ACTG 5205: atazanavir/ritonavir vs efavirenz in treatment naive patients

Simon Collins, HIV i-Base

The few randomised clinical trials using currently licensed antiretrovirals worth highlighting from CROI partly stood out because there were fewer comparative studies that at previous CROI meetings.

Of these, ACTG 5205, and its metabolic sub-study ACTG 5224s, generated most attention.

ACTG 5205

ACTG 5205 enrolled over 1850 treatment naïve patients from 2005-7 and followed them through to September 2009. The study was designed to compare efavirenz to atazanavir and tenofovir/FTC to abacavir/3TC by randomising equally to one of four groups. Patients were stratified by baseline viral load (above vs below 100,000 copies/mL). [1, 2]

Baseline demographics included 83% men/17% women; 40% white (non-Hispanic), 33% black, 23% Hispanic; median age 38 years median viral load ~50,000 copies/mL, and median CD4 230 cells/mL, Primary efficacy endpoints were time to confirmed virological failure (>1,000 copies/mL) at or after 16 weeks and before week 24 or viral load >200 copies/mL at week 24. Safety endpoints included time to first grade 3/4 event or laboratory abnormality at least one grade higher than baseline (excluding unconjugated hyperbilirubinaemia). Tolerability was assessed as time to change of treatment.

In February 2008, following a DSMB review, patients with baseline viral load >100,000 copies/mL who were using abacavir/3TC were unblinded and recommended to change to tenofovir/FTC due to significantly poorer virological responses. [3, 4]

Interpreting the final results now is further complicated because HLA-B*5701 testing was not available when the study started. Also, and perhaps more surprisingly, baseline resistance testing was performed in less than half the study group. In practice, this means that the most relevant results from the whole study relate to the comparisons between efavirenz vs atazanavir/r.

Using the primary virologic endpoint, there were no significant differences between atazanavir/r and efavirenz with either abacavir/3TC (HR 1.13, 95%CI 0.82–1.56) or tenofovir/FTC (HR 1.01, 95%CI 0.70, 1.46). The differences for the safety endpoint report benefits for atazanavir/r over efavirenz only with abacavir/3TC (HR 0.81, 95%CI 0.66–1.0; p=0.05), probably driven by a caution over management of rash and a potential hypersensitivity reaction when NNRTIs are prescribed with abacavir. This was seen even more in the tolerability analysis (HR 0.69, 95%CI 0.55, 0.86; p=0.0008).

Cardiovascular and renal events, non-AIDS malignancies and bone fractures were broadly similar in each group.

Again, as seen in other PI vs NNRTI studies, patients with virological failure were significantly more likely to develop resistance on the NNRTI compared to the PI regimen (approximately 60% vs 20% of virological failures included >1 major mutation in the efavirenz vs atazanavir/r group respectively).

The only signficant differences in CD4 responses were seen when atazanvir/r group was compared to efavirenz, but only when tenofovir/FTC was used as the nucleoside backbone (+252 vs +221 cells/mm3 at 96 weeks, p=0.002).

Lipid differences were more complicated, and while statistically significant for some values, may or may not be of clinical relevance.

In an on-treatment analysis at week 48, efavirenz was consistently associated with significantly greater increases in total cholesterol, LDL and HDL regardless of nucleosides (all comparisons p<0.001, except LDL with TDF/FTC, p=0.002). Increases were also consistently greater with abacavir/3TC compared to tenofovir/FTC. There were no significant differences for triglycerides although there was a trend for greater increases with atazanavir/r compared to efavirenz, when used with tenofovir/FTC (p=0.07). Despite this, there were no significant differences in total cholesterol:HDL ratio in any comparisons.

Finally, creatinine clearance dropped by approximately -3.0 mL/min at week 96 (as-treated analysis) when atazanavir/r was used with tenfovir/FTC compared to slightly higher increases in the other three groups (p<0.001). These differences were described as modest and <5% of patients in any arm experienced changes of greater than 25% decline.

ACTG 5224s

The metabolic substudy ACTG 5224s provided data on bone mineral density (BMD) and limb fat changes in 269 patients in ACTG 5202 (approximately 65 from each of the four comparative regimens). [5, 6]

Exclusion criteria for the substudy included diabetes or other complications including use of medication related to bone or body composition. DEXA evaluations (whole body and bone) were taken at baseline and at 24, 48 and 96 weeks, then annually. CT abdominal scans were taken at baseline and at week 96.

When no interaction was seen between either the RTI component or third drug components, factorial analyses were performed comparing pooled results for each dual RTI, and for atazanavir/r to efavirenz.

Primary endpoints included percentage changes in hip and lumber spine between the two RTI components, and changes of >10% loss of limb fat. Secondary analysis included fracture rates and the same bone and fat changes in the PI vs NNRTI groups.

Baseline demographics broadly reflected the main study and were balanced between groups. Of note, baseline rates of osteopenia (T-score <1.0) were 35% at lumbar spine and 23% at the hip.

Mean values for BMD at lumbar spine dropped in all groups over the first year of treatment and then recovered by about 50% over the subsequent year. These declines were more significant in the tenofovir/FTC compared to abacavir/3TC (approximately -3.5%% vs -1.5% at week 96, p=0.004) group and in patients using atazanavir/r compared to efavirenz (-3.2% vs -1.7, p=0.035).

Tenofovir/FTC was associated with a greater drop in hip BMD at 96 weeks compared to abacavir/3TC (-4.0% vs -2.6%, p=0.025) with no difference between atazanavir/r and efavirenz (p=0.59, each approx -3.0%). Early declines in hip BMD did not appear to reverse over time.

Fracture rates were similar in all groups, with an incidence of 1.7 per 100 patient years, all of which were reported as traumatic (ie expected in general life). No difference by regimen was seen in the main study (where 12% of fractures were without trauma).

Changes in fat distribution was complicated by the study decision to select a relatively marginal 10% cut-off for fat loss as the endpoint. No differences were seen between arms using this criteria (approximately -16%), with a post hoc analysis using >20% reported in <5% patients with no clear RTI or third drug association.

Absolute mean values increased in both RTI arms (approximately +1 kg gain, no statistical difference) but this was higher in the atazanavir/r vs efavirenz groups (approximately +2 kg vs +1 kg;  +30% vs +15%, both p=0.008).

Trunk fat increases were similar in each RTI group, but atazanavir/r had greater increases at 96 weeks compared to efavirenz (approximately +2.4 kg vs +1.2 kg, p=0.023).


The clearest outcome from these complicated results is likely to be a stronger recommendation for atazanavir/ritonavir as a clinical option in for first-line therapy.

Absolute differences in side effects, tolerability and metabolic differences are more complicated to interpret in clinical terms, even when they are statistically significant, but would be a factor to consider in individual patients at higher risk.

The reductions in bone mineral density in all groups are concerning, especially given the high percentage of patients with low levels at baseline. While fracture rates in this study were low, other studies at CROI suggested that the concern that ageing will uncover reduced BMD as a greater complication in HIV-positive people compared to the general population, may, unfortunately, be well founded.

Fat changes are difficult to interpret given the choice of endpoint for fat loss, although ~5% patients lost >20% fat across all arms. The significant increases in trunk fat, largely interpreted as a return to health effect, were based on DEXA results. The analysis of CT results, not included in the CROI presentation, is needed to determine whether this is an accumulation of visceral or subcutaneous fat. 


Unless stated otherwise, all references are to the Programme and Abstracts of the 17th Conference on Retroviruses and Opportunistic Infections, 16-19 February 2010, San Francisco. All oral abstracts are available as webcasts.

  1. Daar E et al. ACTG 5202: Final results of ABC/3TC or TDF/FTC with either EFV or ATV/r in treatment-naive HIV-infected patients. 17th CROI 2010. Oral abstract 59LB.
  2. Webcast: Advances in ART. Wednesday 9.30am.
  3. US study changes use of abacavir/3TC in naive patients with viral load >100,000 copies/mL based on DSMB recommendation. HIV Treatment Bulletin, April 2008.
  4. Sax P et al. Abacavir-lamivudine versus tenofovir-emtricitabine for initial HIV-1 therapy. N Engl J Med. 2009;361. DOI: 10.1056/NEJMoa0906768.
  5. McComsey G et al. Bone and limb fat outcomes of ACTG A5224s, a substudy of ACTG A5202: a prospective, randomized, partially blinded Phase III trial of ABC/3TC or TDF/FTC with EFV or ATV/r for initial treatment of HIV-1 infection. 17th CROI 2010. Oral abstract 106LB.
  6. Webcast: Complications of HIV and ART. Thursday 9.30am.

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