Abacavir-based triple nucleoside regimens for maintenance therapy in patients with HIV

  • Review
  • Intervention

Authors


Abstract

Background

Regimen simplification can be defined as a change in established effective therapy to reduce pill burden and dosing frequency, to enhance tolerability, or to decrease specific food and fluid requirements. Many patients on suppressive antiretroviral therapy may be considered candidates for a simplification strategy and, among them, those who have achieved virologic suppression. Several clinical trials have evaluated the efficacy of triple nucleoside combination as a simplification therapy in patients who achieved virologic suppression

Objectives

The aim of this review is to combine randomised, controlled trials to examine whether in patients with undetectable viraemia on a Protease inhibitor (PI) based regimen simplification treatment with abacavir (ABC)-based triple-nucleoside combinations has similar rates of efficacy and tolerability compared with a PI regimen or simplification with a NNRTIs (efavirenz-EFV- or nevirapine-NVP) containing regimen. Studies were included if they had at least two of the three interventions, including one 3NRTI arm.

Search methods

Electronic databases and conference proceedings were searched (1996-2012) with relevant search terms without limits to language.

Selection criteria

Randomised controlled trials (RCTs) only are included in this review. Patients population is represented by HIV-infected adult patients treated with a PI-containing regimen (PI or boosted PI),  with undetectable viral load. Patients on a PI-containing regimen had three possibilities: continue the PI regimen or switch to a simplification maintenance regimen, including switch to a NNRTI (EFV or NVP) containing regimen, or switch to a triple-NRTI regimen (ABC-zidovudine-lamivudine)

Data collection and analysis

The primary outcomes were: proportion of patients discontinuing or switching antiretroviral therapy due to virologic failure or to adverse events; death (all cause) and AIDS defining illness; occurrence of myocardial infarction and cardiovascular disease. Secondary outcomes  were: proportion of patients maintaining an undetectable viral load (e.g. HIV-RNA <50 or <400 copies/mm3); change in mean CD4+ cell count; occurrence of lipodystrophy. We applied Cochrane Collaboration tools to assess each individual study for risk for bias.

Main results

We included eight RCT, for a total of 1,610 patients. All the studies included HIV-1 infected patients virologically suppressed after a successful treatment with PI containing ART. Articles included in the analysis were published between 2001 and 2010, and could be classified as low risk of bias trials in most of the domains considered. Overall, there was no significant difference between the participants on triple nucleoside combination and controls, either PI-based or NNRTI based in terms of overall failures, death and AIDS related events, and rates of patients with viral load below the detectability cut-off. For the outcomes discontinuation for adverse events and virologic failures, the RRs were not significant , albeit  being not far from the alpha level of 0.05, thus suggesting a weak evidence of lower incidence of side effects  and an higher incidence of virologic failure in the 3NRTI group compared to controls . Change in lipids and in CD4 cells from baselines were reported in 7 studies, but inconsistency in reporting these data did not allow quantitative analysis. However, all agreed that simplification with ABC had a favourable and significant impact on lipid metabolism compared to control group. An increase in CD4 cells count from baseline was evident in all analysed studies, without significant differences between ABC and controls in individual studies.

Authors' conclusions

The strategy of switching to triple nucleoside regimens shows weak evidence of lower incidence of side effects and a higher incidence of virologic failure in the 3NRTI group compared to controls. Simplification with 3NRTI holds the advantages of preserving other classes of antiretroviral drugs, to lower blood lipids, and to be cost effective and simple to administer.Thus, simplification with triple nucleoside regimens AZT + 3TC + ABC should be still considered for individuals who are unable to tolerate or have contraindications to NNRTI or PI based regimens. Additional data are needed on longer-term efficacy of triple NRTI regimens, particularly on the development of antiretroviral resistance. Though studies in the current review were conducted between 2001 and 2010, the large majority of patients from studies analysed received old PI regimens (e.g., indinavir, ritonavir, nelfinavir, saquinavir) not longer recommended by International Guidelines. Since current guidelines recommend new "lipid -friendly" PI, future studies should compare regimens containing these news PIs to triple NRTI regimens. More realistically, however, there are opportunities to examine these issues in existing cohorts.

Résumé scientifique

Trithérapies comportant trois nucléosides à base d'abacavir pour le traitement d'entretien chez les patients infectés par le VIH

Contexte

La simplification de la posologie peut être définie par un changement au niveau du traitement efficace établi en vue de réduire le fardeau des comprimés et la fréquence d'administration, de renforcer la tolérance ou de réduire les besoins spécifiques en aliments et liquides. De nombreux patients sous traitement antirétroviral suppresseur peuvent être considérés comme candidats à une stratégie de simplification et, entre autres, ceux qui présentaient une suppression virale. Plusieurs essais cliniques ont évalué l'efficacité de l'association de trois nucléosides en tant que traitement simplifié chez les patients qui présentaient une suppression virale.

Objectifs

L'objectif de cette revue est de combiner les essais contrôlés randomisés pour examiner si chez les patients ayant une virémie indétectable traités par un régime à base d'inhibiteur de la protéase (IP), un traitement simplifié par des associations de trois nucléosides à base d'abacavir (ABC) entraîne des taux similaires d'efficacité et de tolérance comparé à un régime contenant un IP ou la simplification avec un régime contenant un inhibiteur non nucléotidique de la transcriptase inverse (INNTI) (éfavirenz-EFV- ou névirapine-NVP). Les études ont été incluses si elles comprenaient au moins deux des trois interventions, incluant un bras sous trithérapie d'INTI.

Stratégie de recherche documentaire

Nous avons effectué une recherche dans les bases de données électroniques et les actes de conférences (1996-2012) avec des termes de recherche appropriés sans limite de langue.

Critères de sélection

Nous avons uniquement inclus des essais contrôlés randomisés (ECR) dans cette revue. La population de patients est représentée par des patients adultes infectés par le VIH traités par un régime contenant un inhibiteur de la protéase (IP ou IP potentialisé), dont la charge virale est indétectable. Les patients traités par un régime contenant un IP avaient trois possibilités : continuer le régime contenant un IP ou passer à un régime d'entretien simplifié, y compris passer à un régime contenant un INNTI (EFV ou NVP), ou passer à une trithérapie d'INTI (ABC-zidovudine-lamivudine)

Recueil et analyse des données

Les principaux critères de jugement étaient : la proportion de patients arrêtant ou passant au traitement antirétroviral en raison de l'échec virologique ou d'événements indésirables ; la mortalité (toutes causes confondues) et la maladie définissant le SIDA ; l'occurrence d'infarctus du myocarde et de maladies cardiovasculaires. Les critères de jugement secondaires étaitent : la proportion de patients maintenant une charge virale indétectable (par exemple ARN-VIH < 50 ou < 400 copies/mm3) ; la variation de la numération lymphocytaire CD4+ moyenne ; la présence de lipodystrophie. Nous avons appliqué l'outil de la Cochrane Collaboration pour évaluer les risques de biais de chaque étude individuelle.

Résultats principaux

Nous avons inclus huit ECR impliquant un total de 1 610 patients. Toutes les études ont inclus des patients infectés par le VIH présentant une suppression virale après un traitement efficace par un traitement antirétroviral (TAR) contenant un inhibiteur de la protéase (IP). Les articles inclus dans l'analyse ont été publiés entre 2001 et 2010, et peuvent être classés dans les essais présentant un faible risque de biais dans la plupart des domaines explorés. Globalement, il n'y avait aucune différence significative ente les participants assignés à l'association de trois nucléosides et les témoins, soit à base d'IP soit à base d'INNTI en termes d'échec global, de mortalité et d'événements liés au SIDA et de taux de patients dont la charge virale est inférieure à la valeur limite pouvant être détectée. Pour les résultats concernant l'arrêt en raison d'événements indésirables et d'échecs virologiques, le RR n'était pas significatif, bien que n'étant pas éloigné du niveau alpha de 0,05, suggérant ainsi des preuves peu probantes de l'incidence plus faible d'effets secondaires et de l'incidence plus élevée d'échecs virologiques dans le groupe sous trithérapie d'INTI comparativement aux témoins. Des variations des niveaux de lipides et de lymphocytes CD4 par rapport aux valeurs initiales ont été rapportées dans 7 études, mais l'incohérence dans les données rapportées n'a pas permis de réaliser une analyse quantitative. Toutefois, tous les chercheurs étaient d'accord pour affirmer que la simplification avec ABC a eu un impact favorable et significatif sur le métabolisme des lipides par rapport au groupe témoin. L'augmentation des numérations lymphocytaires CD4 par rapport aux valeurs initiales était évidente dans toutes les études analysées, sans aucune differénce significative entre ABC et les témoins dans les études individuelles.

Conclusions des auteurs

La stratégie du passage à des régimes de trois nucléosides révèle des preuves peu probantes de l'incidence plus faible d'effets secondaires et de l'incidence plus élevée d'échecs virologiques dans le groupe sous trithérapie d'INTI comparativement aux témoins. La simplification avec des trithérapies d'INTI offre les avantages de préserver d'autres classes de mécaments antirétroviraux, d'abaisser les niveaux de lipides dans le sang et d'être rentables et simples à administrer. Ainsi, la simplification avec des régimes de trois nucléosides AZT + 3TC + ABC doit toujours être envisagée pour les individus qui ne tolèrent pas ou présentent des contre-indications pour les régimes à base d'INNTI ou d'IP. Des données supplémentaires sont nécessaires sur l'efficacité à plus long terme des trithérapies d'INTI, en particulier sur le développement de la résistance antirétrovirale. Même si les études incluses dans la revue actuelle ont été menées entre 2001 et 2010, la grande majorité des patients recrutés dans les études analysées ont reçu des régimes contenant des IP anciens (par exemple indinavir, ritonavir, nelfinavir, saquinavir) qui ne sont plus recommandés par les directives internationales. Puisque les directives actuelles recommandent de nouveaux « IP inoffensifs pour les lipides », les futures études devront comparer des régimes contenant ces nouveaux IP aux trithérapies d'INTI. Cependant, il serait plus réaliste d'exploiter les occasions d'examiner ces questions dans des cohortes existantes.

Plain language summary

Abacavir-based triple nucleoside regimens for maintenance therapy in patients with HIV

Regimen simplification can be defined as a change in established effective therapy to reduce pill burden and dosing frequency, to enhance tolerability, or to decrease specific food and fluid requirements. Many patients on suppressive antiretroviral therapy may be considered candidates for a simplification strategy and, among them, those who have achieved virologic suppression. We have reviewed clinical trials evaluating the efficacy and safety of abacavir-containing triple nucleoside combination as a simplification therapy in HIV-infected adult patients treated with a Protease-Inhibitor (PI)-containing regimen and  with undetectable viral load. Patients on a PI-containing regimen had three possibilities: continue the PI regimen or switch to a simplification maintenance regimen with triple nucleoside combination (abacavir-zidovudine-lamivudine) or with non-nucleoside (efavirenz or nevirapine) containing regimens. The review included 8 RCTs and 1675 HIV infected patients. Simplification with triple nucleoside regimen showed an overall failure rate comparable to that of  continuing  PI regimen or  to simplification with non-nucleoside regimens. Rates of failure due to adverse events with triple nucleoside combinations were lower compared to controls, but the difference was not statistically significant. By contrast, rates of virologic failures   were more frequent with  triple nucleoside combination that with PI or NNRTI, but in both the comparisons the differences were  not statistically significant. Simplification with abacavir had a favourable and significant impact on lipid metabolism compared to control group. Simplification with triple nucleoside regimens should be still considered for individuals who are unable to tolerate or have contraindications to NNRTI or PI based regimens

Résumé simplifié

Trithérapies comportant trois nucléosides à base d'abacavir pour le traitement d'entretien chez les patients infectés par le VIH

La simplification de la posologie peut être définie par un changement au niveau du traitement efficace établi en vue de réduire le fardeau des comprimés et la fréquence d'administration, de renforcer la tolérance ou de réduire les besoins spécifiques en aliments et liquides. De nombreux patients sous traitement antirétroviral suppresseur peuvent être considérés comme candidats à une stratégie de simplification et, entre autres, ceux qui présentaient une suppression virale. Nous avons examiné les essais cliniques évaluant l'efficacité et l'innocuité de l'association de trois nucléosides contenant l'abacavir en tant que traitement simplifié chez les patients adultes infectés par le VIH traités par un régime contenant un inhibiteur de la protéase (IP) et dont la charge virale est indétectable. Les patients traités par un régime contenant un IP avaient trois possibilités : continuer le régime contenant un IP ou passer à un régime d'entretien simplifié avec l'association de trois nucléosides (abacavir-zidovudine-lamivudine) ou des régimes ne contenant pas de nucléosides (éfavirenz ou névirapine). Nous avons inclus dans la revue 8 ECR et 1 675 patients infectés par le VIH. La simplification avec un régime de trois nucléosides a entraîné un taux global d'échecs comparable à celui obtenu en continuant le régime contenant un IP ou par la simplification avec des régimes ne contenant pas de nucléosides. Les taux d'échecs dus aux événements indésirables avec des associations de trois nucléosides ont été inférieurs comparés aux témoins, mais la différence n'était pas statistiquement significative. En revanche, les taux d'échecs virologiques ont été plus fréquents avec l'association de trois nucléosides qu'avec un inhibiteur de la protéase (IP) ou un inhibiteur non nucléotidique de la transcriptase inverse (INNTI), mais dans les deux comparaisons les différences n'étaient pas statistiquement significatives. La simplification avec l'abacavir a eu un impact favorable et significatif sur le métabolisme des lipides par rapport au groupe témoin. La simplification avec des régimes de trois nucléosides doit toujours être envisagée pour les individus qui ne tolèrent pas ou présentent des contre-indications pour les régimes à base d'INNTI ou d'IP.

Notes de traduction

Traduit par: French Cochrane Centre 16th July, 2013
Traduction financée par: Pour la France : Minist�re de la Sant�. Pour le Canada : Instituts de recherche en sant� du Canada, minist�re de la Sant� du Qu�bec, Fonds de recherche de Qu�bec-Sant� et Institut national d'excellence en sant� et en services sociaux.

Ringkasan bahasa mudah

Rejimen tiga nucleoside berasaskan Abacavir untuk terapi penyelenggaraan bagi pesakit HIV

Pemudahan rejimen boleh di definis sebagai perubahan dalam terapi yang sudah terbukti berkesan untuk mengurangkan beban pengambilan pil dan kekerapan dos, meningkatkan toleransi, atau untuk mengurangkan keperluan makanan dan cecair tertentu. Ramai pesakit yang menjalani terapi antiretroviral 'suppressive' boleh dipertimbang sebagai calon untuk strategi pemudahan, termasuklah mereka yang telah mencapai 'suppression' virologi. Kami telah menyemak kajian-kajian klinikal untuk menilai keberkesanan dan keselamatan gabungan tiga nucleoside yang mengandungi abacavir sebagai terapi pemudahan bagi pesakit dewasa HIV yang telah dirawat dengan rejimen yang mengandungi Protease-Inhibitor (PI) dan juga pesakit yang mempunyai beban virus yang tidak dapat dikesan. Pesakit yang menggunakan rejimen yang mengandungi PI ada tiga kemungkinan: teruskan rejimen PI atau tukar kepada rejimen pengekalan pemudahan dengan gabungan tiga nucleoside (abacavir-zidovudine-lamivudine) atau dengan rejimen yang mengandungi bukan-nucleoside (efavirenz or nevirapine). Ulasan ini merangkumi lapan kajian rambang terkawal dan 1675 pesakit yang dijangkiti HIV. Pemudahan dengan regimen tiga nucleoside menunjukkan kadar kegagalan keseluruhan setanding dengan meneruskan regimen PI atau pemudahan menggunakan rejimen bukan nucleoside. Kadar kegagalan akibat kesan sampingan dengan gabungan tiga nucleoside adalah lebih rendah berbanding dengan kumpulan kawalan, tetapi perbezaan itu tidak signifikan secara statistik. Sebaliknya, kadar kegagalan virologi adalah lebih kerap dengan menggunakan gabungan tiga nucleoside berbanding dengan PI atau NNRTI (bukan nucleoside); tetapi perbezaan tidak signifikan secara statistik dalam kedua-dua perbandingan. Pemudahan dengan menggunakan abacavir mempunyai kesan yang baik dan signifikan ke atas metabolisme lemak berbanding dengan kumpulan kawalan. Pemudahan dengan menggunakan rejimen tiga nucleoside harus dipertimbangkan untuk individu yang tidak dapat bertahan atau mempunyai kontraindikasi kepada rejimen berdasarkan NNRTI (bukan neucloside) atau PI.

Catatan terjemahan

Diterjemahkan oleh Raymond Chieng Siang Ching (Melaka Manipal Medical College).Disunting oleh Noorliza Mastura Ismail (Melaka Manipal Medical College) Untuk sebarang pertanyaan mengenai terjemahan ini sila hubungi raymondchieng@gmail.com

Summary of findings(Explanation)

Summary of findings for the main comparison. 
Abacavir-containing triple nucleoside simplification compared with PI or NNRTI containing regimens for HIV infection

Patient or population: HIV-infected adult patients treated with a PI-containing regimen  with undetectable viral load

Settings: Treatment simplification in patients who achieved viral suppression

Intervention: Triple nucleoside combination (Abacavir-Zidovudine-Lamivudine)

Comparison: PI or NNRTI containing ART regimens

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
PI or NNRTI regimensTriple Nucleoside Regimen
Failure (overall) 330 per 1000 284 per 1000
(212 to 369)
RR 0.88 (0.74, 1.04)1606
(8)
⊕⊕⊕⊝
moderate
 
Discontinuation for adverse events 177 per 1000 117 per 1000
(86-157)
RR 0.68 (0.44, 1.07)1597
(8)
⊕⊕⊕⊝
moderate
 
Virologic failure 77 per 1000 102 per 1000
(64 to 160)
RR 1.39 (0.95, 2.02)1587
(8)
⊕⊕⊕⊝
moderate
 
Death and AIDS related events 13 per 1000

25 per 1000

(12 to 49)

RR 1.66 (0.62, 4.42)

1044

(4)

⊕⊕⊝⊝

Low quality

 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk Ratio;
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Because adherence is crucial for the success of antiretroviral treatment, regimens with fewer pills, simpler dosing schedules, and fewer adverse events have become the first choice for antiretroviral therapy. Regimen simplification can be defined broadly as a change in established effective therapy to reduce pill burden and dosing frequency, to enhance tolerability, or to decrease specific food and fluid requirements (DHHS 2008). Exposure to PIs has been associated with metabolic complications (e.g.,dyslipidemia, insulin resistance) hepatotoxicity, gastrointestinal  adverse effects,  risk of cardiovascular events, and potential for drug interactions DAD 2008; DHHS 2011.  Moreover, the decrease in  bone mineral density  has been reported more pronounced in patients receiving  boosted PI-containing regimen compared with NNRTI and NRTIs (Duvivier 2009). Many patients on suppressive antiretroviral therapy may be considered candidates for a simplification strategy and, among them, those who have achieved virologic suppression. Results of a previous meta-analysis show a higher rate of virologic failure in patients switching to an abacavir-based triple nucleoside regimen (abacavir (ABC)-zidovudine (AZT)-lamivudine (3TC)) (Trizivir®) compared to patients continuing a protease inhibitor (PI) regimen (Bucher 2003). Since these results were found, however, several trials comparing simplified maintenance therapy and PI continuation have been published, and the present review includes these studies.

An association between combination antiretroviral therapy (cART) and the risk of myocardial infarction has been observed in some cohort studies. Data from the “Data collection of Adverse events of Anti-HIV Drugs Group” (DAD) study reported an association between cumulative exposure to some PIs and an increased risk of myocardial infarction (DAD 2007; DAD 2008; Lundgren 2009). Current or recent (within the last 6 months) use of the NRTIs abacavir (ABC) and didanosine (ddI) was  associated with an increased risk of cardiovascular disease within the setting of the DAD cohort study and the viral suppression arm of the SMART (DAD 2008; SMART 2008).By contrast, the results of a  case-control study nested within the French Hospital Database on HIV showed that any association of short-term/recent exposure to ABC with risk of MI  was lost when subjects were case-matched to controls by cocaine or intravenous drug use (Lang 2009). Similarly, results from the Veterans Administration’s Clinical Case registry,  show that cumulative exposure to ABC was not associated with significant increase in MI. Moreover, any trend towards an association of ABC use with MI was weaker after adjustment for traditional cardiovascular risk factors including renal function.(Ribaudo 2011). More recently, a modest association was found with the cumulative use of ABC and risk of myocardial infarction (Worm 2010). However, no excess risk of myocardial infarction with ABC therapy has been observed among 14,174 patients in the aggregated clinical trials database maintained by the manufacturer of ABC (Brothers 2009) and in a meta-analysis of 28 RCTs  (9,233 subjects) comparing ABC  containing cART to  other regimens not-containing ABC (Cruciani 2011)

Recent data have offered conflicting evidence concerning the relative effectiveness of NRTI backbone combinations. The ACTG A5202 study showed that among patients with a high baseline viral load (HIV RNA >100,000 copies/mL), those taking abacavir plus lamivudine with either boosted atazanavir or efavirenz were more likely to experience virologic failure than were those taking tenofovir plus emtricitabine (Sax 2008). By contrast,  the HEAT study (Smith 2008), a head-to-head comparison of abacavir plus lamivudine with tenofovir plus emtricitabine, in combination with lopinavir-ritonavir and a retrospective analysis of six previous abacavir trials, found that the backbones had similar efficacy (Pappa 2008).

Description of the condition

With the advent of antiretroviral drugs, especially highly active antiretroviral therapy (HAART), HIV infection has been transformed from an almost uniformly fatal disease to a chronic and manageable condition. The field of HIV medicine continues to evolve rapidly and, to maintain a high standard of care, treatment guidelines must be refined continually to assist the complex decision-making process. Treatment can provide durable virologic, immunologic, and clinical benefits while minimizing toxicities and drug resistance, improving quality of life, and potentially allowing for a normal life span.  

Description of the intervention

The use of PI-sparing regimens has gained interest because many of the older PI-containing regimens often increase fasting cholesterol and triglyceride values and may cause disorders in lipid metabolism. In order to decrease pill burdens, preserve future treatment options, and potentially lower the risk of drug-induced adverse effects, such as lipodystrophy and hyperlipidaemia, many clinicians and investigators have switched patients from a fully suppressive PI-containing regimen to a simplification strategy of NNRTIs or NRTIs (Kessler 2005, Rizzardini 2006). This strategy has been proven effective only in patients who had been virologically suppressed for an extended period at the time of the switch (Clumeck 2001, Katlama 2003, Moyle 2002). One of the most commonly investigated switch strategies has involved replacing a PI with ABC among patients receiving two other NRTIs, resulting in a triple-nucleoside regimen. ABC-AZT-3TC (Trizivir®) is the only triple-NRTI combination for which randomised, controlled trials are available (DHHS 2011). ABC-AZT-3TC demonstrated comparable antiretroviral activity to IDV-based (Staszewsky 2001; Vibhagool 2004) and NFV-based regimens (Kumar 2006), but it was found of inferior virologically activity compared to an EFV-based regimen (Gulick 2004).

How the intervention might work

This combination is generally not recommended as initial ARV regimen and should be used only when a preferred, an alternative, or an acceptable NNRTI-, PI-, or INSTI- based regimen is less desirable because of concerns about toxicities, drug interactions, or regimen complexity (WHO 2010; DHHS 2011). A triple-NRTI combination regimen, however, has multiple potential advantages: fewer drug-drug interactions, low pill burden, and availability of a fixed-dose combination. Moreover, although Trizivir® has decreased activity when used as initial therapy in patients with baseline viral loads >100,000 copies/mL, it appears that once the patient's virus is fully controlled, the combination may provide durable treatment benefit in a simplification or switch strategy, or both (Kessler 2005; Rizzardini 2006).

Why it is important to do this review

A regimen simplification can be defined broadly as a change in established effective therapy to reduce pill burden and dosing frequency, to enhance tolerability, or to decrease specific food and fluid requirements. Many patients on cART may be considered candidates for regimen simplification. Current simplification strategies include within-class simplifications (e.g.,NRTI and PIs substitutions) and out-of-class (commonly a change from a PI-based to an NNRTI-based regimen). However, simplification with triple nucleoside regimens AZT + 3TC + ABC should be considered for individuals who are unable to tolerate or have contraindications to NNRTI or PI based regimens (DHHS 2011). Subsets of patients who may benefit most of triple-nucleoside are those with HIV/TB coinfection, pregnant women, chronic viral hepatitis B, and HIV-2 infection (WHO 2010).

Objectives

The aim of this review is to combine randomised, controlled trials to examine whether in patients with undetectable viraemia simplification treatment with ABC-based triple-nucleoside combinations has similar rates of efficacy and tolerability compared with a continued PI regimen or simplification with NNRTIs (efavirenz or nevirapine) containing regimen. Meanwhile, it offers the opportunity to address the risk of cardiovascular disease in ABC-treated patients and comparators.

Methods

Criteria for considering studies for this review

Types of studies

Randomised, controlled trials (RCTs) only are included in this review.

Types of participants

Chronically HIV-infected adult patients treated with a PI-containing regimen (PI or boosted PI),  with undetectable viral load (HIV-1 RNA  below the cut-off value as defined in individual studies). Study populations include patients starting a first antiretroviral PI-containing regimen as well as antiretroviral-experienced participants on a stable PI-containing regimen.

Types of interventions

Patients on a PI-containing regimen had three possibilities:

  1. Continue the PI regimen or switch to a simplification maintenance regimen, including

  2. Switch to a NNRTI (EFV or NVP) containing regimen, or

  3. Switch to a triple-NRTI regimen (ABC-AZT-3TC (Trizivir®))

1 and 2 are controls, 3 is the experimental intervention. If a trial had all three options, we compared the experimental group to each of the 2 control groups (1 vs 3, and 2 vs 3); in this case, to avoid double-counts in the experimental group, we split the ‘shared’ group (3NRTI) into two groups with smaller sample size, and include two independent  comparisons (Higgins 2011).

Types of outcome measures

Primary outcomes

The following are the primary outcomes for this review.

  • Proportion of patients discontinuing or switching antiretroviral therapy due to virologic failure.

  • Rates of patients with adverse events requiring treatment interruption or switching, or both.

  • Overall rates of treatment interruption or switching, or both.

  • Death (all cause).

  • Occurrence of myocardial infarction and cardiovascular disease.

  • Occurrence of new HIV-related event (death or AIDS-defining illness).

Secondary outcomes

The following are the secondary outcomes for this review.

  • Proportion of patients maintaining an undetectable viral load, as defined by the authors (e.g. HIV-RNA <50 or <400 copies/mm3).

  • Change in mean CD4+ cell count (mean relative change (percent) or mean absolute change, compared with baseline, and standard deviation).

  • Occurrence of lipodystrophy.

  • Quality of life indicators, as reported in the studies.

  • Any adverse events.

Search methods for identification of studies

We followed standard Cochrane systematic review methodology. A comprehensive, unbiased search strategy was developed to ensure that as many relevant studies are found as possible. Relevant studies were sought regardless of language and publication status. The literature search was conducted with the assistance of the Trials Search Coordinator of the Cochrane HIV/AIDS Group . Hand searches of the reference lists of all the relevant reviews and studies found was undertaken. Standard HIV/AIDS Cochrane Collaborative Group search terms were used.  Major subject heading terms included: highly active antiretroviral therapy, antiretroviral agents, and individual drug names (trizivir®, abacavir, ziagen®). An example of search strategy is shown on appendix.(Appendix 1).

Electronic searches

The following electronic journal and trial databases were searched for relevant trials: 1. Cochrane Central Register of Controlled Trials (CENTRAL) 2. Cochrane Database of Systematic Reviews 3. EMBASE 4. PubMed 5. Clinical Trials.gov 6. WHO International Clinical Trials Registry Platform.

Searching other resources

The following electronic database were searched:

AIDSInfo® (http://www.aidsinfo.nih.gov/).

Moreover,  abstracts from the following  international conferences were searched using  the International AIDS society electronic database (available at http://www.iasociety.org/Default.aspx?pageId=7): International AIDS Society Conference on HIV Pathogenesis and Treatment (2001, 2003, 2005, 2007, 2009, 2011) and International AIDS Conference (2002, 2004, 2006, 2008, 2010). Abstracts of the Conference on Retroviruses and Opportunistic Infections (CROI), Interscience Conference on Antimicrobial Agents & Chemptherapy (ICAAC), and The European AIDS Clinical Society (EACS)-2001-2011, were also searched. 

Data collection and analysis

Selection of studies

Two authors (MC, SGP) independently screened the titles and abstracts of references identified by the search for potential relevance. We resolved disagreement or doubt by discussion. We obtained in full any study that potentially met the inclusion criteria based on the title or abstract, or both, and assessed these studies against the inclusion criteria. We recorded the reasons for exclusion of any study previously considered for inclusion

Data extraction and management

The following data were abstracted independently by two reviewers (MC, SGP) from the selected trials using standardised data extraction form:

  •  Administrative details: authors, year of publication.

  •  Details of participants: number, setting, baseline characteristics by group.

  •  Details of the study: study design, country of origin, type and duration of follow up.

  •  Details of the antiretroviral regimen used.

  •  Primary and secondary outcome descriptions and outcome measures.

  •  Number of withdrawals in each group, with reasons for withdrawal.

Assessment of risk of bias in included studies

Two authors (MC, SGP) independently assessed the risk of bias of each trial using a simple form and following the domain-based evaluation as described in the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 (Higgins 2011). We compared the assessment results and discussed any discrepancies between ourselves in order to achieve agreement on the final assessment for each criteria .

We assessed the following domains as '‘Low risk’ of bias, ‘‘Unclear risk’ of bias, or High risk’ of bias.

  • Randomisation

  • Concealment of allocation

  • Blinding (of participants, personnel and outcome assessors)

  • Incomplete outcome data

  • Selective outcome reporting

  • Free of other bias (e.g., baseline imbalance, early stopping).

We used the following definitions:

  • Generation of the allocation sequence

  • "low risk of bias", if the allocation sequence was generated by a computer or random number table. Drawing of lots, tossing of a coin, shuffling of cards, or throwing dice will be considered as adequate if a person who was not otherwise involved in the recruitment of participants performed the procedure.

  • "unclear risk of bias", if the trial was described as randomised, but the method used for the allocation sequence generation was not described.

  • "high risk of bias", if a system involving dates, names, or admittance numbers was used for the allocation of patients.

 Allocation concealment

  • "low risk of bias", if the allocation of patients involved a central independent unit, on-site locked computer, identically appearing numbered drug bottles or containers prepared by an independent pharmacist or investigator, or sealed envelopes.

  • "unclear risk of bias", if the trial was described as randomised, but the method used to conceal the allocation was not described.

  • "high risk of bias", if the allocation sequence was known to the investigators who assigned participants or if the study was quasi-randomised.

 Blinding (or masking)

We assessed each trial (as '‘Low risk’ of bias, ‘‘Unclear risk’ of bias, or High risk’ of bias) with regard to the following levels of blinding:

  • blinding of clinician (person delivering treatment) to treatment allocation;

  • blinding of participant to treatment allocation;

  • blinding of outcome assessor to treatment allocation.

  • Incomplete outcome data

  • "low risk of bias", the numbers and reasons for dropouts and withdrawals in all intervention groups were described or if it was specified that there were no dropouts or withdrawals.

  • "unclear risk of bias", the report gave the impression that there had been no dropouts or withdrawals, but this was not specifically stated.

  • "high risk of bias", the number or reasons for dropouts and withdrawals were not described.

We have further examined the percentages of dropouts overall in each trial and per randomisation arm and evaluated whether intention-to-treat analysis has been performed or could be performed from the published information.

 Were all randomised participants analysed in the group to which they were allocated? (ITT analysis)

  • "low risk of bias": Specifically reported by authors that ITT was undertaken and this was confirmed on study assessment, or not stated but evident from study assessment that all randomised participants are reported/analysed in the group they were allocated to for the most important time point of outcome measurement (minus missing values) irrespective of non-compliance and co-interventions.

  • "high risk of bias": Lack of ITT confirmed on study assessment (patients who were randomised were not included in the analysis because they did not receive the study intervention, they withdrew from the study or were not included because of protocol violation) regardless of whether ITT reported or not

  • ‘As-treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation.

  •  "Unclear risk of bias": Described as ITT analysis, but unable to confirm on study assessment, or not reported and unable to confirm by study assessment.

 Selective outcome reporting

  • "low risk of bias": adequate, pre-defined, or clinically relevant and reasonably expected outcomes are reported on.

  • "Unclear risk of bias": not all pre-defined, or clinically relevant and reasonably expected outcomes are reported on or are not reported fully, or it is unclear whether data on these outcomes were recorded or not.

  • "high risk of bias":inadequate, one or more clinically relevant and reasonably expected outcomes were not reported on; data on these outcomes were likely to have been recorded.

Free of other bias (e.g., Baseline imbalance, early stopping,):

  • "low risk of bias": the trial appears to be free of other components that could put it at risk of bias.

  • "unclear risk of bias": the trial may or may not be free of other components that could put it at risk of bias.

  • "high risk of bias": there are other factors in the trial that could put it at risk of bias, e.g., no sample size calculation made, early stopping, industry involvement, or an extreme baseline imbalance.

Measures of treatment effect

Review Manager 5.1 (RevMan 2011) was used to analyse the data.   For binary outcomes, we calculated the relative risk (RR) with 95% confidence intervals (CI) for each outcome. Virologic outcome were assessed as rates of patients with VL below the predefined cut-off (<50 copies/mL and/or 200-500 copies/mL) at 48 weeks and/or at 96 weeks. Where available, we reported the results for this endpoint  according to the missing=failure (MEF) approach and  to the  time to loss of virologic response (TLOVR) algorithm .Continuous outcomes, such as variation in CD4+ cell counts, are presented as weighted mean differences (WMD) with 95% CIs.

Unit of analysis issues

A unit of analysis was the patient. If data were available on outcome measures at two or more periods, they were analysed according to the different period of time.

Dealing with missing data

We assessed the percentages of dropouts overall for each included trial and per each randomisation arm and we evaluated whether an intention to treat analysis has been performed or could be performed with the available published information.

When additional data were needed, we contacted the corresponding author of each  study by e-mail in order to access further information.When missing data were not available, we limited our analysis to the available data

Assessment of heterogeneity

We used the I² statistic to measure statistical heterogeneity among the trials in each analysis. In the presence of substantial heterogeneity we explored it by pre-specified subgroup analysis. The I2 statistic describes the percentage of total variation across trials that is due to heterogeneity rather than sampling error (Higgins 2003). An heterogeneity if I2 >50% was considered to be statistically significant (Higgins 2003;Higgins 2009; Higgins 2011)

Assessment of reporting biases

To inspect for publication bias visually, we generated graphic funnel plots (Egger 1997).

Data synthesis

 In the case of no heterogeneity (I²=0), studies were pooled using a fixed-effect model. Where values of I² were greater than zero, a random effects analysis was undertaken . Where values of I² greater than 75% indicated a very high level of heterogeneity, we refrained to pool data from different studies and we undertook a narrative overview.

Subgroup analysis and investigation of heterogeneity

We performed subgroup analyses according to the type of drugs in the control group (e.g. nevirapine, efavirenz), timing of simplification to nucleoside therapy (e.g. simplification in patients after a first antiretroviral PI-containing regimen and in antiretroviral-experienced participants). Where available, the frequency and severity of adverse events were compared between groups.

Sensitivity analysis

We conducted a sensitivity analysis to determine how the adequacy of allocation concealment affected the results of the review

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies

Results of the search

Searches were conducted on November 19, 2009, and the last update on July 6, 2012. The search identified 1051 titles (Figure 1). MC and SGP independently conducted the selection of potentially relevant studies by scanning the titles and abstracts. Irrelevant reports were discarded, and the full article was obtained for all potentially relevant or uncertain reports. Studies were reviewed for relevance, based on study design, types of participants, exposures and outcomes measures.Tweenty-eight manuscripts were reviewed by two reviewers (MC, SGP), who independently applied the inclusion criteria. CM acted as arbiter where there was disagreement. Studies were reviewed for relevance, based on study design, types of participants, exposures and outcome measures. Seventeen trials/studies, corresponding to 20 publications, dealing with HIV-1 infected patients with undetectable viraemia after successful ART and switched to triple NRTI combination (co formulation of zidovudine/lamivudine plus abacavir or trizivir®) were identified.

Figure 1.

Flow diagram.

Included studies

We included eight randomised controlled trials (Bonjoch 2005; Clumeck 2001; Katlama 2003; Keiser 2005; Maggiolo 2003; NEFA Study; Opravil 2002; Sprenger 2010). Data of the NEFA study (NEFA Study) were reported in 4 publications (Martínez 2003; Martínez 2007; Ochoa de Echanguen 2005;  Fisac 2005), the primary reference being Martínez 2003. All the studies included HIV-1 infected patients virologically suppressed after a successful treatment with PI containing ART (though in the Katlama 2003 trial and Bonjoch 2005 trial prior treatment was mostly but not exclusively based on PI containing ART) and then switched to triple NRTI combination. Baseline characteristic of the selected studies are summarized in appendix (Table 1). The Katlama 2003 trial included also subsets of patients previously treated with 2 NRTI+ NNRTI (19 %), or with 3 NRTI (17 %), while in the Bonjoch 2005 trial less than 10 % of randomised patients were not receiving PIs at inclusion; all the patients in the remaining studies were naive to NNRTI and abacavir. The Clumeck 2001 study recruited 211 participants from 29 centres in Europe and Canada; participants had median age of 38 years (range 20-66 years), an HIV-1 RNA screening level <50 copies/mL , and median CD4 cell count of 504 cells/mm3 (range 72 to 1565) and 507 cells/mm3 (range 67 to 1457) in the abacavir and PI arms, respectively. The Opravil 2002 study recruited 228 patients from 8 centres in Switzerland and Italy; participants had median age of 41 years (range 27-65) and 38 years (range 23-65 years) in the continuation and simplified arms, respectively; an HIV-1 RNA screening level <50 copies/mL , and median CD4 cell count of 504 cells/mm3 (range 72 to 1565) and 507 cells/mm3 (range 67 to 1457) in the abacavir and PI arms, respectively. The Katlama 2003 trial enrolled 219 patients from 47 centres in 9 European countries; participants had median age of 37 years (range 26-82) and 40 years (range 20-69 years) in the continuation and simplified arms, respectively; an HIV-1 RNA screening level <50 copies/mL , and median CD4 cell count of 504 cells/mm3 (range 133 to 1324) and 482 cells/mm3 (range 134-1246) in the continuation and simplified arms, respectively. The Maggiolo 2003 trial enrolled 209 patients from a single centre in Bergamo (Italy); participants had mean age of 36.1 ± 7.9, 36.6 ± 7.2, and 37.7 ± 5.7 years in the abacavir, efavirenz and PI continuation arms, respectively; an HIV-1 RNA screening level <50 copies/mL , and mean CD4 cell count of 600 ± 248, 597 ± 272, and 578 ± 267 cells/mm3 in the abacavir, efavirenz and PI continuation arms, respectively. The Martinez 2003 trial (NEFA study) enrolled 460 patients from 15 centres in Spain; participants had median age of 39 (range 23-73), 38 (range 21-69), and 40 (range 24-71) years in the nevirapine, efavirenz and abacavir arms, respectively; an HIV-1 RNA screening level <200 copies/mL , and median CD4 cell count of 508, 558, and 544 cells/mm3 in the nevirapine, efavirenz and abacavir arms, respectively. The Bonjoch 2005 trial included 134 participants from several centres in Spain; participants had median age of 39 years in both arms; an HIV-1 RNA screening level <200 copies/mL , and median CD4 cell count of 530 and 554 cells/mm3 in the abacavir and nevirapine arms, respectively. The Keiser 2005 trial included 144 participants from 44 centres in USA; participants had median age of 43 (range 23-64) and 42 (range 25-62) years in simplified and PI-continuation arms, respectively; an HIV-1 RNA screening level <50 copies/mL , and mean CD4 cell count of 551 ± 226 and 531 ± 233 cells/mm3 in the simplified and PI-continuation arms, respectively. The Sprenger 2010 trial (FREE study) enrolled 207 participants from 10 centres in the Netherlands; participants had median age of 40.3 (median 24.7-62.0) and 43.1 (median 21.9-68.6) years in PI-continuation and simplified arms, respectively; an HIV-1 RNA screening level <50 copies/mL , and median CD4 cell count of 207 (range 10-370) and 195 (range 10-437) cells/mm3 in the PI-continuation and simplified arms, respectively. The Bonjoch 2005 trial did not report information on gender. A higher proportion of men than women, ranging from 76 to 90 %, was reported in the remaining studies but one: the Maggiolo 2003 study included 33 % of male patients.

Table 1. Baseline characteristics of included studies
  1. * data are medians; ** data are means. Legenda: PI= protease inhibitors; ABC=abacavir; EFV= efavirenz; NVP= nevirapine; ns= not stated

Study Viral load (copies/ml) Duration (months) of  last ART treatment* CD4 cells/mm3* Age in yrs* Male (%)
Clumeck 2001Undetectability since initiation of therapy and <50 at screening20  in each arm504 (ABC) vs 507 (PI)28 in each arm78 (PI) vs 87 (ABC)
Opravil 2002<400 for ≥6 months and <50 at screening27 (PI) vs 21 (ABC)513 (PI) vs 507 (ABC)41 (PI) vs 28 (ABC)76 (PI) vs 79 (ABC)
Katlama 2003<400 since initiation of therapy and < 50 within 14 days of study entry27 (ABC) vs 24482 (ABC) vs 50440 (ABC) vs 3780 (ABC) vs 83
Maggiolo 2003<50 over the last 6 months and at screening21 ( ABC), 20.6 (EFV), 21 (PI)600(ABC),  597 (EFV),  598 (PI)**36.1 (ABC),  36.7 (EFV), 37.7 (PI)**33 (ABC), 29 (EFV), 39 (PI)
NEFA study<200 for ≥6 months30 (ABC), 31(EFV), 29 (NVP)544 (ABC), 558 (EFV), 508 (NVP)40 (ABC), 38 (EFV), 39 (NVP)76 (ABC), 75 (EFV), 77 (NVP)
Keiser 20052 most recently tests <400 and <50 at screening

ns; on a well-tolerated PI-containing

ART for >3 months

551 (ABC) vs 531 (PI)**43 (ABC) vs 42 (PI)46 (ABC) vs 47 (PI)
Bonjoch 2005<200 for ≥6 monthsns530 (ABC) vs  544 (NVP)39 in each arm82
Sprenger 20102 tests <50 between wks 12 and wks 24 before randomisation≥6 months310 (ABC) vs  350 (PI)43.1 (ABC) vs 40.3 (PI)82 (ABC) vs 88 (PI)

The study by Maggiolo et al. had  the 3 intervention groups (3NRTI, PI, NNRTI), while the NEFA study had  ABC group  compared to nevirapine or efavirenz containing regimens. To avoid double counts in the experimental group, we have  split the 3NRTI in the study by Maggiolo et al. and in the NEFA study for the outcomes overall failure, discontinuation for adverse events, and virologic failure

Excluded studies

Reasons for excluding potentially relevant trials are given in the characteristics of excluded studies table (Characteristics of excluded studies). Overall we excluded 17 studies. Nine non-randomised cohort studies comparing a triple NRTI regimen containing abacavir as a simplification regimen were found (Moyle 2002; Chiesa 2003; Ibarra Barrueta 2004;Cozzi-lepri 2006; Abgrall 2006; Abgrall 2007; Palma 2007; Wolbers 2007; Bommenel 2011). There were 7 prospective observational cohort studies and two retrospective cohort studies (Moyle 2002, Ibarra Barrueta 2004).

We also excluded 8 randomised studies with triple nucleoside regimens as a first line treatment in naive patients (Staszewsky 2001; Matheron 2003; Orkin 2004; Vibhagool 2004; Gulick 2004; Kumar 2006; Kumar 2009; Munderi 2010)

Risk of bias in included studies

We applied Cochrane Collaboration tools to assess each individual study for risk for bias ( see Assessment of risk of bias in included studies). The Cochrane approach assesses risk of bias in individual studies across seven domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, incomplete outcome data, selective outcome reporting and other potential biases. See: Methodological quality graph and summary (Figure 2; Figure 3).

Figure 2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Generation of allocation sequence

In three studies (Clumeck 2001; Maggiolo 2003; Keiser 2005) investigators stated that the participants were randomly assigned to interventions and control, but the method used to achieve randomizations was not explicit; thus we have judged this domain as 'unclear risk of bias'. The others studies reported how they generated the allocation sequence (e.g., computer generated, centrally randomised) and we have judged this domain as "low risk of bias".

Allocation concealment

We made a judgment of 'low risk of bias' for this domain in 4 studies with central randomizations (Bonjoch 2005; Katlama 2003; NEFA Study; Opravil 2002). In 4 studies (Clumeck 2001; Maggiolo 2003; Keiser 2005; Sprenger 2010) it was unclear if adequate measures were taken to ensure that investigators were unaware of the upcoming assignment. Therefore we judged this domain as 'unclear risk' of bias for these studies.

Blinding

All the included trials were reported as open-label studies, and therefore at risk of bias for some of the outcomes analysed (e.g., adverse events, discontinuation for adverse events, overall failure. One study clearly stated that a particular attention was paid to limiting potential judgment biases, and we made a judgment of 'low risk of bias' for this domain (Maggiolo 2003). We made a judgment of "high risk of bias" for the remaining domains in all studies.

Incomplete outcome data

All eight trials clarified the number of and the reasons for drop-outs.An intention-to-treat (ITT) analysis was performed in all studies. We made a judgment of 'low risk of bias' for this domain in all studies.

Selective reporting

There was no evidence of selective outcome reporting in the included trials and it appeared that the outcomes reported were comparable to those specified in the methods section of the reports.

Other potential sources of bias

The authors of seven trials reported having been sponsored by pharmaceutical companies (Clumeck 2001, Katlama 2003, Opravil 2002, NEFA Study; Bonjoch 2005; Keiser 2005; Sprenger 2010), but the investigators provided reasonable reassurances that the manufacturers had no, or a very limited, active role in influencing the design and conduct of most of the studies.

Effects of interventions

See: Summary of findings for the main comparison

See also Table 2

Table 2. Summary estimates and heterogeneity for selected outcomes
  1. ABC= abacavir containing triple nucleoside combination; PI= protease inhibitor; NNRTI= non-nucleoside

Outcome Risk Ratio 95 % CI P value for effect estimates Heterogeneity (I2 ) P value for heterogeneity
Overall Failure0.880.74-1.040.1517 %0.29
-ABC vs PI0.800.62-1.030.0829 %0.22
-ABC vs NNRTI0.990.79-1.240.950 %0.69
Discontinuation for adverse events0.680.44-1.070.0957 % 0.02
-ABC vs PI0.770.39-1.530.4665 % 0.01
-ABC vs NNRTI0.630.34-1.180.1554 %0.11
Virologic failure1.390.95-2.020.0918 %0.28
-ABC vs PI1.490.72-3.080.2839 %0.14
-ABC vs NNRTI1.320.89-1.970.170 %0.47
Death & AIDS related events1.660.62-4.420.310 %0.65
-ABC vs PI1.440.41-5.040.570 %0.48
-ABC vs NNRTI2.120.43-10.380.35Not applicable 

Overall Failure (Analysis 1.1, Figure 4):

Figure 4.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.1 Failure (overall).

Eight trials with 1,606 participants reported on overall failure. Triple nucleoside combination (693 participants) was compared to PI continuation (466 participants) or to NNRTI simplification (477 participants). There was some degree of heterogeneity, and a random effect model was applied. Overall, there was no significant difference between the participants on triple nucleoside combination and controls (RR 0.88, 95% CI 0.74 to 1.04), either PI-based (RR 0.80, 95% CI 0.62 to 1.03) or NNRTI based (RR 0.99, 95% CI 0.79 to 1.24).

Discontinuation for Adverse Events ( Analysis 1.2;Figure 5):

Figure 5.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.2 Discontinuation for adverse events.

Eight trials with 1,597 participants reported on discontinuation for adverse events. Triple nucleoside combination (689 participants) was compared to PI continuation (461 participants) or to NNRTI simplification (447 participants). There was substantial heterogeneity (I2 =57 %), and a random effect model was applied. Overall, although there was no significant difference between the participants on triple nucleoside combination and controls (RR 0.68, 95% CI 0.44 to 1.07), either PI-based (RR 0.77, 95% CI 0.39 to 1.53) or NNRTI based (RR 0.63, 95% CI 0.34 to 1.18), the test for overall effect (p=0.09) was closed to the level of significance, thus suggesting a weak evidence of lower incidence of side effects  in the experimental group

Virologic Failure (Analysis 1.3;Figure 6):

Figure 6.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.3 Virologic failure.

Eight trials with 1,587 participants reported on virologic failure. Triple nucleoside combination (689 participants) was compared to PI continuation (461 participants) or to NNRTI simplification (437 participants). A random effect model was applied (I2 =18 %). Overall, there was no significant difference between the participants on triple nucleoside combination and controls (RR 1.39, 95% CI 0.95 to 2.02), either PI-based (RR 1.49, 95% CI 0.72 to 3.08) or NNRTI based (RR 1.32, 95% CI 0.89 to 1.97), though the test for overall effect (p=0.09) was closed to the level of significance, thus suggesting a weak evidence of higher incidence of virologic failure in the 3NRTI group compared to controls

Death (Analysis 1.4; Figure 7):

Figure 7.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.4 Death.

Four trials with 1,044 participants reported on death.Triple nucleoside combination was compared to PI continuation (579 participants) or to NNRTI simplification (465 participants). A fixed effect model was applied (I2 =0 %). Overall, there was no significant difference between the participants on triple nucleoside combination and controls (RR 1.80, 95% CI 0.43 to 7.61), either PI-based (RR 3.87, 95% CI 0.44 to 34.22) or NNRTI based (RR 0.71, 95% CI 0.07 to 6.74).

Death & AIDS related events (Analysis 1.5):

The outcome AIDS related evens was also available from the studies that reported the outcome death. Again, there was no significant difference between the participants on triple nucleoside combination and controls (RR 1.66, 95% CI 0.62 to 4.42), either PI-based (RR 1.44, 95% CI 0.41 to 5.04) or NNRTI based (RR 2.12, 95% CI 0.43 to 10.38).

Any adverse events (Analysis 1.6;figure 9):

Three trials only reported overall adverse events, with a substantial heterogeneity (I2 =63 %) and no significant difference between the participants on triple nucleoside combination and controls (RR 0.92, 95% CI 0.73 to 1.15),

Rate of patients with viral load < 50 copies/ml ( Analysis 1.7; figure 10)

Five studies (Bonjoch 2005; Clumeck 2001; Katlama 2003; Keiser 2005; Sprenger 2010) reported rate of participants with viral load below the detectability cut-off (<50 copies/ml).. There were not significant differences between ABc recipients and controls (nevirapine in the Bonjoch study, PI in the other studies.)

Changes in lipids.

Change in lipids from baselines was reported in 7 studies (Clumeck 2001; Opravil 2002, Martínez 2003, Katlama 2003; Maggiolo 2003; Keiser 2005; Bonjoch 2005) , but inconsistency in reporting these data did not allow quantitative analysis.

A significant reduction in plasma lipid profiles was observed at week 48 in the abacavir arm for the ITTexposed population in the Clumeck 2001 study. The median change from baseline for the abacavir and PI arms, respectively, was -0.14 mmol/l versus +0.04 mmol/l for triglycerides (P =0.035) and -0.51 mmol/l versus -0.11 mmol/l for cholesterol (P < 0.001).

In the simplified group of the Opravil 2002 study, non fasting serum cholesterol levels decreased from baseline by a median of 35 mg/dL at week 4 and remained reduced thereafter. The changes from baseline differed between the simplified and the PI-continuation group at all time points (P < 0.002). There was also a significant difference in the change from baseline in non fasting triglyceride levels, with a reduction by 43 mg/dL at week 48 in the simplified group. Only 10 and 3 patients in the continuation and simplified groups, respectively, received lipid-lowering drugs during the study. Weight and waist-to-hip ratio did not change throughout the study in either group

Significant reductions in cholesterol and triglyceride plasma levels were observed in the Katlama 2003 study in the Trizivir arm compared to continuation arm (P < 0.001 and P = 0.006, respectively). Moreover, by week 48 the proportions of patients with at least one lipodystrophy manifestation was significantly lower in the simplification compared to the continuation arm (27 vs. 42 %, respectively; p< 0.03)

In the Martinez 2003 study the median fasting plasma triglyceride values at each time point were not significantly different among the groups, though the proportion of patients with plasma triglyceride levels above 400 mg per decilitre was significantly smaller (p=0.05) at 12 months in the abacavir group (4 %) than in the nevirapine group (2 %) or the efavirenz group (13 %). The median fasting plasma cholesterol values were significantly lower in the abacavir group than in the other two groups at all follow-up visits (P<0.001 by the Kruskal–Wallis test) . Moreover, the proportion of patients with plasma cholesterol levels above 240 mg per decilitre was also significantly lower in the abacavir group than in the other two groups at each time point. There were no significant differences among the groups in the proportions of patients with moderate or severe lipoaccumulation or lipoatrophy during the study.

In the Maggiolo 2003 study a persistent increase in the mean triglyceride blood levels in the PI group was observed, with a maximum increase of 88.7 mg/dL from baseline values 104 weeks after initiation of therapy. At all time points, the difference from baseline values was statistically significant (according to a paired Student’s t test). A slight increase (that was not statistically significant) in triglyceride levels was also observed in the efavirenz group (maximum increase from baseline values, 15.9 mg/dL after 24 months). By contrast, ABC-based therapy did not influence this parameter. Cholesterol levels were significantly lowered by abacavir containing treatment throughout the study period; the maximum decrement (27.6 mg/dL) was observed 4 months after initiation of therapy. In contrast, a steady and statistically significant increase in cholesterol levels was observed in the PI group, whereas patients who received efavirenz showed a less marked, but persistent, increase that was more pronounced (and statistically significant) 20 and 24 months after initiation of therapy

In the Keiser 2005 study patients in the abacavir-switch arm had a significantly greater decrease from baseline in total cholesterol (Least Square Mean -LSM, 42 vs -10 mg/dL; P <0.001), LDL-cholesterol ( -14 vs +5 mg/dL; P = 0.016), LDL direct-cholesterol (-15 vs +1 mg/dL; P = 0.028), VLDL-cholesterol (-27 vs -7 mg/dL; P = 0.019), triglycerides (-134 vs -36 mg/dL; P = 0.019), apolipoprotein B (-23 vs -11 mg/dL; P = 0.031), and apolipoprotein E (-2 vs -1 mg/dL; P = 0.021) versus the PI-continuation arm. The LSM value for total cholesterol was below the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) goal value (<200 mg/dL) in the ABC-switch arm (198 mg/dL); however, it remained above this value in the PI-continuation arm (230 mg/dL).Changes in serum triglycerides in the two treatment groups paralleled the changes in total cholesterol.

In the Bonjoch 2005 study median total cholesterol decreased in both groups : from 5.5 mM at baseline to 4.75 mM at week 48 in the abacavir group (p = 0.018) and from 5.6mM to 5.06mM in the nevirapine group (p = 0.03) Also, low density (LDL) lipoprotein cholesterol decreased from 3.59 mM to 2.68 mM (p< 0.001) in the abacavir group, and from 3.6 mM to 3.11 mM (p = 0.046) in the nevirapine group. Likewise, the proportion of patients with hypercholesterolaemia declined significantly (p=0.019) in both groups:from 55.6% at baseline to 30.5% at week 48 in the abacavir group, and from 68.9% to 42.2% in the nevirapine group. The proportion of patients with elevated LDL cholesterol decreased from 60% to 20.6% (P = 0.003) in the abacavir group, and from 59.3% to 41.7% (P > 0.05) in the nevirapine group . All the variations in lipid profile were comparable between groups except for LDL cholesterol at week 48 (P = 0.049), with
a more pronounced decrease in the abacavir group. No significant changes were demonstrated in body fat distribution throughout the study, and no new cases of lipodystrophy were detected.

Change in CD4 cells

Change in CD4 cells count from baselines was reported in 7 studies (Clumeck 2001; Opravil 2002, Martínez 2003, Katlama 2003; Maggiolo 2003; Keiser 2005; Bonjoch 2005) , but inconsistency in reporting these data did not allow quantitative analysis.

In the Clumeck 2001 study CD4 cell counts increase was reported as similar, with a median increase of 26 cells/mm3 in the ABC arm compared with 13 cells/mm3 in the PI (ITT-exposed) arm. Likewise, CD4 count was reported as not significantly different between simplification and continued arms in the Katkama 2003 study: 26 cells/mm3 in both groups. In the Opravil 2002 study CD4 cells continued to increase throughout the study, with little difference between groups. There were no significant differences among the groups in the Martinez 2003 study: the median CD4 cell counts (P>0.40 at each time point by the Kruskal–Wallis test). At 12 months, the median increases from base line were 50, 49, and 39 CD4 cell/mm3 in the nevirapine, efavirenz, and ABC groups, respectively (P=0.48 by the Kruskal–Wallis test). In the Maggiolo 2003 study the immunologic response by week 104 was similar in the ABC, efavirenz, and PI groups, with a mean increase of 146 cells/mL, 127 cells/mL, and 123 cells/mL from baseline values, respectively. In each group, the increase was statistically significant (P <0.001). In the Keiser 2005 study, at week 28, the LSM CD4+ cell counts were 514 and 526 cells/mm3 in the ABC arm and in the PI-continuation arm, respectively, with the LSM difference from baseline being -13 and -1 cells/mm3, respectively (P = 0.597). In the Bonjoch 2005 study, a non statistically significant median CD4+ count increase from baseline of 96 cells/mL and 43 cells/mL was observed in ABC and nevirapine groups, respectively.

Other Outcomes.

Myocardial infarction and cardiovascular events were reported in a study only (Opravil 2002): 1 myocardial infarction episode was observed among 84 participants continuing PI, and no episodes were recorded among 79 ABC recipients.

The study by Opravil et al (Opravil 2002) excluded patients if an archived reverse transcriptase mutation at codon 215 was detected in their proviral DNA at screening or in the plasma RNA–derived sequence from the most recently stored sample with an HIV-1 RNA level of >1000 copies/mL for those without recoverable proviral DNA. However, resistance was not assayed at screening in the remaining studies

The occurrence of Hypersensitivity reactions (HSR) to the administered drugs was unequivocally reported in 6 studies. There were 29 possible HSR, all in patients receiving ABC. Prevalence of HSR varied considerably among studies:2/104 in the Clumeck study, 1/84 in the Opravil study, 5/69 in the Maggiolo study, 11/110 in the Katlama study, 8/149 in the Martinez 2003 study, and 2/52 in the Keiser 2005 studies. The Bonjock study did not mention explicitly HSR, though 5 rash episodes not requiring treatment discontinuation in each treatment groups were reported. Of note, HLA B5701 screening test ( Mallal 2008) was not available or not performed in the included studies.

Quality of life indicators were reported in 2 studies (Keiser 2005; Bonjoch 2005). Health-related quality of life (QOL) was evaluated in the Keiser 2005 study using the Medical Outcomes Study 36-item Short Form Health Survey (SF-36), which patients completed at baseline and at study weeks 12 and 28 (Ware 1993). Healthcare resource utilization (total health care visits, emergency room visits, intensive care hospitalizations, general ward hospitalizations, outpatient clinic visits, home care visits, and long-term care visits) was assessed at weeks 4, 8, 12, 20 and 28.QOL scores did not differ between treatment arms for most domains, except Vitality (in favour of the PI-continuation arm; P = 0.042) and Health (in favour of the ABC switch arm; P = 0.013). There were no differences between treatment arms with respect to any health care utilization parameter. In the Bonjock 2005 study QOL was evaluated by the MOSHIV questionnaire (Medical Outcome Study for HIV-positive patients)(Wu 1997).QOL remained at adequate levels from baseline to week 48 according to the General Health Scale in both groups (group 1—baseline:68.4, week 48: 67.5; and group 2—baseline: 67.1, week 48:67.4). The effort to take medication decreased significantly regarding all the scales evaluated (P , 0.001). A more pronounced decrease in effort to take medication under was observed in group 1 at week 48 with regard to group 2 (P = 0.023).

Subgroups analyses

We performed subgroup analyses according to the type NNRTI drugs in the simplification regimen (Analysis 2.1; Analysis 2.2; Analysis 2.3; additional figg.11-13). Three studies (Maggiolo 2003; NEFA Study; Bonjoch 2005) reported separately results for nevirapine and/or efavirenz. The results of the comparison between triple nucleoside combination and nevirapine or efavirenz containing ART for the outcomes overall failure, discontinuation for adverse events and virologic failure are shown on figures 9-11. There were no significant difference between the participants on triple nucleoside combination and controls, with the exception of a higher number of virologic failure on ABC compared to efavirenz (RR 2.76, 95 % CI, 1.21-6.31). Of the six studies with a PI continuing arm, one (Sprenger 2010 ) had a lopinavir/ritonavir arm, the others a variety of PIs (indinavir, nelfinavir, ritonavir, saquinavir, amprenavir) (Clumeck 2001; Katlama 2003; Keiser 2005; Maggiolo 2003; Opravil 2002). Therefore, it was not possible to perform subgroup analysis according to the prescribed PIs.

A subgroup analysis according to timing of simplification to nucleoside therapy was also performed, and studies in naives patients with simplification after a first-line antiretroviral PI-containing regimen (Bonjoch 2005; Sprenger 2010) analysed separately from those  in antiretroviral-experienced participants (data not shown). This latter group included patients with sub-optimal previous treatment (e.g., NRTI mono- or dual therapy). Of note, subgroup analyses of studies with simplification after first-line therapy did not affect significantly the results of the meta-analysis, though for the outcome virologic outcome the meta-analytical estimates and the test for overall effect(RR 0.84; 95 % CI, 0.35/2.07; p=0.7) were not longer closed to the level of significance as in the overall comparison(1.3; 0.95/2.02; p=0.09)

 Sensitivity analysis

We conducted a sensitivity analysis to determine how the adequacy of allocation concealment affected the results of the review. We made a judgment of 'low risk of bias' for this domain in 4 studies with central randomizations (Bonjoch 2005; Katlama 2003; NEFA Study; Opravil 2002) and of 'unclear risk' of bias in the remaining 4 studies (Clumeck 2001; Maggiolo 2003; Keiser 2005; Sprenger 2010). Subgroup analyses stratifying the trials according to allocation concealment  did not demonstrate significant differences on pooled estimates for the outcomes analysed between trials with adequate and with unclear allocation concealment.

Discussion

This systematic review evaluated RCT in which clinicians use a triple NRTI combination containing abacavir as simplification treatment in a patient with virologic suppression. The major rationales behind regimen simplification are to improve the patient’s quality of life, improve medication adherence, avoid long-term toxicities, and reduce the risk of virologic failure.

Summary of main results

The present systematic review included 8 RCTs that summarized data on pre-defined endpoints from a total sample size of 1675 HIV infected patients. Simplification with ABC triple nucleoside regimen showed an overall failure rate comparable to that of  continuing  PI regimen or  to simplification with NNRTI. Pooled analysis of treatment effects for the endpoint failure due to adverse events  showed lower rate of failure  with triple nucleoside combinations compared to controls, but the difference was not statistically significant (RR 0.68, 95 % CI 0.44-1.07). By contrast, rates of virologic failures   were more frequent with  triple nucleoside combination that with PI or NNRTI, but in both the comparisons the differences were  not statistically significant (overall RR 1.39, 95 % CI 0.95-2.07). However, subgroup analysis of studies (2 RCT for a total of 444 patients) comparing  ABC or efavirenz in combination with 2 NRTI showed an increase of virologic failure in ABC recipients (RR 2.65, 95 % CI 1.26-5.61). No differences in treatment effects were observed with regard to the endpoints death, death and AIDS related events, rate of patients with viral load < 50 copies/ml.

Seven studies reported metabolic parameters and all agreed that simplification with ABC had a favourable and significant impact on lipid metabolism compared to control group. A significant reduction in plasma lipid profiles (triglycerides and cholesterol) was consistently observed in ABC recipients. Moreover, a positive effect of ABC containing regimens on lipodystrophy was reported in 2 of the analysed studies (Katlama 2003; Maggiolo 2003). An increase in CD4 cells count from baseline was evident in all analysed studies, without significant differences between ABC and controls in individual studies. Unfortunately there was inconsistency in reporting metabolic and immunologic data and we could not perform a quantitative analysis.

Overall completeness and applicability of evidence

All the studies included HIV-1 infected patients virologically suppressed after a successful treatment with PI containing ART and then switched to triple NRTI combination. A higher failure rate of ABC/3TC/ZDV in patients with earlier exposure to NRTI mono or dual therapy in comparison with patients without suboptimal therapy has been observed in several studies [Wolbers 2007; Opravil 2002; Bucher 2003]. Failure rates decreased with increasing exposure to ABC/3TC/ZDV and most of the observed events occurred in the first year of therapy. Patients who are likely to harbour archived resistance mutations against NRTI, for example those who had previous zidovudine-containing mono or dual NRTI therapy, have a high risk of virological failure with this strategy [Opravil 2002; Martínez 2003]. On the other hand, a switch to ABC/lamivudine/zidovudine maintains viral suppression in patients never exposed to zidovudine mono or dual therapy [Martínez 2003, Opravil 2004]. The incidence of virological failure per 100 patient-years was 3.4 in patients without versus 13.5 in those with a history of zidovudine mono/dual therapy. After achieving viral suppression, a switch to ABC/lamivudine/zidovudine maintains long-term antiviral efficacy in patients without previous episodes of incomplete viral suppression on nucleoside analogues (Opravil 2004).Thus, the strategy of switching only those patients who are unlikely to harbour NRTI resistance mutations is supported by the available evidence and holds the advantages of preserving other classes of antiretroviral drugs that will thus be available for second-line treatment, to lower blood lipids, and to be cost effective and simple to administer.

There are concerns about ABC use and myocardial infarction that are mainly based on data from non-randomised, observational cohort studies (DAD 2008). By contrast, results of non randomised and randomised studies show that exposure to ABC was not associated with significant increase in MI (Lang 2009; Bedimo; Cruciani 2011). Moreover, any trend towards an association of ABC use with MI was weaker after adjustment for traditional cardiovascular risk factors including renal function (Bedimo). The selection of studies for the current review adds little to this topic, since the RCTs included in the analysis were not designed specifically to investigate differences in cardiovascular disease, and myocardial infarction and cardiovascular events were reported in one study only (Opravil 2002)

Quality of the evidence

Articles included in the analysis were published between 2001 and 2010, and could be classified as low risk of bias trials in most of the domains considered. All the studies described detail on drop-outs, exclusions, missing data, and performed intention to-treat-analysis. However, it was unclear from 4 of the 8 studies if adequate measures were taken to ensure that investigators were unaware of the upcoming assignment, and we judged this domain as 'unclear risk' of bias for these studies. Lack of concealment may have influenced the assessment particularly when dealing with subjective outcomes (e.g., stopping study medication "for any reason"). .However, there were many objective outcomes in this review (e.g., CD4 cell count, proportion of patients with virologic failure, proportion of patients with Viral load below the cut-off, lipids levels). Blinding was not performed in any of the studies included. Likewise, blinding may be of particular importance when outcome measures involve some subjectivity, but probably less important for more objective outcomes, because the risk of ascertainment bias is limited.

Potential biases in the review process

Electronic, online trial and manual searches were conducted to search for relevant articles, but there may still have been papers we didn't find. As a result, selective biases may exist in our review.

Agreements and disagreements with other studies or reviews

Cohort studies showed that patients switching from a virologically effective first-line boosted PI-cART regimen to a cART regimen containing efavirenz or nevirapine have similar virological outcomes to patients who continue on the first-line regimen, whereas patients switching to a regimen containing ABC have an increased risk of virological failure (Abgrall 2006, Bommenel 2011). However, in cohort studies differences in virological outcome associated with different regimens do not necessarily reflect differences in efficacy, as bias due to residual confounding factors cannot be excluded despite adjustments for factors known to be associated with outcome. Moreover, in RCTs simplification of PI-containing HAART in patients with sustained virological response had a higher probability of maintaining the suppression of viral replication after 3 years of follow-up when nevirapine or efavirenz were substituted for PI compared with ABC ( Opravil 2002; Martínez 2003). However, rates of viral suppression among patients who had not had previous suboptimal therapy with NRTI were similar for the three simplification regimens (Martínez 2007).

Results of a previous meta-analysis show a higher rate of virologic failure in patients switching to ABC based triple nucleoside regimen compared to patients continuing protease inhibitor regimen (Bucher 2003). However this meta-analysis had a somewhat different design, and included 3 studies only with a simplification group containing ABC based triple nucleoside regimen.· Since  several studies have been published thereafter, we undertook the current meta-analysis of trials comparing simplified maintenance therapy  with 3- NRTI and  non-nucleoside reverse transcriptase inhibitor , and/or  PI continuation.

Authors' conclusions

Implications for practice

Current simplification strategies include within-class simplifications and out-of-class (commonly a change from a PI-based to an NNRTI-based regimen). ABC-AZT-3TC (Trizivir®) is the only triple-NRTI combination for which randomised, controlled trials are available.This combination is generally not recommended as initial ARV regimen and should be used only when a preferred, an alternative, or an acceptable NNRTI-, PI-, or INSTI- based regimen is less desirable because of concerns about toxicities, drug interactions, or regimen complexity (WHO 2010; DHHS 2011; EACS 2009). Zidovudine/Lamivudine + Tenofovir is another triple NRTI combinations that has been studied (DART 2006); however, because comparative data with standard regimens are not available, this combination cannot be recommended in routine clinical practice (DHHS 2011I).The future use of triple NRTI regimens as simplification strategies need careful long term evaluation, since it is apparent that selection pressure can lead to the development of both resistance to a single agent and cross-resistance to other drugs in the same combination. However, simplification of ART after a successful induction period to a single-class regimen offers potential advantages such as convenient dosing regimens leading to better adherence, favourable lipid profiles, and fewer potentially serious drug interactions than standard cART. Thus, simplification with triple nucleoside regimens AZT + 3TC + ABC should be still considered for individuals who are unable to tolerate or have contraindications to NNRTI or PI based regimens (DHHS 2011). Subsets of patients who may benefit most of triple-nucleoside are those with HIV/TB coinfection, pregnant women, chronic viral hepatitis B, and HIV-2 infection (WHO 2010).

Implications for research

Additional data are needed on longer-term efficacy of triple NRTI regimens, particularly on the development of antiretroviral resistance and quality of life. Though studies in the current review were conducted between 2001 and 2010, the large  majority of patients from studies analysed received old  PI regimens (e.g., indinavir, ritonavir, nelfinavir, saquinavir) not longer recommended by International Guidelines. Since current guidelines recommend new "lipid -friendly" PI, future studies should compare regimens containing these news PIs to triple NRTI regimens. Moreover, there are multiple opportunities to examine these issues in existing cohorts.

The high efficacy of antiretroviral therapy has resulted in more trials that switch or simplify existing therapy in patients with fully controlled HIV. Simplification of a suppressive triple antiretroviral therapy to boosted protease inhibitor monotherapy has demonstrated safety and efficacy in a high proportion of patients. However, simplification with protease inhibitor monotherapy has been associated with an increased risk of virologic failure (Arribas 2009). If simplification to monotherapy is selected to treat some patients, twice-daily lopinavir/ritonavir, or preferably once-daily darunavir/ritonavir, should be chosen as data with other boosted protease inhibitors are inconclusive or lacking. The MONOI study establishes darunavir/ritonavir monotherapy as durable and efficacious for maintaining virological suppression in HIV-1 patients (Katlama 2010; Valantin 2012). Darunavir/ritonavir monotherapy should be considered as a treatment option for standard triple-therapy patients who have had a substantial period of viral suppression.   Despite a higher proportion of intermittent viremia in monotherapy arm, a similar evolution of cellular HIV-1 DNA level was observed between mono and triple therapy arm (Lambert-Niclot 2012). Nevertheless, more studies focusing on the control of HIV replication in viral reservoirs with monotherapy, as with triple therapy are warranted

Acknowledgements

We would like to thank Tara Horvath and the editorial team of the Cochrane HIV group for continuous support. We would also like to thank Joy Oliver, Trials Search Coordinator of the Cochrane HIV group, who designed the search strategy and undertook the electronic search for studies.

Data and analyses

Download statistical data

Comparison 1. Abacavir vs PI continuation or NNRTI
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Failure (overall)81606Risk Ratio (M-H, Random, 95% CI)0.88 [0.74, 1.04]
1.1 ABC vs PI6908Risk Ratio (M-H, Random, 95% CI)0.80 [0.62, 1.03]
1.2 ABC vs NNRTI3698Risk Ratio (M-H, Random, 95% CI)0.99 [0.79, 1.24]
2 Discontinuation for adverse events81597Risk Ratio (M-H, Random, 95% CI)0.68 [0.44, 1.07]
2.1 ABC vs PI6899Risk Ratio (M-H, Random, 95% CI)0.77 [0.39, 1.53]
2.2 ABC vs NNRTI3698Risk Ratio (M-H, Random, 95% CI)0.63 [0.34, 1.18]
3 Virologic failure81587Risk Ratio (M-H, Random, 95% CI)1.39 [0.95, 2.02]
3.1 ABC vs PI6899Risk Ratio (M-H, Random, 95% CI)1.49 [0.72, 3.08]
3.2 ABC vs NNRTI3688Risk Ratio (M-H, Random, 95% CI)1.32 [0.89, 1.97]
4 Death41044Risk Ratio (M-H, Fixed, 95% CI)1.80 [0.43, 7.61]
4.1 ABC vs PI3579Risk Ratio (M-H, Fixed, 95% CI)3.87 [0.44, 34.22]
4.2 ABC vs NNRTI1465Risk Ratio (M-H, Fixed, 95% CI)0.71 [0.07, 6.74]
5 Death & AIDS related events41044Risk Ratio (M-H, Fixed, 95% CI)1.66 [0.62, 4.42]
5.1 ABC vs PI3579Risk Ratio (M-H, Fixed, 95% CI)1.44 [0.41, 5.04]
5.2 ABC vs NNRTI1465Risk Ratio (M-H, Fixed, 95% CI)2.12 [0.43, 10.38]
6 Any Adverse Events3837Risk Ratio (M-H, Random, 95% CI)0.92 [0.73, 1.15]
6.1 ABC vs PI2372Risk Ratio (M-H, Random, 95% CI)1.01 [0.80, 1.27]
6.2 ABC vs NNRTI1465Risk Ratio (M-H, Random, 95% CI)0.75 [0.61, 0.94]
7 Rate of patients with Viral Load < 50 cp/ml5774Risk Ratio (M-H, Random, 95% CI)1.04 [0.92, 1.18]
7.1 ABC vs PI continuation4640Risk Ratio (M-H, Random, 95% CI)1.05 [0.91, 1.21]
7.2 ABC vs NNRTI1134Risk Ratio (M-H, Random, 95% CI)0.98 [0.76, 1.27]
Analysis 1.1.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 1 Failure (overall).

Analysis 1.2.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 2 Discontinuation for adverse events.

Analysis 1.3.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 3 Virologic failure.

Analysis 1.4.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 4 Death.

Analysis 1.5.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 5 Death & AIDS related events.

Analysis 1.6.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 6 Any Adverse Events.

Analysis 1.7.

Comparison 1 Abacavir vs PI continuation or NNRTI, Outcome 7 Rate of patients with Viral Load < 50 cp/ml.

Comparison 2. ABC vs NNRTI
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Failure (overall)3 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
1.1 ABC vs Nevirapine2364Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.74, 1.32]
1.2 ABC vs efavirenz2369Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.72, 1.42]
2 Discontinuation for Adverse Events3 Risk Ratio (M-H, Random, 95% CI)Subtotals only
2.1 ABC vs Nevirapine2364Risk Ratio (M-H, Random, 95% CI)0.63 [0.28, 1.42]
2.2 ABC vs Efavirenz2369Risk Ratio (M-H, Random, 95% CI)0.54 [0.21, 1.40]
3 Virologic failure3 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
3.1 ABC vs Nevirapine2364Risk Ratio (M-H, Fixed, 95% CI)1.11 [0.71, 1.75]
3.2 ABC vs Efavirenz2369Risk Ratio (M-H, Fixed, 95% CI)2.76 [1.21, 6.31]
Analysis 2.1.

Comparison 2 ABC vs NNRTI, Outcome 1 Failure (overall).

Analysis 2.2.

Comparison 2 ABC vs NNRTI, Outcome 2 Discontinuation for Adverse Events.

Analysis 2.3.

Comparison 2 ABC vs NNRTI, Outcome 3 Virologic failure.

Appendices

Appendix 1. Search strategy

Database:     PubMed  2009 - 2012

 

Date:              6 July 2012

  

SearchQueryItems found
#9Search (#3 AND #4 AND #7) AND ("2009/12/01"[Date - Publication] : "2012/07/06"[Date - Publication]) 166
#8Search #3 AND #4 AND #7 885
#7Search #5 OR #6 1304
#6Search trizivir[tiab] 40
#5Search abacavir[tiab] OR ziagen[tiab] 1286
#4Search (randomised controlled trial [pt] OR controlled clinical trial [pt] OR randomised [tiab] OR placebo [tiab] OR drug therapy [sh] OR randomly [tiab] OR trial [tiab] OR groups [tiab]) NOT (animals [mh] NOT humans [mh]) 2570936
#3Search #1 AND #2 76288
#2Search antiretroviral therapy, highly active[MeSH] OR anti-retroviral agents[MeSH] OR antiviral agents[MeSH:NoExp] OR ((anti[tiab]) AND (hiv[tiab])) OR antiretroviral*[tiab] OR ((anti[tiab]) AND (retroviral*[tiab])) OR HAART[tiab] OR ((anti[tiab]) AND (acquired immunodeficiency[tiab])) OR ((anti[tiab]) AND (acquired immuno-deficiency[tiab])) OR ((anti[tiab]) AND (acquired immune-deficiency[tiab])) OR ((anti[tiab]) AND (acquired immun*[tiab]) AND (deficiency[tiab])) 121189
#1Search HIV Infections[MeSH] OR HIV[MeSH] OR hiv[tiab] OR hiv-1*[tiab] OR hiv-2*[tiab] OR hiv1[tiab] OR hiv2[tiab] OR hiv infect*[tiab] OR human immunodeficiency virus[tiab] OR human immunedeficiency virus[tiab] OR human immuno-deficiency virus[tiab] OR human immune-deficiency virus[tiab] OR ((human immun*[tiab]) AND (deficiency virus[tiab])) OR acquired immunodeficiency syndrome[tiab] OR acquired immunedeficiency syndrome[tiab] OR acquired immuno-deficiency syndrome[tiab] OR acquired immune-deficiency syndrome[tiab] OR ((acquired immun*[tiab]) AND (deficiency syndrome[tiab])) OR "sexually transmitted diseases, Viral"[MeSH:NoExp] 283939

Appendix 2. additional figures

figures 7-13 (Figure 7; Figure 8; Figure 9; Figure 10; Figure 11; Figure 12; Figure 13)

Figure 8.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.5 Death & AIDS related events.

Figure 9.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.6 Any Adverse Events.

Figure 10.

Forest plot of comparison: 1 Abacavir vs PI continuation or NNRTI, outcome: 1.7 Rate of patients with Viral Load < 50 cp/ml.

Figure 11.

Forest plot of comparison: 2 ABC vs NNRTI, outcome: 2.1 Failure (overall).

Figure 12.

Forest plot of comparison: 2 ABC vs NNRTI, outcome: 2.2 Discontinuation for Adverse Events.

Figure 13.

Forest plot of comparison: 2 ABC vs NNRTI, outcome: 2.3 Virologic failure.

Contributions of authors

All authors contributed to the design and conduct of this review, as well as with manuscript drafting and submission.

Declarations of interest

MC: has been an advisor/consultant for Bayer, Cephalon and ViiV Health care, and has received honoraria for educational lectures from Abbott, ViiV Healthcare and Novartis. These activities were not related to his work with this review.

Sources of support

Internal sources

  • ULSS 20 Verona, Italy.

  • University of Padua, Italy.

External sources

  • No sources of support supplied

Differences between protocol and review

The risk of bias assessment in the review follows the domain-based evaluation as described in the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 (Higgins 2011), while in the protocol we refer to the criteria given in The Cochrane Handbook for Systematic Reviews of Intervention 5.0.0 (Higgins 2008)

Characteristics of studies

Characteristics of included studies [ordered by year of study]

Clumeck 2001

MethodsProspective, multicenter, open-label, comparative, randomised study
ParticipantsHIV-1 infected patients treated with successful first-line therapy (2NRTI plus PI); mostly on 1st line PI ART, but around 10 % with previous 2 NRTI or monotherapy (AZT)
Interventionstriple NRTI combination (zidovudine/lamivudine/abacavir) vs continuation of 2NRTI plus PI (indinavir, nelfinavir, ritonavir, saquinavir)
Outcomestreatment failure, defined as virologic failure (HIV RNA >400 cp/ml) or premature discontinuation of randomised study treatment; virologic suppression <50 cp/ml; evolution of CD4; change in metabolic parameters, development of adverse events
Characteristics of Included studies 
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskpatients were randomly allocated
Allocation concealment (selection bias)Unclear riskInsufficient information to permit judgment of ‘Low risk’ or ‘High risk’
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Opravil 2002

MethodsProspective, multicenter, open-label, comparative, randomised study
ParticipantsHIV-1 infected patients treated with successful first-line therapy with 2NRTI plus PI (indinavir, nelfinavir, ritonavir, saquinavir, saquinavir/r)
InterventionsPatients were randomised to continuing their current PI-containing antiretroviral regimen or to switch to triple NRTI combination (co formulation of zidovudine/lamivudine plus abacavir or trizivir®)
OutcomesTreatment failure, defined as virologic failure (HIV RNA >200 cp/ml), any change of randomised study treatment, or loss to follow-up; tolerability; change in metabolic parameters.
Characteristics of Included studies 
NotesPatients who have the reverse transcriptase 215 mutation were excluded; study included patients ( ̴ 50 %) with sub-optimal previous treatment (e.g., mono- or dual NRTI treatment)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskCentralized randomisation
Allocation concealment (selection bias)Low riskcentral allocation
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Maggiolo 2003

MethodsProspective, randomised, single centre, Open-Label study.
ParticipantsHIV-1 infected patients treated with successful therapy with 2NRTI plus PI (indinavir, nelfinavir, saquinavir); 20 % only were on first line treatment, but not specified previous treatment
InterventionsPatients were randomised to either continued therapy with their current PI-containing antiretroviral regimen or to the same regimen with abacavir or efavirenz, substituted for the PI.
OutcomesTreatment failure, defined as virologic failure (HIV RNA >500 cp/ml) or premature discontinuation of randomised study treatment; virologic suppression <50 cp/ml; evolution of CD4; change in metabolic parameters.
Characteristics of Included studies 
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskpatients were randomly allocated
Allocation concealment (selection bias)Unclear riskInsufficient information to permit judgment of ‘Low risk’ or ‘High risk’
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
Low riskNo blinding, but particular attention paid to limiting potential judgment biases
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Katlama 2003

MethodsProspective, multicenter, open-label, randomised study
ParticipantsHIV-1 infected patients treated with successful first-line therapy :2NRTI plus PI (61 %; 21 % indinavir + combivir®), 2NRTI plus NNRTI, 3NRTI
Interventionsco formulated zidovudine/lamivudine/abacavir vs continuation of triple-therapy regimen (2NRTI plus PI, 2NRTI plus NNRTI, 3NRTI)
OutcomesTreatment failure, defined as virologic failure (HIV RNA >540 cp/ml) or premature discontinuation of randomised study treatment; virologic suppression <50 cp/ml; evolution of CD4; change in metabolic parameters, development of adverse events
Characteristics of Included studies 
NotesPatients with previous NRTI mono-therapy not included
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskcentrally randomised
Allocation concealment (selection bias)Low riskcentral allocation
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

NEFA Study

MethodsProspective, multicenter, open-label, comparative, randomised study
ParticipantsHIV-1 infected patients treated with successful ART regimen containing a PI and 2 NRTI
InterventionsPatients were randomised to switch from their current PI-containing antiretroviral regimen to the same double NRTI regimen in combination with abacavir, nevirapine or efavirenz as substituted for the PI.
OutcomesDeath, progression to AIDS, virologic failure (RNA >200 cp/ml); CD4 count, incidence of side effects, metabolic changes.
Characteristics of Included studies 
NotesStudy included patients with sub-optimal previous treatment (e.g., mono- or dual NRTI treatment)
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskCentralized randomizations
Allocation concealment (selection bias)Low riskEach patient's identification number and treatment group assigned at the coordinating centre after the centre had received the randomizations form.
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Keiser 2005

MethodsProspective, Randomized, Multicenter, Open-Label study
ParticipantsHIV-1 infected experienced patients treated with successful therapy with 2NRTI plus PI (indinavir, nelfinavir, saquinavir, amprenavir)
InterventionsPatients were randomised to either continued therapy with their current PI-containing antiretroviral regimen or to the same regimen with abacavir, substituted for the PI(s).
OutcomesThe primary endpoint was the change in fasting total cholesterol from baseline at Week 28.The secondary endpoints were the proportion of subjects with plasma HIV-1 RNA <50 copies/mL and the proportion of subjects with plasma HIV-1 RNA <400 copies/mL at each scheduled post-baseline assessment
Characteristics of Included studies 
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskpatients were randomly allocated
Allocation concealment (selection bias)Unclear riskInsufficient information to permit judgment of ‘Low risk’ or ‘High risk’
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding,and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Bonjoch 2005

MethodsProspective, multicenter, open-label, comparative, randomised, non-inferiority study
ParticipantsHIV-1 infected patients treated with successful first-line therapy. Up to 89% of the patients in ABC group and 94% of the subjects in nelfinavir group were receiving PIs at inclusion in the study. Overall, the most frequent PIs were nelfinavir (35%), ritonavir-boosted indinavir (21.6%), and unboosted indinavir (17.2%). Most frequent NRTI strategies were: lamivudine + stavudine (45%) and zidovudine + lamivudine (20%).
Interventionscoformulated zidovudine/lamivudine/abacavir vs coformulated zidovudine/lamivudine + nevirapine
Outcomesvirologic suppression <200 cp/ml at 48 wks; virologic suppression <50 cp/ml; evolution of CD4; change of lipid metabolism
Characteristics of Included studies 
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskMulticenter, open label, blinded, centrally randomised
Allocation concealment (selection bias)Low riskcentrally randomised
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and all of the study’s pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Sprenger 2010

MethodsProspective, multicenter, open-label, comparative, randomised study
ParticipantsHIV-1 infected patients treated with successful first-line therapy (2NRTI plus lopinavir/ritonavir)
InterventionsPatients were randomised to continuing their current PI-containing antiretroviral regimen or to switch to triple NRTI combination (co formulation of zidovudine/lamivudine/ abacavir)
OutcomesPrimary Outcome Measures: Plasma HIV-RNA < 400 cop/ml at week 96 for the Intent- To-Treat (ITT). Secondary Outcome Measures: HIV-RNA <50 cop at week 96
HIV-RNA <400 and <50 cop/ml at week 48 Time to virological failure Immunological efficacy at week 48 and 96 measured by absolute change from baseline in CD4 cell counts. Duration of change in CD4 cell count from baseline to >200, Proportion of subjects experiencing one or more predefined values of fasting glucose and triglycerides, LDL and LDL/HDL ratio. Development of adverse events
Characteristics of Included studies 
NotesStudy data refer to the interim analysis at 48 wks
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskon a 1:1 basis by computer generated allocation
Allocation concealment (selection bias)Unclear riskInsufficient information to permit judgement of ‘Low risk’ or ‘High risk’
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding, and the review authors judge that some of the outcomes are likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcome data available from all randomised participants analysed in the group to which they were allocated. ITT analysis.
Selective reporting (reporting bias)Low riskThe study protocol is available and the interim analysis at 48 wks reports the outcomes that are of interest in the review in the pre-specified way
Other biasLow riskThe study appears to be free of other sources of bias.

Characteristics of excluded studies [ordered by year of study]

StudyReason for exclusion
Staszewsky 2001A randomised trial in naive patients of abacavir/lamivudine/zidovudine vs indinavir/lamivudine/zidovudine
Moyle 2002A retrospective study of pretreated patients who switched to a simplified regiment with triple NRTI regimen containing abacavir
Chiesa 2003Observational cohort study of patients with undetectable HIV RNA on PI containing treatment switched to a combination containing nevirapine, efavirenz or abacavir.
Matheron 2003A randomised trial in naive patients of abacavir/lamivudine/zidovudine vs nelfinavir/lamivudine/zidovudine
Vibhagool 2004Randomized trial of abacavir/lamivudine/zidovudine vs indinavir/lamivudine/zidovudine in ART-naive patients
Ibarra Barrueta 2004A retrospective study of pretreated patients who switched to a simplified regimen with trizivir®
Gulick 2004A randomised trial of triple-nucleoside regimens vs efavirenz containing regimens as a first line treatment in naive patients
Orkin 2004A randomised trial comparing a three- and four-drug HAART regimen in first-line therapy (QUAD study)
Cozzi-lepri 2006A multi cohort prospective observational study comparing abacavir and efavirenz as the third drug used in combination with a background therapy regimen of 2 nucleoside reverse-transcriptase inhibitors in patients with initially suppressed viral loads
Kumar 2006A randomised trial of trizivir, combivir/nelfinavir, and lamivudine/stavudine/nelfinavir in naive patients
Abgrall 2006A cohort study of patients with undetectable HIV RNA who switched from a first PI containing treatment to a combination containing efavirenz, nevirapine or abacavir
Abgrall 2007A cohort study of patients who switched from a first unsuccessful PI containing treatment to a combination containing efavirenz, nevirapine or abacavir
Palma 2007A cohort of paediatric patients switched from protease inhibitor-based HAART to triple nucleoside regimens
Wolbers 2007An observational cohort study of patients switched to triple NRTI combination with abacavir after successful previous treatment with a PI or NNRTI containing combination treatment
Kumar 2009RCT of initial anti-retroviral therapy with triple NRTI combination vs atazanavir/lamivudine/zidovudine
Munderi 2010A randomised trial in naive patients of abacavir/lamivudine/zidovudine vs nevirapine/lamivudine/zidovudine
Bommenel 2011A cohort study of patients who continued on a virologically successful first-line boosted protease inhibitor (PI)-containing combination antiretroviral therapy regimen or who switched to a PI-free cART including efavirenz, nevirapine or abacavir.

Ancillary