Maraviroc fails to show non-inferiority to efavirenz in treatment-naive patients: 48 week results

Simon Collins, HIV i-Base

Several studies presented new information on the CCR5 inhibitor maraviroc.

Michael Saag from University of Alabama at Birmingham and colleagues presented 48-week results from a randomised, double-blind study comparing maraviroc to efavirenz, each with background AZT+3TC, in 721 treatment-naive drug-sensitive patients (28% women). The study was designed to assess non-inferiority between efavirenz (QD) and maraviroc (BID). [1]

Baseline median CD4 (range) was 241 (5-1422) and 254 (8 – 1053) cells/mm3 and mean viral load was about 4.9 logs (SD+/- 0.7) respectively in the maraviroc and efavirenz arms.

By 48 weeks, efficacy results favoured efavirenz. 11.9% versus 4.2% discontinued treatment due to viral failure (with 70% vs 73% patients suppressed to <400 copies/mL and 65% vs 69% suppressed to <50 copies/mL) in the maraviroc and efavirenz arms respectively.

The lower confidence interval was –9.5% for suppression to <400 and –10.9% for <50 copies/mL. As the predefined lower bound was set at 10%, maraviroc could only be shown to be non-inferior compared to efavirenz for suppression to <400 copies/mL, but not for the more important <50 copies/mL viral endpoint. When stratified by baseline CD4 count >100,000 copies/mL, suppression to <50 copies/mL occurred in 60% of maraviroc patients vs 67% with efavirenz.

However, mean CD4 count increase from baseline was greater in the maraviroc arm (+170 vs +143 cells/mm3) with a significant difference (+26.0 95%CI, 7.0, 46.0). Tolerability also favoured maraviroc (4.1% versus 13.6% discontinued due to side effects, with the efavirenz having a higher report of grade3/4 events. Four malignancies were reported in the efavirenz arm (2 Hodgkins disease, 1 NHL, 1 KS) compared to one case of NHL in the maraviroc arm.

The incidence of grade 3/4 transaminase abnormalities was similar between the two groups.

Change in median lipids (TC, HDL, LDL and TG) were stable or non-significant for maraviroc arms, and increased in the efavirenz groups, similar to that reported in previous naive studies.

One as yet unexplained finding was that these results were significantly different by North/South region. There was similar suppression to <50 copies/mL in patients from North America and Europe (68.0% vs 67.8%), but not in patients from Argentina, South Africa and Australia (62.1 vs 71.0), and this has driven the difference in the overall study.

Table 1: Week 48 results marviroc BID vs efavirenz QD

MVC BID + CBV (n=360) EFV + CBV (n=361) Difference* (lower bound of 1-sided 97.5% CI)
VL <400 copies/mL, % 70.6 73.1 -3.0 (-9.5)
VL <50 copies/mL, % 65.3 69.3 -4.2 (-10.9)
Mean change from BL in CD4+ cell count, cells/mm3 +170 +143 +26 (+7, +46)
Patients with catgeory C events 6 (1.7) 12 (3.3) N/A

* ITT analysis; Lower bound of 1-sided 97.5% confidence interval; non-inferiority margin = ­10%


Failure to achieve non-inferiority at suppression to <50 copies/mL in the study was dependent on the choice of –10% as a lower cut-off, although other FDA regulatory studies have set this at 12%. However, this still leave the scenario of having a drug that is 9% less potent than standard of care.

This, together with poorer response rates when starting with baseline viral load >100,000 copies/mL and a BID formulation, will limit interest in use for first-line treatment. However, as risk of developing X4 tropism increases with duration of infection, defining an optimum time to use maraviroc remains unclear.

Wider use may also be more likely if the geographical differences can be explained, perhaps relating to the sub-type or clade of HIV infection, and in patient dependent on a higher CD4 response.

It is difficult to understand why treatment naive trials still enroll patients with very advanced HIV in to an investigational trial (both arms included at least one patient with CD4 <10 cells/mm3). It is also difficult to understand why anyone with a CD4 count over 1000 cells/mm3 would be encouraged to enter the study.

Treatment advocates have long argued for upper and lower CD4 cut-offs (around 100-500 cells/mm3), both to guarantee that people most in need of treatment receive at least the minimum standard of care, and that the results are generally applicable for people who will use the results in a real world setting.


Saag M, Ive P, Heera J et al. A multicentre, randomised, double-blind, comparative trial of a novel CCR5 antagonist, maraviroc versus efavirenz, both in combination with Combivir (zidovudine [ZDV]/lamivudine [3TC]), for the treatment of antiretroviral naive subjects infected with R5 HIV-1: week 48 results of the MERIT study. 4th IAS conference, Sydney, 2007. Abstract WESS104.

Links to other websites are current at date of posting but not maintained.