trials at the moment of inclusion in the systematic review had to be 48 weeks.
  • Efficacy data had to be reported for the 48-week time-point using the TLOVR algorithm or the FDA Snapshot algorithm for the virological response (the precentage of patients with a plasma viral load HIV-1 RNA < 50 copies/mL) [16, 17].
  • They had to evaluate, in at least one treatment arm, highly active antiretroviral therapy (HAART) regimens comprising two nucleoside reverse transcriptase inhibitors (NRTIs) plus either a nonnucleoside reverse transcriptase inhibitor (NNRTI) (nevirapine, efavirenz, etravirine or rilpivirine), a boosted protease inhibitor (PI) (ritonavir-boosted darunavir, lopinavir, atazanavir or fosamprenavir), an integrase inhibitor (raltegravir or elvitegravir/cobicistat) or a chemokine (C-C motif) receptor 5 (CCR5) antagonist (maraviroc).

In each trial, treatment success was defined as HIV-1 RNA suppression <50 copies/mL at week 48, using the TLOVR algorithm. With the TLOVR algorithm, treatment failure is virological rebound at any time up to week 48, discontinuation of treatment for adverse events or discontinuation for other reasons [16]. This was the primary efficacy endpoint in most of the clinical trials selected. In more recent clinical trials, the TLOVR method has been replaced by the FDA ‘Snapshot’ algorithm. However, this new endpoint provides very similar estimates of response rate, compared with the TLOVR method [17]. The FDA ‘Snapshot’ algorithm evaluates HIV RNA response using only the results at the week 48 time-point – rebound at earlier time-points is not classified as treatment failure, unless it leads to discontinuation of randomized medication. However, patients who discontinue before week 48 for any reason are still classified as treatment failures using the FDA Snapshot method, consistent with the TLOVR algorithm.

Data collection

Information on the following trial characteristics and results was abstracted for each trial: (1) trial design, treatment regimens compared in the trial and the daily dosages of their components; (2) baseline characteristics: the number of individuals enrolled, the percentage of participants who were male, the percentage who were Caucasian, log10 plasma HIV RNA and CD4 cell counts; and (3) response rates: the percentages of patients with plasma HIV RNA < 50 copies/mL at 48 weeks in the intent-to-treat (ITT) TLOVR or FDA Snapshot analyses. Abstractions were performed by one reviewer and were confirmed by a second; any discrepancies were reconciled by conference with the study team.

The primary analysis determined the difference in efficacy (HIV-1 RNA < 50 copies/mL at week 48) between the subgroups with baseline HIV-1 RNA < 100 000 and ≥100 000 copies/mL. It was determined whether this difference in efficacy between the subgroups varied among treatment arms according to:

  • the third drug used (NNRTI, PI or integrase/CCR5);
  • the NRTI backbone [tenofovir (TDF)/emtricitabine (FTC) or TDF/lamivudine (3TC) vs. abacavir (ABC)/3TC, zidovudine (ZDV)/3TC or stavudine (d4T)/3TC].

In the estimation of average effects across groups of studies, inverse-variance weights were used. All analyses used the generalized linear models (PROC GLM) procedure in sas v9.1 (SAS Institute, Cary, NC).

Results

The MEDLINE search produced 21 clinical trials [5-7, 9-11, 18-30] with standardized efficacy data at week 48, analysed by baseline HIV-1 RNA (Table 1). Of the trials selected for the analysis, eight were comparisons of NNRTIs (Table 2a), eight were comparisons of boosted PIs (Table 2b) and five were comparisons of an NNRTI with a boosted PI (n = 1), an integrase inhibitor (n = 3) or a CCR5 antagonist (n = 1) (Table 2c).

Table 1. Baseline characteristics of trials included in the meta-analysis
TrialTreatment armnCD4 countHIV RNAMaleWhite
[ref](cells/mL)(log10 copies/mL)(%)(%)
  1. Nucleoside reverse transcriptase inhibitors (NRTIs): TDF, tenofovir; FTC, emtricitabine; 3TC, lamivudine; ZDV, zidovudine; d4T, stavudine; ABC, abacavir. Nonnucleoside reverse transcriptase inhibitors: EFV, efavirenz; NVP, nevirapine; ETR, etravirine; RPV, rilpivirine. Protease inhibitors (Pis): LPV/r, lopinavir/ritonavir; DRV/r, darunavir/ritonavir; ATV/r, atazanavir/ritonavir; FPV/r, fosamprenavir/ritonavir. Integrase inhibitors: RAL, raltegravir; ELV/c, elvitegravir/cobisistat. Chemokine (C-C motif) receptor 5 (CCR5) antagonists: MVC, maraviroc.
  2. qd, once daily; bid, twice daily; xr, extended release; ir, immediate release; ND, no data available.
NNRTI trials      
ASSERTABC/3TC/EFV1922405.08386
[40]TDF/FTC/EFV1932305.18084
CNA30021 ABC(qd)/3TC/EFV3842644.98454
[9]ABC(bid)/3TC/EFV3862594.97954
CNA30024 ABC/3TC/EFV3242674.88051
[44]ZDV/3TC/EFV3252584.88251
EPV2001ZDV/3TC(qd)/EFV2783404.68249
[18]ZDV/3TC(bid)/EFV2763864.77651
VERXVETDF/FTC/NVP(xr)5052304.78577
[19]TDF/FTC/NVP(ir)5082284.78574
SENSE2 NRTIs/ETR793194.88592
[20]2 NRTIs/EFV782734.87785
THRIVE2 NRTIs/RPV3402635.07461
[5]2 NRTIs/EFV3382635.07260
ECHOTDF/FTV/RPV3462405.07762
[6]TDF/FTC/EFV3442575.08060
PI trials      
Abbott 730TDF/FTC/LPV/r(qd)3332164.98078
TDF/FTC/LPV/r(bid)3312155.17773
ALERTTDF/FTC/FPV/r531614.97964
TDF/FTC/ATV/r531884.98949
ARTEMISTDF/FTC/DRV/r3432284.97040
TDF/FTC/LPV/r3462184.87044
BMS-089d4T/3TC/ATV1051945.17057
d4T/3TC/ATV/r952014.87353
CASTLETDF/FTC/ATV/r4402055.069ND
TDF/FTC/LPV/r4432045.069ND
HEATABC/3TC/LPV/r3432144.98452
TDF/FTC/LPV/r3451934.88050
KLEANABC/3TC/FPV/r4341885.17861
TDF/FTC/LPV/r4441945.17856
SHAREABC/3TC/ATV/r1112085.19154
Trials of NNRTI vs. PI, integrase or CCR5      
ARTENTDF/FTC/ATV/r1931825.18480
TDF/FTC/NVP3761885.18480
MERITZDV/3TC/MVC3602414.97157
ZDV/3TC/EFV3612544.97255
STARTMRKTDF/FTC/RAL2812195.08141
TDF/FTC/EFV2822175.08244
QUAD 1TDF/FTC/ELV/c3483914.88861
TDF/FTC/EFV3523824.89064
QUAD 2TDF/FTC/ELV/c3503644.99281
TDF/FTC/ATV/r3503754.98978
Table 2. HIV RNA < 50 copies/mL at week 48 by baseline HIV RNA levels. (a) Trials of nonnucleoside reverse transcriptase inhibitor (NNRTI)-based treatments; (b) trials of protease inhibitor (PI)-based treatments; (c) trials of NNRTIs vs. PIs, integrase inhibitors or chemokine (C-C motif) receptor 5 (CCR5) antagonists
(a)
TrialTreatment armHIV RNA < 50 copies/mL at week 48 [n/total (%)]Difference (%)
HIV RNA < 100 000HIV RNA 100 000
copies/mLcopies/mL
ASSERTABC/3TC/EFV61/95 (64%)53/97 (55%)−10
TDF/FTC/EFV62/83 (75%)75/110 (68%)−7
CNA30021 ABC(qd)/3TC/EFV141/217 (65%)112/167 (67%)+2
ABC(bid)/3TC/EFV145/217 (67%)116/167 (69%)+2
CNA30024 ABC/3TC/EFV142/198 (72%)84/126 (67%)−5
ZDV/3TC/EFV140/199 (70%)84/126 (67%)−3
EPV2001ZDV/3TC(qd)/EFV119/202 (59%)45/76 (59%)0
ZDV/3TC(bid)/EFV129/196 (66%)38/80 (48%)−18
VERXVETDF/FTC/NVP(xr)267/311 (86%)142/194 (73%)−13
TDF/FTC/NVP(ir)240/303 (79%)144/203 (71%)−8
SENSE2 NRTIs/ETR40/52 (77%)20/27 (74%)−3
2 NRTIs/EFV40/51 (78%)18/27 (67%)−12
THRIVE2 NRTIs/RPV170/187 (91%)121/153 (79%)−12
2 NRTIs/EFV140/167 (84%)136/171 (80%)−4
ECHOTDF/FTC/RPV162/181 (90%)125/165 (76%)−14
TDF/FTC/EFV 136/163 (83%)149/181 (82%)−1
(b)
TrialTreatment armHIV RNA < 50 copies/mL at week 48 [n/total (%)]Difference (%)
HIV RNA < 100 000HIV RNA ≥ 100 000
copies/mLcopies/mL
Abbott 730TDF/FTC/LPV/r (qd)106/138 (77%)137/193 (71%)−6
TDF/FTC/LPV/r (bid)135/171 (79%)115/160 (72%)−7
ALERTTDF/FTC/FPV/r13/29 (45%)17/24 (71%)+26
TDF/FTC/ATV/r26/29 (90%)18/24 (75%)−15
ARTEMISTDF/FTC/DRV/r195/226 (86%)92/117 (79%)−7
TDF/FTC/LPV/r192/226 (85%)80/120 (67%)−18
BMS-089d4T/3TC/ATV48/55 (87%)25/40 (63%)−25
d4T/3TC/ATV/r41/50 (82%)32/55 (58%)−24
CASTLETDF/FTC/ATV/r178/217 (82%)165/223 (74%)−8
TDF/FTC/LPV/r177/218 (81%)162/225 (72%)−9
HEATABC/3TC/LPV/r133/188 (71%)98/155 (63%)−8
TDF/FTC/LPV/r142/205 (69%)91/140 (65%)−4
KLEANABC/3TC/FPV/r132/197 (67%)154/237 (65%)−2
TDF/FTC/LPV/r134/209 (64%)155/235 (66%)−2
SHAREABC/3TC/ATV/r 37/49 (76%)48/62 (77%)+2
(c)
TrialTreatment armHIV RNA < 50 copies/mL at week 48 [n/total (%)]Difference (%)
HIV RNA < 100 000HIV RNA ≥ 100 000
copies/mLcopies/mL
  1. TDF, tenofovir; FTC, emtricitabine; 3TC, lamivudine; ZDV, zidovudine; d4T, stavudine; ABC, abacavir; EFV, efavirenz; NVP, nevirapine; ETR, etravirine; RPV, rilpivirine; LPV/r, lopinavir/ritonavir; DRV/r, darunavir/ritonavir; ATV/r, atazanavir/ritonavir; FPV/r, fosamprenavir/ritonavir; RAL, raltegravir; ELV, elvitegravir; MVC, maraviroc.
  2. qd, once daily; bid, twice daily; xr, extended release; ir, immediate release.
ARTENTDF/FTC/ATV/r71/78 (91%)71/115 (62%)−29
TDF/FTC/NVP114/146 (78%)150/230 (65%)−13
MERITZDV/3TC/MVC151/211 (72%)99/150 (66%)−6
ZDV/3TC/EFV142/204 (70%)93/156 (60%)−10
STARTMRKTDF/FTC/RAL111/127 (87%)130/154 (84%)−3
TDF/FTC/EFV114/139 (82%)116/143 (81%)−1
QUAD 1TDF/FTC/ELV207/230 (90%)99/118 (84%)−6
TDF/FTC/EFV201/236 (85%)95/116 (82%)−3
QUAD 2TDF/FTC/ELV187/201 (93%)129/152 (85%)−8
TDF/FTC/ATV/r192/213 (90%)116/142 (82%)−8

In addition, there were three trials [31-34] with data analysed using the standardized efficacy endpoints, but at different time-points (weeks 24, 96 and 144). These are included in Table 3.

Table 3. Summary results from studies not included in the meta-analysis
Trial [ref]Treatment armHIV RNA < 50 copies/mL at week 48 [n/total (%)]Difference (%)
HIV RNA < 100 000HIV RNA ≥ 100 000
copies/mLcopies/mL
  1. TDF, tenofovir; FTC, emtricitabine; 3TC, lamivudine; ZDV, zidovudine; d4T, stavudine; EFV, efavirenz; RAL, raltegravir.
  2. qd, once daily; bid, twice daily.
QDMRK [31]TDF/FTC/RAL(qd)205/230 (89%)113/152 (74%)−15
Week 24TDF/FTC/RAL(bid)215/234 (92%)128/152 (84%)−6
Gilead 903 [32]TDF/3TC/EFV128/161 (80%)104/138 (75%)−5
Week 96d4T/3TC/EFV127/172 (74%)95/129 (74%)0
Gilead 934 [34]TDF/FTC/EFV67/107 (63%)79/120 (66%)+3
Week 144ZDV/3TC/EFV67/118 (57%)63/113 (56%)−1

Other trials were excluded from the meta-analysis: several trials had not been analysed by baseline HIV-1 RNA level [35, 36]. Others had not used the standardized TLOVR or FDA Snapshot algorithms and had not used the 50 copies/mL cut-off for the analysis [37-39].

For the 21 clinical trials included in the meta-analysis, the median baseline CD4 counts ranged from 161 to 391 cells/uL, and mean baseline log10 HIV-1 RNA levels ranged from 4.6 to 5.1 log10 copies/mL (Table 1). Overall, of the 12 370 patients included the analysis, 6814 (55%) had baseline HIV-1 RNA levels <100 000 copies/mL, while 5556 (45%) had baseline levels ≥100 000 copies/mL. The clinical trials recruited mainly male patients. The median percentage of White (Caucasian) patients ranged from 40 to 92% across the trials.

Across the 21 trials included in the meta-analysis, the mean percentage of patients achieving HIV-1 RNA < 50 copies/mL at week 48 was 81.5% (5322 of 6814) for patients with baseline HIV-1 RNA < 100 000 vs. 72.6% (3949 of 5556) for patients with HIV-1 RNA ≥ 100 000 copies/mL. Across the three types of trial included (Table 2a–c), most treatment arms showed a numerically lower response rate for patients with baseline HIV-1 RNA levels ≥100 000 copies/mL.

In the meta-analysis, when data were stratified by the third agent or the NRTI backbone (Fig. 1), the absolute difference in efficacy between the low and high HIV-1 RNA subgroups was 7.3% [95% confidence interval (CI) 5.9% to 8.9%; P < 0.001]. This difference in efficacy between subgroups was consistent in trials of NNRTI-based treatments (difference = 6.9%; 95% CI 4.3% to 9.6%), PI-based treatments (difference = 8.4%; 95% CI 6.0% to 10.8%) and integrase Inhibitor- or CCR5-based treatments (difference = 6.0%; 95% CI 2.1% to 9.9%). The difference in efficacy was maintained when only the seven trials using TDF/FTC as the NRTI backbone were included (difference = 8.4%; 95% CI 6.0% to 10.8%). There was no evidence for heterogeneity of this difference between classes of drugs used or NRTI backbones (Cochran's Q tests; not significant).

Figure 1.

Difference in HIV RNA suppression rates at week 48 between patients with HIV RNA < 100 000 copies/mL and those with HIV RNA > 100 000.

Of the six treatment arms in three trials with equivalent analyses run at different time-points (Table 3), there were numerically lower response rates for patients with baseline HIV-1 RNA levels ≥100 000 copies/mL in four treatment arms (Table 3).

Discussion

In this meta-analysis of 21 randomized trials, the percentage of patients with HIV-1 RNA < 50 copies/mL at week 48 was significantly lower for patients with baseline HIV-1 RNA levels >100 000 copies/mL (P < 0.001). This difference in efficacy was consistent across trials of different NRTI backbones and treatment classes (NNRTIs, PIs, integrase inhibitors and CCR5 antagonists).

This analysis used standardized efficacy endpoints for analysis of HIV-1 RNA at week 48 – this endpoint was typically the primary efficacy parameter for analysis of each clinical trial. The meta-analytic approach maximizes the statistical power to analyse the association between baseline HIV-1 RNA and response rates. Individual clinical trials may be too small to show these associations.

There are several limitations to this analysis. First, the 100 000 copies/mL cut-off was used for this analysis because the published results were presented in this way. In the clinical trials selected, the median baseline HIV-1 RNA levels were often close to 100 000 copies/mL. However, there may be other more important cut-off levels for analysing response, above or below the 100 000 copy level. Secondly, the TLOVR and FDA Snapshot endpoints used in this analysis include discontinuation of randomized treatment as failure. If possible, this analysis should be repeated using only virological endpoints. This type of analysis has been conducted in trials from the AIDS Clinical Trials Group (ACTG), and has also shown effects of baseline HIV-1 RNA on long-term suppression rates [37, 38]. Thirdly, it was not possible to determine whether high baseline HIV-1 RNA levels were independently predictive of poor response rates, or were a marker for other factors, such as HCV coinfection, and poor adherence to trial medication. In a recent analysis of ACTG trials, baseline HIV-1 RNA levels were found to be independently predictive of long-term HIV-1 RNA suppression [2].

There was a wide range of HIV-1 RNA suppression rates at week 48 in different trials, even looking at the same treatment. For example, among patients randomized to TDF/FTC/efavirenz (EFV) with baseline HIV-1 RNA levels ≥100 000 copies/mL, the percentage with HIV-1 RNA suppression <50 copies/mL at week 48 ranged from 68% in the ASSERT trial [40] to 82% in the QUAD 1 trial [29]. This analysis was not designed to compare response rates between treatments in different trials, which could differ in baseline characteristics, trial conduct and adherence to randomized treatment. However, there was a consistent association between baseline HIV-1 RNA levels and response rates, across this wide range of trials.

Several recently reported trials have shown differences in efficacy between treatments only for patients with baseline HIV-1 RNA levels ≥100 000 copies/mL. For example, the QDMRK trial, shown in Table 3, showed higher response rates for twice daily raltegravir over once daily raltegravir for patients with HIV-1 RNA ≥ 100 000 copies/mL, but not for patients with lower baseline levels [31]. Similarly, the ACTG 5202 trial was modified by the Data Safety Monitoring Board, after first-line ABC/3TC-based treatments showed significantly higher rates of virological failure among patients with baseline HIV-1 RNA ≥ 100 000 copies/mL compared with TDF/FTC-based treatments [29]. Some treatment guidelines restrict the recommended use of certain antiretrovirals to patients with lower pretreatment HIV-1 RNA levels [41-43].

If there is an excess risk of resistance for people with high baseline HIV RNA levels when taking certain treatment classes, one option would be to start treatment with boosted PIs, which are associated with the lowest rates of treatment-emergent drug resistance [13]. Patients could then be switched to simplified treatments, such as fixed-dose combinations of NRTIs and NNRTIs, when the HIV-1 RNA was fully suppressed and the risk of treatment-emergent drug resistance lower.

In summary, this meta-analysis of clinical trials shows a consistent association between higher baseline HIV-1 RNA levels and a higher risk of treatment failure. This analysis underscores the need for robust and simple initial treatment strategies for patients with high baseline HIV-1 RNA levels.

Acknowledgements

We wish to thank all the patients and their families, as well as all study groups contributing to the published reports that were evaluated for this meta-analysis.

Funding and conflicts of interest: This work was both supported by an unrestricted study grant from Janssen to MetaVirology Ltd and investigator sponsored. CS received no financial support for this project, but a travel grant for presentation of this study at the World AIDS Conference in Washington DC, USA in July 2012. CS and AH have served in the past as independent advisors for Janssen Company (responsible for antiretroviral compounds including darunavir, rilpivirine and etravirine in Europe). YvD and CM are employees of Janssen, and AH is the employer of WS. The Johann Wolfgang Goethe University has received integration grants and is a partner of the European AIDS Treatment Network (NEAT), a project funded by the European Union under the 6th Framework programme (contract number: LSHP-CT-2006-037570). NEAT had no role in study design, in the collection, management, analysis or interpretation of the data, in the preparation, review or approval of the manuscript, or in the decision to submit the article for publication.

Ancillary