HIV Articles  
Clinical Use of HIV Integrase Inhibitors: A Systematic Review and Meta-Analysis
  Download the PDF here
PLOS One Jan 2013
Peter Messiaen1, Annemarie M. J. Wensing2, Axel Fun2, Monique Nijhuis2, Nele Brusselaers1, Linos Vandekerckhove1* 1 General Internal Medicine and Infectious Diseases, Ghent University Hospital, Ghent, Belgium, 2 Virology, Department of Microbiology, University Medical Center Utrecht, Utrecht, The Netherlands

Optimal regimen choice of antiretroviral therapy is essential to achieve long-term clinical success. Integrase inhibitors have swiftly been adopted as part of current antiretroviral regimens. The purpose of this study was to review the evidence for integrase inhibitor use in clinical settings.
MEDLINE and Web-of-Science were screened from April 2006 until November 2012, as were hand-searched scientific meeting proceedings. Multiple reviewers independently screened 1323 citations in duplicate to identify randomized controlled trials, nonrandomized controlled trials and cohort studies on integrase inhibitor use in clinical practice. Independent, duplicate data extraction and quality assessment were conducted.
48 unique studies were included on the use of integrase inhibitors in antiretroviral therapy-naive patients and treatment-experienced patients with either virological failure or switching to integrase inhibitors while virologically suppressed. On the selected studies with comparable outcome measures and indication (n = 16), a meta-analysis was performed based on modified intention-to-treat (mITT), on-treatment (OT) and as-treated (AT) virological outcome data. In therapy-naive patients, favorable odds ratios (OR) for integrase inhibitor-based regimens were observed, (mITT OR 0.71, 95% CI 0.59-0.86). However, integrase inhibitors combined with protease inhibitors only did not result in a significant better virological outcome. Evidence further supported integrase inhibitor use following virological failure (mITT OR 0.27; 95% CI 0.11-0.66), but switching to integrase inhibitors from a high genetic barrier drug during successful treatment was not supported (mITT OR 1.43; 95% CI 0.89-2.31). Integrase inhibitor-based regimens result in similar immunological responses compared to other regimens. A low genetic barrier to drug-resistance development was observed for raltegravir and elvitegravir, but not for dolutegravir.
In first-line therapy, integrase inhibitors are superior to other regimens. Integrase inhibitor use after virological failure is supported as well by the meta-analysis. Careful use is however warranted when replacing a high genetic barrier drug in treatment-experienced patients switching successful treatment.
Since the first reports on Acquired Immunodeficiency Syndrome (AIDS), the human immunodeficiency virus (HIV) has caused a devastating pandemic with yearly 2.6 million new infections worldwide [1]. The stable integration of the reverse transcribed viral genome into host chromatin forms an important point-of-no-return during HIV infection. Raltegravir is the first representative of a new class of antiretroviral drugs targeting the strand transfer reaction during this integration process. Strand transfer integrase inhibitors bind in the catalytic core domain of the enzyme and compete for binding with host DNA. Introduction of raltegravir in 2008 appeared almost simultaneously with approval of second generation drugs of existing therapeutic classes as the protease inhibitor (PI) darunavir and the non-nucleoside reverse transcriptase inhibitor (NNRTI) etravirine. Combined use of these drugs has resulted in high levels of virological suppression in treatment-experienced populations [2], [3]. As a result, the treatment goals in highly experienced patients have been redefined towards successful suppression of plasma viral load [4], [5]. In addition to high efficacy, the initial use of this first integrase inhibitor (INI) also suggested good tolerability, a favorable safety profile and absence of significant drug-drug interactions. Following this success, raltegravir has been explored in a divergent setting of clinical indications such as therapy-naive populations, once-daily formulations, simplification strategies, nucleoside/nucleotide reverse transcriptase inhibitors sparing regimens and maintenance therapy. Conflicting results were reported in several clinical situations, hampering uniform conclusions for successful use of raltegravir. Meanwhile other INIs with a similar mechanism of action such as elvitegravir and dolutegravir have been clinically evaluated. Elvitegravir has been approved in the US and dolutegravir has entered advanced stages of clinical development (Table 1). The objective of this study was to perform a systematic review and meta-analysis of current evidence regarding the use of integrase inhibitors in various clinical settings.
Systematic review

The systematic review resulted in 48 eligible studies on the clinical use of integrase inhibitors, of which 15 abstract-only reports (Figure 1). These studies include in total more than 9400 HIV-infected patients. Of these studies, 38 (79%) described interventions regarding raltegravir use. Elvitegravir and dolutegravir were respectively investigated in 5 (10%) studies each. The average study population size was 202 (IQR 28-222), the average study duration 48 weeks (IQR 24-48). All but four of the included studies were prospective, the majority randomized (59%) and multi-centered (59%). Blinding was performed in 48% of the studies, 20 studies were single-armed (Figure 2). Study characteristics of all studies with latest result updates and evidence levels per category can be found in Table 2, the studies and data used in the meta-analysis are listed in Table 3.
Subsequently a meta-analysis of virological outcome (number of patients achieving HIV RNA below 50 copies/ml) was performed on the 16 controlled studies that compared an INI-based regimen with placebo or other drug classes for similar indications and in which similar endpoints could be evaluated (same measures and same available time-point results). This resulted in three subcategories (treatment-naive patients, treatment-experienced patients with virological failure and patients switching successful suppressive therapy) and the exclusion of studies on treatment intensification, due to the absence of comparable endpoints. The results of the meta-analysis are visualized in Forest plots (Figure 3 and Figure S1). Low heterogeneity in the outcome was seen in the treatment-naive subgroup (mITT, I2 0.0%) and the patients switching successful suppressive therapy group (mITT, I2 23.6). Higher heterogeneity was seen in the studies for patients experiencing virological failure (mITT, I2 83.7%), which points to a higher inter-study variation on virological outcome (Figure S2).
Clinical outcome in antiretroviral-naive patients
Based on our pre-defined criteria for meta-analysis, we included ten studies on treatment naïve patients. Overall, INI based regimens showed a better virological outcome, which reached significance in the mITT analysis (OR 0.71, 95% CI 0.59-0.86; Figure 3A) and OT analysis (OR 0.63, 95% CI 0.47-0.84; Figure S1A). The meta-analysis using AT data (OR 0.86, 95% CI 0.61-1.22; Figure S1A) showed a similar but non-significant favourable trend for INI-based regimens. For one study, no OT or AT data could be obtained [13], for another study AT data were lacking [14].
Comparison of INI versus NNRTI both in combination with dual NRTI.
A sub-analysis of the virological outcome data at 48 weeks comparing INI versus NNRTI showed an OR favoring INIs over efavirenz in the mITT meta-analysis (OR 0.67, 95% CI 0.54-0.84) and OT meta-analysis (OR 0.59, 95% CI 0.43-0.81).
In STARTMRK, raltegravir twice-daily (n = 281) was compared to once-daily efavirenz (n = 282) with a backbone of tenofovir/emtricitabine [15]-[19]. Raltegravir showed non-inferiority based on the primary virological endpoint from 48 up to 240 weeks (mITT 48 week treatment difference +4.2%, 95% CI -1.9 to 10.3). Moreover viral decline in the early treatment phase was significantly more rapid in the raltegravir arm. In the rare cases resistance was observed, multiple raltegravir resistance associated mutations were detected (Table S1). In Protocol 004, an initial dose-ranging trial comparing raltegravir (n = 160) to efavirenz with tenofovir/lamivudine (n = 38) as backbone, similar virological and immunological results at 48 weeks (mITT) were observed as in STARTMRK at all doses [20]-[22]. Few but high-level raltegravir resistance was detected.
Amongst the studies with raltegravir in antiretroviral-naive patients which could not be incorporated in the meta-analysis, QDMRK, comparing once-daily raltegravir (800 mg qd) versus twice-daily raltegravir (400 mg bd), yields important additional information. Despite high levels of suppression in both arms, the once-daily arm was inferior compared to the twice-daily arm (mITT) [23]. This higher virological failure rate was observed mainly in patients starting with high baseline viral load and low C-through levels at 24 hours. Resistance was rare but more frequent in the once-daily arm. Also not included was the uncontrolled SHIELD study, which evaluated raltegravir in combination with abacavir/lamivudine (n = 35) and reported a high proportion (77%) of patients reaching undetectable viral load at 96 weeks in mITT analysis [24]. In the GS-236-0102 phase 3 study, elvitegravir combined with the booster cobicistat and a backbone of emtricitabine and tenofovir (QUAD) (n = 348) was compared to efavirenz with the same backbone (n = 352) both formulated as single tablet regimens (STR). The QUAD STR showed non-inferiority based on the primary virological endpoints up to 48 weeks (mITT 48 weeks treatment difference: +3.6% CI -1.6 to +8.8%) [25]. As has been reported for studies with raltegravir, a more rapid initial HIV RNA decline with elvitegravir was observed compared to the efavirenz arm. In both arms, similar small proportions of patients developed drug resistance upon therapy failure (both arms n = 8). In case of INI resistance in the QUAD failure group, NRTI resistance was observed as well, while in the comparator arm the detection of resistance was mainly limited to NNRTIs.
In the smaller GS-236-014 phase 2 study, the elvitegravir containing QUAD STR (n = 48) was also compared to an efavirenz containing STR (n = 23) with the same NRTI-backbone [26]. Although more patients in the elvitegravir arm achieved an undetectable viral load after one year of follow-up (mITT +8.4%, 95% CI -8.8 to +25.6), this was not statistically significant. Treatment failures were rare and no drug resistance could be assessed. The INI based regimen was well tolerated and fewer adverse events were reported.
In SPRING-1, a phase II dose-ranging randomized trial, a third INI dolutegravir was evaluated. Three different once-daily dosing arms (n = 51 each) were tested against efavirenz (n = 50) with either abacavir/lamivudine or tenofovir/emtricitabine [27]-[29]. Interim results at 48 weeks of follow-up provide favorable virological outcome in all dolutegravir arms driven by better tolerability (mITT, % of patients with HIV RNA <50 copies/ml: 87% dolutegravir arm versus 82% efavirenz arm). In the few cases of treatment failure, the interim analysis at 48 weeks from SPRING-1 did not detect mutations associated with dolutegravir resistance.
The follow-up study SPRING-2 compared dolutegravir 50 mg (n = 413) versus raltegravir 400 mg (n = 414), both combined with dual NRTI backbone [30]. This study was not included in the meta-analysis since it compared two INIs. After 48 weeks, similar proportions of patients on both INI regimens achieved undetectable viremia (ITT, % of patients with HIV RNA <50 copies/ml: 88% versus 85%). In the dolutegravir treated group, no resistance was detected upon failure, while in the raltegravir treated patients one INI and four NRTI mutations were observed.
The SINGLE trial compared two STR in therapy-naïve patients: dolutegravir combined with abacavir/lamivudine (n = 414) versus efavirenz combined with tenofovir/emtricitabine (n = 419) [14]. A significantly better virological outcome after 48 weeks for the dolutegravir-treated group was reported (mITT treatment difference +7.4%; 95% CI +2.5 to +12.3; p = 0,003), while also a significant better immunological response (CD4 increase 267 cells/μl versus 204 cells/μl). The dolutegravir STR was very well tolerated and no INI or NRTI resistance was detected.
Comparison of INI versus PI both in combination with dual NRTI.
In the GS-236-0103 study, the elvitegravir containing QUAD STR (n = 353) was evaluated against ritonavir boosted atazanavir combined with emtricitabine/tenofovir (n = 355) [31]. At 48 weeks, the QUAD regimen showed non-inferiority based on suppression below 50 copies/ml (mITT treatment difference +3.0% CI -1.9 to +7.8%). In five cases in the QUAD group, resistance was observed versus none in the boosted PI treated group. Of those five, primary INI resistance was seen in four and NRTI resistance in three patients.
Comparison of INI versus dual NRTI both in combination with PI.
In the search for simplification strategies, INI based NRTI-sparing regimens have been explored. A sub-analysis of this simplification approach indicated a favorable but non-significant OR in favor of INI compared to 2 NRTI when both are combined with a PI at 24 weeks using mITT data (OR 0.88, 95% CI 0.52-1.50) (Figure 3A).
In the randomized open-label SPARTAN trial, raltegravir with unboosted atazanavir (n = 63) was compared to ritonavir boosted atazanavir plus tenofovir/emtricitabine (n = 31) [13]. Through week 24, both arms achieved comparable efficacy rates (ITT % of patients with HIV RNA <50 copies/ml: 74.6% versus 63.3%). The higher rates of hyperbilirubinemia with twice-daily atazanavir and an increased development of raltegravir resistance in the INI-treated group, prompted early termination of the study. Upon virological failure, four out of five evaluable raltegravir treated patients developed high-level raltegravir resistance. In the PROGRESS trial, raltegravir with ritonavir boosted lopinavir (n = 101) was compared to ritonavir boosted lopinavir (n = 105) with tenofovir/emtricitabine [32] and reported non-inferiority of the study regimen at reaching HIV RNA <40 copies/ml at week 48 (ITT, 81.2% versus 85.7%; difference -4.5%; 95% CI, -15.1% to 5.9%). Upon virological failure, two out of four evaluable raltegravir treated patients developed high-level raltegravir resistance.
The randomized open-label RADAR trial, where raltegravir with ritonavir boosted darunavir (n = 40) was compared with ritonavir boosted darunavir and tenofovir/emtricitabine (n = 40), reported inferior virological outcome for the raltegravir containing regimen after 24 weeks (mITT 75% versus 82.5% of patients with HIV RNA <50 copies/ml) [33].
The ACTG A5262 trial evaluating raltegravir with ritonavir boosted darunavir (n = 112) could not be included in the meta-analysis due to lack of a control arm [34]. Unexpected high levels of virological failure were observed at 48 weeks (mITT 26% or 28 on 112 subjects). Virologic failure and the emergence of INI resistance upon failure was associated with a baseline viral load >100.000 copies/ml. Of note, in this trial a high percentage of patients harbored NRTI resistance mutations in their viral population at baseline, which may serve as an indicator of undisclosed treatment experience or more extensive archived transmitted resistance than detected using regular pol resistance tests.
Clinical outcome in treatment-experienced patients with virological failure
Data from three studies on ART-experienced patients with virological failure could be incorporated in the meta-analysis. Overall, a statistical significant OR in favour of INI use in this population was observed in the mITT meta-analysis (OR 0.27; 95% CI 0.11-0.66) (Figure 3B). The meta-analyses using OT data (OR 0.28, 95% CI 0.20-0.38) and AT data (OR 0.16, 95% CI 0.04-0.61) included two studies and confirmed this result (Figure S1B). The higher I2 (mITT: 83.7%) and the Funnel plot (Figure S2B) point to a difference in study outcome, most likely influenced by differences in trial design and analyses.
Comparison of INI versus placebo both in combination with optimized background.
In the BENCHMRK 1 and 2 trials comparing raltegravir (n = 459) to placebo (n = 237) in patients experiencing therapy failure, superior and sustained viral suppression was observed up to 96 weeks (mITT, 24 weeks: 62% versus 33% virological suppression, p<0.001) [2], [35]. The majority of virus isolates of the raltegravir failure patients harbored integrase resistance mutations, most of which were already detected by 24 weeks of therapy. In the Protocol 005 study, the efficacy of several raltegravir dosages (200 mg, 400 mg or 600 mg bd) (n = 134) with optimized background versus placebo (n = 45) were evaluated in highly experienced patients with HIV RNA >5000 copies/ml [36], [37]. At 24 weeks, 62.0% in all raltegravir treated groups versus 11.3% in the placebo group reached an undetectable viral load (mITT). Integrase mutations were observed in 35/38 INI treated patients failing therapy.
Comparison of INI versus PI both in combination with optimized background.
In the GS-183-105 dose ranging study, elvitegravir plus optimized background regimen was compared with ritonavir boosted PI plus optimized background regimen (n = 63) in predominantly high PI experienced patients [38]. In the 50 mg qd arm (n = 75) elvitegravir was non-inferior while in the 125 mg arm (n = 73) elvitegravir was superior at reaching successful virological outcomes after 24 weeks (mITT time-weighted average change in log10 HIV-1 RNA (DAVG) treatment difference: -0.42, 95% CI -0.77 to -0.06, p = 0.021). This time-weighted endpoint has not been further validated in other clinical trials. No OT and AT data could be extracted for this study.
A sub-set of additional trials were reviewed but could not be included in the meta-analysis for various reasons. Three uncontrolled trials evaluated raltegravir in treatment-experienced patients in combination with other relatively new compounds. The first study is an open-label study describing the use of raltegravir with etravirine (400 mg bid) and an optimized background in patients experiencing treatment failure with darunavir (n = 20) [39]. 65% of participating patients obtained viral suppression at week 24 (mITT). The second study is ANRS 139 TRIO, which combined raltegravir and etravirine with ritonavir boosted darunavir (n = 103) in highly treatment-experienced patients experiencing virological failure. In this landmark study, the combination of three new compounds resulted in - at that time - fascinating high virological suppression rates (mITT: 86% at 48 weeks) among treatment-experienced patients and these results persisted up to 96 weeks [3], [40]. Lastly, raltegravir in combination with etravirine and maraviroc (n = 28) was tested in an uncontrolled study among patients experiencing therapy failure. This approach resulted in high levels of virological suppression (mITT: % of patients HIV RNA <50 copies/ml: 92% at 48 weeks) and no virological failure [41].
Functional mono-therapy with raltegravir in triple-resistant patients (n = 67) was evaluated in one observational trial and showed high efficacy at 48 weeks (mITT: 64% of patients with HIV RNA <50 copies/ml) [42]. Raltegravir resistance was observed in all patients experiencing virological failure.
The GS-183-0145 study compared once-daily elvitegravir versus twice-daily raltegravir in combination with a fully active ritonavir boosted PI and a second agent in patients with virological failure [43]. This study could not be included in the meta-analysis because of comparison of two INI. Elvitegravir was non-inferior to raltegravir regarding virological response at 48 weeks (mITT). In case of failure, HIV-1 integrase resistance patterns by both drugs were comparable, indicating a similar genetic barrier and cross-resistance between both drugs [44].
The VIKING trials evaluating dolutegravir in raltegravir treatment-experienced patients could not be included in the meta-analysis due to their uncontrolled design [45]-[47]. In VIKING I dolutegravir 50 mg once daily (n = 27) and in VIKING II dolutegravir 50 mg twice daily (n = 24) were applied as functional mono-therapy for 10 days, followed by a subsequent replacement of the failing regimen by an optimized backbone. Use of dolutegravir in these cohorts resulted in a very high level of virological suppression at day 11 (mITT HIV RNA <50 copies/ml: 78% respectively 96%). Persistent viral suppression was observed in 41% respectively 52% of these highly experienced patients at week 24. On 15 paired viral isolates from day 1 and day 11 of VIKING II, 3/15 patients harboured additional raltegravir associated mutations. It could not be distinguished whether those mutations had been selected during initial raltegravir treatment or de novo during subsequent dolutegravir use [47].
Clinical outcome in treatment-experienced patients with suppressed viral load
Three studies on ART-experienced patients switching to INI while virologically suppressed were included in the meta-analysis: the analysis based on mITT data indicated a non-significant unfavourable OR when an antiretroviral drug was switched to an INI (OR 1.43; 95% CI 0.89-2.31) (Figure 3C). The meta-analysis based on AT data (OR 1.73, 95% CI 1.01-2.97) demonstrates inferiority after such switch (Figure S1C).
Comparison of INI versus PI both in combination with dual NRTI.
In the SWITCHMRK 1 and 2 studies (n = 347 both arms), a switch from ritonavir boosted lopinavir towards raltegravir based ART was evaluated compared to continuation of the lopinavir based therapy [48]. Baseline genotyping was not performed as patients had an undetectable viral load at screening. Archived resistance data were not taken into account. The studies were terminated prematurely because non-inferiority of raltegravir to ritonavir boosted lopinavir was not established at week 24 (mITT: treatment difference -6.2%, 95% CI -11.2 to -1.3). The lower success rate in the raltegravir arm was most likely due to inclusion of patients with a history of therapy failure and pre-existing resistance against the NRTI backbone. Patients without previous virological failure had similar virological response rates at week 24 in both arms. The majority of the assessable patients who rebounded on raltegravir-based therapy harbored raltegravir resistant virus.
In SPIRAL, a second open-label trial, a switch of ritonavir boosted PI towards raltegravir (n = 139) versus continuation of the ritonavir boosted PI (n = 134) was evaluated in patients with well documented treatment history and long-term virological suppression [49]. In the mITT analysis at week 32, the switch from any ritonavir boosted PI to raltegravir in patients with undetectable viral load resulted in comparable high rates of virological suppression. Low-level raltegravir resistance was observed in only one patient.
Comparison of INI versus enfuvirtide both with background regimen.
In the EASIER-ANRS 138 trial two switch strategies were compared: one group switched immediately to raltegravir (n = 85), the second group continued the low genetic barrier drug enfuvirtide and switched only after 24 weeks (n = 85). When analyzing the mITT 24 week data, the switch from enfuvirtide to raltegravir in heavily pretreated patients with a viral load <400 copies/ml at inclusion, resulted in similar rates of viral suppression [50]. No raltegravir resistance was detected upon virological failure.
Four smaller observational single-armed studies - hence not include in the meta-analysis - evaluated the switch from enfuvirtide to raltegravir in patients with an undetectable viral load and reported high virological success rates at weeks 16 to 48 [51]-[54].
The ODIS trial evaluated two dosage schemes of raltegravir - not included in the meta-analysis - while switching from a protease inhibitor and found that the 800 mg once daily arm (n = 149) had higher rates of virological failure at 24 weeks compared to 400 mg twice-daily (n = 73). In patients with prior NRTI resistance, significant higher failure rates were seen in both arms [55]. RASTA (Raltegravir Simplification for Toxicity or Adverse events) compared switching to raltegravir 400 mg either with tenofovir/emtricitabine (n = 21) or with abacavir/lamuvidine (n = 19) in patients on PI, NNRTI or NRTI-based therapy with suppressed viral load and found comparable virological suppression rates at 24 weeks. Only one patient experienced therapy failure after switch [56]. Anecdotal data from another small study (n = 20) which could not be included in the meta-analysis, showed high virological suppression up to 48 weeks in 96% of patients following regimen simplification towards a low genetic barrier regimen with raltegravir plus nevirapine (n = 20). Prior to the simplification, these patients were long term suppressed on a regimen containing nevirapine most likely without a history of therapy failure [57].
Although several studies have been performed investigating the intensification effect of adding an INI to a successful regimen, the body of evidence from those studies is graded as insufficient [58]-[66]. The heterogeneous nature of the studies, using different outcome measures to assess clinical outcome, residual immune activation and viral replication, and the duration of intensification makes comparison and inclusion in a meta-analysis impossible.
Five other uncontrolled studies describing a switch to raltegravir and boosted or unboosted PI reported good results but the evidence graded as insufficient [67]-[71]. A varying percentage of participants with an undetectable viral load at start of those studies (63% to 100%), different outcome measures and study duration, all hampered uniform conclusions.
Pooled analysis of immunological efficacy, adverse events and emergence of drug resistance
When assessing the immunological response after start of INIs, the majority of the controlled studies with raltegravir, elvitegravir or dolutegravir indicate a similar median CD4 increase compared to other regimens. However, in therapy-naive patients, GS-236-0102 (48w), SINGLE (48w) and the long-term follow-up of STARTMRK (240w), all reported significantly higher CD4 increments compared to efavirenz-based therapies [14], [19], [25]. In the subgroup of treatment-experienced patients with virological failure, use of raltegravir resulted in significant better immunological outcome in BENCHMRK 1 and 2 compared to placebo (96w) [35]. ODIS reported similar significant results after switching to raltegravir from a boosted PI (24w) [55].
The INIs are generally well tolerated and rarely Grade 3 or 4 treatment-emerging adverse events are reported. Compared to efavirenz, discontinuation from INIs due to clinical adverse events is infrequent, while compared to PIs, less severe and life-threatening laboratory abnormalities are observed. An overview of the major adverse events of all INIs can be found in Table 1.
In case of treatment failure in therapy-naive patients, few but high-level raltegravir and elvitegravir resistance was observed, which often conferred cross-resistance to these drugs. No resistance for dolutegravir in this patient population was detected. When combined with dual NRTI, the occurrence of raltegravir or elvitegravir resistance-associated mutations (RAM) was associated in 50% of cases with resistance to NRTI (Table S1).
We performed a systematic review on all published clinical data concerning integrase inhibitors and subsequently meta-analyses on the virological outcome of those studies which included a controlled arm. Based on the meta-analyses, treatment with INIs in combination with dual NRTI showed to be more beneficial for treatment-naive patients compared to other currently used treatment strategies. Also in treatment-experienced patients with virological failure, use of INIs proved to be beneficial as well. However, in successfully treated patients with a history of therapy failure, switching a high genetic barrier drug towards an INI was not supported. More in depth, the following indications for use of integrase inhibitors can be summarized:
Initial therapy
The meta-analysis shows a significant OR in favor of INI combined with dual NRTI based on mITT and OT data, with a similar favorable trend when AT data are used. As both mITT and OT based meta-analyses show a similar significant OR, the clinical benefit of INIs is not only driven by improved tolerability, but also by higher antiviral efficacy. The non-significance of the AT-based meta-analysis can be due to small differences between OT and AT study populations, or might be influenced by the non-availability of AT data from a large dolutegravir trial.
In recent European and US treatment guidelines, raltegravir with a tenofovir/emtricitabine backbone is listed among the preferred regimens for antiretroviral-naive HIV infected individuals [4], [5], [72]. This is supported by the meta-analyses. Raltegravir showed comparable high virological efficacy compared to efavirenz as first line antiretroviral regimen, but was found to be superior driven by its good toxicity profile and tolerability [17], [18]. Besides its good tolerability, raltegravir has a limited risk for drug-drug interactions [73], [74]. Disadvantages of raltegravir are the non-availability of a single tablet regimen and the twice-daily dosing schedule, as supported by the QDMRK study [23]. Raltegravir showed a low genetic barrier to drug resistance upon failure. The emergence of raltegravir resistance was infrequent, but often of high-level (at least two INI RAMs) and transferring cross-resistance to elvitegravir, confirming resistance profiles observed in earlier vitro studies [75].
More recently developed INIs like elvitegravir and dolutegravir hold promise as part of a single tablet regimen (STR) in first-line therapy. Boosted elvitegravir as part of a STR revealed promising results in two large trials, but caution is needed because of increased INI and NRTI resistance. A similar low genetic barrier to drug resistance upon failure was seen for elvitegravir. Raltegravir and elvitegravir based regimens showed comparable or superior immunological response compared to other regimens. Dolutegravir combined with abacavir/lamuvidine has been the first combination reported to be virologically and immunologically superior compared to an efavirenz-based regimen. No drug resistance was detected suggesting a high genetic barrier to resistance development.
In this patient population, novel treatment strategies have been explored, such as the combination of INI with a PI, sparing the NRTIs. Individual studies are underpowered or failed to show superiority. Also the mITT, OT and AT-based meta-analysis failed to show significant odds ratios in favor of these nucleoside-sparing regimens. For stronger conclusions, more data are needed. Currently a large trial evaluating this concept (NEAT-001, tenofovir/emtricitabine/boosted darunavir versus raltegravir/boosted darunavir) is underway.
Virological failure
The meta-analyses demonstrate convincing evidence for a treatment change towards a regimen containing raltegravir or elvitegravir compared to placebo and PI in PI pre-treated individuals with virological failure, based on mITT as well as OT and AT data. For dolutegravir, no randomized trials were available to include in the meta-analysis. Cross-resistance observed in these studies suggests no additional value for sequential use of raltegravir and elvitegravir. In contrast, dolutegravir has limited cross-resistance to other INIs based on a recent report, which could infer a potential role for this drug in INI-experienced patients depending on the resistance profile [45]-[47], [76], [77]. A timely switch or even interruption of raltegravir or elvitegravir may prevent accumulation of resistance and should be considered in order to maximize the potential effect of dolutegravir [78]. Similar or superior immunological response was observed for raltegravir and elvitegravir based regimens compared to other regimens.
Regimen simplification
Switching from enfuvirtide to raltegravir resulted in high levels of durable suppression in several uncontrolled trials [51]-[54] indicating that substitution of a low genetic barrier component of combination antiretroviral therapy by raltegravir in patients with documented or suspected drug resistance can be safely performed [50]. In contrast, the switch from a high genetic barrier PI towards raltegravir in a similar population resulted in a unfavorable OR in the OT-based meta-analysis, and thus higher levels of therapy failure in the raltegravir arm. When adding the effect of adherence and tolerability (mITT), the unfavorable effect was less evident. Two major studies revealed conflicting results possibly influenced by duration of suppression and documentation of treatment history [48], [49]. The results indicate that when switching virologically suppressed patients, individual patient management is needed to assess history of treatment failure, available resistance profiles and duration of the current suppressive regimens in order to perform a safe switch.
Limitations of the systematic review and meta-analysis
One of the limitations of this systematic review and meta-analysis are potential variations in efficacy between the individual INIs compared in similar settings, inherent to the study methodology. Furthermore, the virological outcome data were obtained following different protocols. However, direct comparison of these methodologies has not revealed major differences in outcome [79]-[81]. The SINGLE trial could not be incorporated in the AT meta-analysis for therapy-naive patients, since the number of patients failing combination antiretroviral therapy during the study were not reported so far. In reviewing the current literature on the clinical use of INIs, it becomes obvious that certain clinical questions cannot be answered because of insufficient evidence due to the lack of controlled studies. One of these gaps concerns effect of treatment intensification. Another gap concerns use of INI during pregnancy, since good tolerability and rapid decline of HIV RNA in the plasma suggests a place for integrase inhibitors in this setting [82]. However, there is no evidence from large trials on efficacy and teratogenicity. Similarly, insufficient data are available to include raltegravir in standard post exposure prophylaxis regimens. With the anticipated arrival of studies with new available INIs these gaps could be closed. Finally, the review was restricted to English-language reports.
The meta-analyses positioned INI as a preferred drug in the setting of treatment-naive and as beneficial addition in treatment-experienced patients with virological failure, based on virological efficacy. Careful use of INI when replacing a high genetic barrier PI is warranted. The perspectives of new single tablet regimens containing elvitegravir or dolutegravir taken in the absence of food restrictions hold promise for broad use in first line regimens. In general, the addition of the integrase inhibitor class to our armamentarium has strengthened cART regimens, and further rational use can preserve future therapeutic options.

  iconpaperstack view older Articles   Back to Top