Study protocol
Investigators from several US treatment centres, along with those from GlaxoSmithKline (who market Trizivir) undertook this Phase IV, randomised, open-label study that effectively compared abacavir and unboosted atazanavir head-to-head, with a common background of AZT and 3TC in both arms. The study's official title was ESS100327, but it is also known as the ACTION study.
Participants were randomised to receive one of two triple-drug antiretroviral regimens: AZT plus 3TC plus abacavir (taken as one Trizivir tablet twice daily) or AZT plus 3TC plus atazanavir (taken as one Combivir pill twice daily, plus two 200mg capsules of atazanavir once daily).
A switch to tenofovir (Viread) was allowed if participants taking abacavir experienced a suspected hypersensitivity reaction (HSR). Similarly, a switch to unboosted fosamprenavir was allowed if participants taking atazanavir experienced significant jaundice or yellowing of the whites of the eyes – both of which are clinical manifestations of hyperbilirubinaemia.
The study protocol had five different definitions of virological failure:
- viral load reduced less than 1 log from baseline by Week 12
- viral load not below 400 copies/ml by Week 24
- viral load previously below 50 copies/ml rebounding above 400 copies/ml prior to Week 24
- confirmed viral load above 400 copies/ml after Week 24
- unconfirmed viral load above 400 copies/ml at Week 48
None of these definitions meet the current gold standard of a sustained viral load below 50 copies/ml at Week 48, however, which has been used in many recent studies, including those comparing currently recommended first-line therapies: boosted fosamprenavir with boosted lopinavir (Kaletra) and efavirenz with Kaletra.
The triple nucleoside once-daily fixed-dose combination pill Trizivir (containing AZT/3TC/abacavir) is non-inferior to the unboosted protease inhibitor (PI) atazanavir (Reyataz plus Combivir (containing AZT/3TC) as first-line therapy, according to data presented on Thursday to the Forty-Sixth Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in San Francisco.
After 48-weeks, 62% of participants on the triple NRTI achieved viral loads below 400 copies/ml compared with 59% on the unboosted PI, although important differences were seen in patients who initiated therapy with viral loads above 100,000 copies/ml, leading the investigators to conclude that Trizivir should only be considered as first-line therapy in patients with viral loads below this threshold.
Currently, treatment guidelines from both the United States and the United Kingdom do not recommended a triple NRTI combination as first-line therapy due to Trizivir's reduced potency compared with efavirenz. Although unboosted atazanavir is recommended in US guidelines as an option for first-line treatment, the latest IAS-USA international treatment guidelines only recommend the use of ritonavir-boosted atazanavir due to concerns over low trough levels of the unboosted drug that might lead to resistance. Unboosted atazanavir is not approved in the EU for similar reasons, and therefore, UK guidelines do not recommend its use in first-line therapy.
However, the guidelines do state that there are situations when the triple NRTI combination of AZT/3TC/abacavir could be considered as first-line therapy - notably when NNRTI- or boosted PI-based therapy cannot be used, for example when the concomitant treatment of a coinfection - in particular tuberculosis (TB) - is warranted, since TB treatment can interact with NNRTI- or boosted PI combinations.
Baseline characteristics
Demographic, immunologic and virologic characteristics were essentially similar in both arms. This was an ethnically varied group, with an average age of around 37 years; one-in-five were women. Baseline viral loads ranged between 5,000 and 200,000 copies/ml, and the median baseline viral load was around 66,000 copies/ml; median CD4 counts were at least 100 cells/mm3, and the median CD4 count was around 268 cells/mm3.
Intent-to-treat exposed analysis
In total, 279 patients were enrolled in the study, 139 were randomised to receive abacavir and 140 received lopinavir. However, one patient in the abacavir arm did not receive treatment and this led the investigators to provide a slightly modified intent-to-treat analysis which they termed, 'intent-to-treat exposed (ITT-E)'.
Efficacy data
The investigators first presented ITT-E data where switch does not equal failure. This found that, overall, both arms were similar in efficacy. At 48 weeks, 64% of participants initially randomised to the abacavir arm and 63% of participants initially randomised to the atazanavir arm had sustained viral loads below 400 copies/ml. Both arms had a median CD4 count increase of 147 cells/mm3 from baseline to week 48.
However, there were noticeable differences in efficacy at higher viral loads when the participants were stratified according to baseline viral load above and below 100,000 copies/ml.
Only 39% of the 23 participants in the abacavir arm with baseline viral loads above 100,000 copies/ml had sustained viral loads below 400 copies/ml at Week 48. In contrast, 64% of the 25 participants in the atazanavir arm with baseline viral loads above 100,000 copies/ml had sustained viral loads below 400 copies/ml after 48 weeks. However, due to the small number of participants in both arms, this only achieved borderline statistical significance.
During their summing up, the investigators also presented more rigourous ITT-E data, where switch equals failure, and which found slightly reduced efficacy at 48 weeks in the study population as a whole: 62% in the abacavir arm and 59% in the atazanavir arm. No CD4 cell count data were provided for this analysis.
Discontinuation, virological failure, resistance, and adverse events
A total of 103 out of the 138 (75%) who began treatment in the abacavir arm completed the full 48 weeks, compared with 98 out of the 140 (70%)who began treatment with atazanavir. The differences were not statistically significant.
The most common reason for not completing the study was virological failure. A total of 18 (13%) in the abacavir arm and 17 (12%) in the atazanavir arm met one of the five definitions of virological failure. There were no significant differences between the arms in terms of which of these five definitions they met.
Resistance profiles were somewhat similar between the two arms, with ten participants in the abacavir arm harbouring resistant viruses, nine of whom had the M184V mutation (which confers resistance to 3TC and FTC), seven of whom had minor PI mutations. One individual had several TAMS and one individual – strangely, and possibly due to baseline transmitted resistance – K103N, which confers resistance to NNRTIs.
Eleven participants in the atazanavir arm had resistant virus, nine of whom also had the M184V mutation, six of whom harboured minor PI mutations and again one had K103N.
Similar numbers in both arms experienced grade 2-4 adverse events. Notably 21% and 4% in the atazanavir arm experienced hyperbilirubinaemia and yellowing of the whites of the eyes, respectively. The most commonly-reported side-effects in the abacavir arm were nausea (11% versus 4% in the atazanavir arm) and fatigue (5% versus 2% in the atazanavir arm), and 5% had a suspected abacavir HSR.
Again, lipid profiles were comparable at week 48, with very slight increases of total cholesterol and very slight reductions in LDL in both arms. However, slightly higher increases were noted in triglycerides in participants in the abacavir arm (+11% from baseline versus +5% in the atazanavir arm); and slightly higher increases were seen in HDL in the atazanavir arm (+21% versus +16% in the abacavir arm).
Conclusion
The investigators concluded that although abacavir-containing triple nucleoside therapy was non-inferior to atazanavir-containing PI-based combination therapy according to the study protocol, Trizivir “remains a viable option” for “select patients naïve to therapy” with viral loads below 100,000 copies/ml.
However, these data show that neither Trizivir nor unboosted atazanavir appear to be as potent as the currently recommended - or alternative - first-line regimens in the most recent treatment guidelines from IAS-USA or the British HIV Association (BHIVA).
Kumar P et al. ACTION study: efficacy and safety of abacavir/lamivudine/zidovudine [ABC/3TC/ZDV] BID versus lamivudine/zidovudine [3TC/ZDV] BID + atazanavir [ATV] QD in ART-naïve HIV-1 infected subjects. 46th ICAAC, San Francisco, abstract H-1058, 2006.