New fixed-dose combination pills measure up to Atripla - corrected version*

Several studies, or updates of studies, comparing newer against older drug regimens were presented at the Eleventh International Congress on Drug Therapy in HIV Infection last week.

Amongst them were the 48-week results from the STAR study, an open-label study comparing the new-generation NNRTI (non-nucleoside reverse transcriptase inhibitor) drug, rilpivirine, with the first-generation NNRTI efavirenz, not as themselves, but in their one-pill-once-a-day guises as Atripla (tenofovir/FTC plus efavirenz) and Eviplera (tenofovir/FTC plus rilpivirine; called Complera in the US).

This is the first time these two one-pill, fixed-dose combinations have been directly compared with each other.

Also presented were the 96-week results of two placebo-controlled studies comparing Gilead’s new ‘Quad’ pill Stribild against either Atripla or tenofovir/FTC (Truvada) and ritonavir-boosted atazanavir (ATV/r: Reyataz/Norvir). Stribild contains tenofovir and FTC and the integrase inhibitor elvitegravir plus cobicistat, a drug developed by manufacturers Gilead as an alternative to ritonavir and which, like ritonavir, boosts the blood levels of elvitegravir to effective levels rather than acting against HIV itself.

Eviplera superior to Atripla for lower viral loads...

The STAR study was conducted as an open-label study because the licensing studies of rilpivirine, ECHO and THRIVE, did not compare the fixed-dose combinations but efavirenz and rilpivirine separately, in combination with Truvada (tenofovir/FTC) or, in the case of THRIVE, sometimes two other nucleoside (NRTI) drugs. This meant that, because the studies also included placebos of whatever drug the participant was not actually taking, people were taking at least three pills a day and sometimes more.

The STAR study compared viral suppression, CD4 count rises, resistance, adverse events and blood lipids in 392 treatment-naive people taking Eviplera against 392 taking Atripla, without placebos. The mean CD4 count of those starting was 390 cells/mm³ and the mean viral load 63,000 copies/ml (4.8 logs). After 48 weeks, 86% of those on Eviplera and 81% on Atripla had a viral load below 50 copies/ml. This percentage difference was not statistically significant (95% confidence interval -1.2 to +9.2).

But it was significantly different in the 65% of people who started treatment with a viral load below100,000 copies/ml: here 88% of people on Eviplera and 81% on Atripla had an undetectable viral load; this 7% difference had a 95% confidence interval of +0.9 to +13.2, meaning that Eviplera was not just ‘non-inferior’ to Atripla, but superior in terms of the primary outcome of viral suppression.

...but more resistance cases, especially in higher viral loads

Nonetheless, although viral suppression failure rates on rilpivirine were lower, there was concern that there would be more drug resistance seen in those whose regimens did fail. This had been seen in ECHO and THRIVE and was a reason rilpivirine was not immediately adopted as an NNRTI of choice for first-line therapy.

Resistance rates were not as high in STAR. The percentage of people developing drug resistance on Eviplera was 4%, as opposed to 7% in the other two studies (pooled average), and in people starting therapy with a viral load over 100,000 copies/ml it was 5%, as opposed to 8%. In contrast, resistance rates to Atripla were just 1% in STAR and 2% in ECHO and THRIVE.

However, in the small number of people with baseline viral loads over 500,000 copies/ml, 18% on Eviplera in STAR developed resistance compared to 4% on Atripla and this may indicate the need for continued caution in using rilpivirine in people with very high viral loads.

In contrast, there were more adverse events in people taking Atripla.These were mainly due to the well-known psychological and neurological side-effects of efavirenz. Fifty-seven per cent of study participants on Atripla had a psycho-neurological adverse event compared with 30% on Eviplera; especially marked was the difference in dizziness (26% on Atripla versus 8% on Eviplera) and there were also differences in dreams, sleep patterns and feeling sleepy during the day.

Other differences between the two drugs were not so marked; for instance, although the efavirenz in Atripla raised total cholesterol by 22% whereas it stayed largely the same in people on Eviplera, efavirenz also raised ‘good’ HDL-cholesterol, meaning that the total/HDL cholesterol ratio, which is probably the best guide to cardiovascular risk, remained unchanged.

Stribild non-inferior to efavirenz or atazanavir but unexplained failure rate in women

The 96-week results from the 102 and 103 double-blind placebo-controlled trials of Stribild differed little from the 48-week results already presented and analysed, most recently last September.

In study 102, Stribild is being compared with Atripla and in 103, with Truvada/atazanavir/ritonavir. The studies are being continued till week 192.

At 96 weeks, 84% of participants in trial 102 and 83% in trial 103 on Stribild had viral loads under 50 copies/ml compared with 82% taking the other drugs in both trials. Failure due to non-suppression of viral load, as opposed to other reasons like drop-outs and side-effects, was 6% for Stribild in either study, 8% for Atripla and 8% for ATV/r.

In the 103 trial, an analysis was presented of differential success rates for different populations. For instance, younger people (under 40) did better on Stribild with 83% having a viral load under 50 copies/ml at week 96, against 77% on ATV/r. In comparison 88% of older people had a viral load under 50 copies/ml on ATV/r compared with 84% on Stribild.

One safety concern with Stribild is that a number of cases of decline in kidney function have been seen: these seem to be related to cobicistat rather than the other drugs. At 96 weeks, there was a difference in trial 103: serum creatinine (a waste product that kidneys should eliminate and which is a measure of kidney efficiency) had risen by 0.13 mg/dl in Stribild patients and only 0.01 mg/dl in patients on ATV/r.

One result stood out, however, that is worrying and so far has not been explained. There were very few women in these studies; just 12% in 102 and 10% in 103 were women, and the failure to recruit women was criticised in this conference, most notably by community co-chair Brian West, chair of EATG, in his closing remarks.

This means that there were just 154 results from women out of 1400 participants in total. But lower virological suppression rates were seen in women, notably on Stribild and notably in the 70 women in 103, where 74% taking ATV/r achieved a viral load under 50 copes/ml, but only 59% (17 out of 29) taking Stribild. Whether these 12 failures, which are much higher proportionately but lower numerically than in men (47 men experienced treatment failure on Stribild) is just a statistical fluke or linked to a gender characteristic will hopefully be answered when further studies currently being conducted in women and children publish their results.

*Correction

• The initial version of this story said that  Eviplera had superior efficacy to Atripla for patients with baseline viral loads over 100,000 copies/ml,  rather than under 100,000 copies/ml, as is the case.