ACTG 5202: No significant difference in time to failure between Kivexa and Truvada in patients with lower viral loads when paired with efavirenz or atazanavir

Final results from the ACTG 5202 study were presented at the 17^th Conference on Retroviruses and Opportunistic Infections (CROI) in San Francisco on Wednesday morning.

Unlike prior results in patients with viral loads higher than 100,000 copies/ml, in which virologic failure occurred more quickly with abacavir/3TC than with tenofovir/emtricitabine, there was no significant difference in outcomes between these two nucleoside backbones in patients with viral loads under 100,000 copies/ml.

There was also no significant difference in time to failure between efavirenz and ritonavir-boosted atazanavir, when paired with either of the two dual-nucleoside backbones.

ACTG 5202 was a blinded clinical trial in which 1857 treatment-naive patients were randomised to begin one of four different antiretroviral combinations.

Either of two dual-nucleoside combinations – abacavir/lamivudine (ABC/3TC, Kivexa), or tenofovir/emtricitabine (TDF/FTC, Truvada) – were combined with either of two "third drugs" – efavirenz (Sustiva, also marketed as Stocrin) or ritonavir-boosted atazanavir (ATV/r) (Reyataz).

The study design therefore included four different major comparisons, one for each permutation of study drugs:

• ABC/3TC vs TDF/FTC for patients on efavirenz,
• ABC/3TC vs TDF/FTC for patients on ATV/r,
• efavirenz vs ATV/r for patients on ABC/3TC, and
• efavirenz vs ATV/r for patients on TDF/FTC.

Participants were also stratified into two groups: those with high (≥100,000 copies/ml) and low (

• efficacy (time to virologic failure, defined as a viral load ≥1000 copies/ml at 16 to 24 weeks, or ≥200 copies/ml at or after 24 wks);
• safety (defined as time to grade 3 or 4 symptoms or lab toxicities); and
• tolerability (time to change in treatment regimen for any reason).

Early in 2008, interim study data showed that virologic failure (rebound of viral load after initial suppression) was likely to occur significantly sooner with ABC/3TC than with TDF/FTC for participants with viral loads above 100,000 copies/ml. Participants in the high viral load stratum were unblinded – i.e., informed which nucleoside combination they were taking, so they could decide whether to continue unchanged or switch.

Data from this point in the study (comparisons of the two nucleoside combinations for patients with high viral loads), have been presented and widely circulated.

Data from the final analyses were presented at Wednesday's CROI session, including results for participants with lower viral loads and for the comparisons between efavirenz and ATV/r.

Patients were 83% men, 33% black, 23% Hispanic and 40% white, with median age 38 years, median viral load of 4.7 log₁₀ copies/ml, and median CD4 cell count of 230 cells/mm³. Patients were followed until September 2009, 96 weeks after the last patient was enrolled, for a median of 138 weeks’ follow-up.

Nucleoside comparisons for participants with viral loads below 100,000 copies/ml

For study participants with viral loads below 100,000 copies/ml, results were as follows (all figures for ABC/3TC vs TDF/FTC):

Efficacy:

• for patients on ATV/r: probability of sustained virologic suppression at 96 weeks was 88.3% vs 90.3%; difference of -2.0% (95% confidence interval [CI], -7.5-3.4); hazard ratio [HR] for virologic failure of 1.26 (95% CI, 0.76-2.05).
• for patients on efavirenz: probability of sustained suppression, 87.4% vs 89.2%; difference of -1.8% (95% CI, -7.5-3.9); HR for failure of 1.23 (95% CI, 0.77-1.96).

Safety:

• ATV/r: HR for grade 3 or 4 event of 1.13 (95% CI, 0.83-1.54, p = 0.44).
• efavirenz: HR for grade 3 or 4 event of 1.38 (95% CI, 1.03-1.85, p = 0.03). (Significant difference.)

Tolerability:

• ATV/r: HR of 1.43 (95% CI, 1.06-1.92, p = 0.18).
• efavirenz: HR of 1.48 (95% CI, 1.12-1.95, p = 0.005). (Significant difference.)

Boosted atazanavir versus efavirenz

For all participants of any baseline viral load at the end of the study, third-drug comparisons were as follows (all figures for ATV/r vs efavirenz):

Efficacy:

• for patients on ABC/3TC: probability of sustained suppression at 96 weeks was 83.4% vs 85.3%; difference of -1.9% (95% CI, -6.8-2.6); HR for failure of 1.13 (95% CI, 0.82-1.56).
• for patients on TDF/FTC: probability of sustained suppression, 89.0% vs 89.8%; difference of -0.8% (95% CI, -4.9-3.3); HR for failure of 1.01 (95% CI, 0.70-1.46).

Safety:

• ABC/3TC: HR of 0.81 (95% CI, 0.66-1.00, p = 0.05).
• TDF/FTC: HR of 0.91 (95% CI, 0.72-1.15, p = 0.44).

CD4 cell increases:

• ABC/3TC: increase of 250 vs 251cells/mm³ (p = 0.89)
• TDF/FTC: increase of 252 vs 221 cells/mm³ (p = 0.002)

In an as-treated analysis, fasting low-density lipoprotein (LDL), high-density lipoprotein (HDL), and total cholesterol levels were significantly higher with efavirenz compared to ATV/r, with either nucleoside backbone, and significantly higher with ABC/3TC compared to TDF/FTC, with either third drug. Estimated creatinine clearance was not significantly different between EFV and ATV/r when combined with ABC/3TC, but with TDF/FTC there was a significant increase with efavirenz compared to a modest decline with ATV/r.

Conclusions

For participants with viral load below 100,000 copies/ml, the final data from ACTG 5202 show no significant differences in time to virologic failure between ABC/3TC and TDF/FTC when combined with either efavirenz or ATV/r. There were also no significant differences in time to failure between efavirenz and boosted atazanavir, in combination with either of the pairs of nucleosides. (See, however, the discussion of equivalence, below.)

There were some significant differences in time to adverse events and regimen changes. When taken with efavirenz, ABC/3TC led to grade 3/4 adverse events more quickly than TDF/FTC. (Eric Daar noted that many of these were abacavir hypersensitivity reactions [HSR], as HLA screening was not in standard practice at the time.)

Screening-out persons at risk for HSR at baseline would likely have improved the safety profile of the ABC/3TC arm. (A post-hoc analysis, which could retrospectively exclude patients with HSR, has not yet been done.) Lipid profiles were more greatly affected by ABC/3TC than by TDF/FTC, and more by efavirenz than by ATV/r.

Finally, TDF/FTC resulted in greater CD4 cell count increases than ABC/3TC, when used in combination with efavirenz.

Equivalence: a statistical footnote

Although no significant difference appeared in any of the paired comparisons between treatment combinations, ACTG 5202 was not able to formally demonstrate "equivalence" for any comparison. This is due to somewhat involved statistical considerations.

In the first place, an equivalence comparison is more stringent than the "non-inferiority" comparison commonly used in antiretroviral clinical trials.

Secondly, the difference between anticipated and actual study outcomes made the definition of equivalence more demanding than originally intended. (Each comparison looked at the hazard ratio between the two treatment arms, defining the two arms as equivalent if the 95% confidence interval of the hazard ratio fell entirely within the range of 0.71 to 1.40. These hazard ratio boundaries were intended to correspond to a ±10% difference in probability of failure between the two arms. However, this calculation required an assumption of how many patients would have failed by week 96, which was estimated at roughly 32%. In fact, the actual number of failures was considerably smaller, with the statistical effect of making equivalence a more stringent condition.)

In short, although the paired efficacy comparisons fell short of formal equivalence as defined in the study protocol, they still demonstrated a lack of statistically significant difference in outcomes.

Further information

You can view the abstract on the official conference website.

You can also view a webcast and slides of this session on the official conference website.