Efavirenz 006 Analysis of Virologic Failures Only:
2-3 YEAR FOLLOW-UP

When To Begin Therapy Analysis Suggests Baseline CD4 Count Does Matter

BRIEF SUMMARY:
In 2-3 year follow-up, the study looks at viral load failure which is defined as viral load >50 copies/ml. Patients on the efavirenz regimen (+AZT/ 3TC) were more likely to still be <50 copies/ml than patients on the indinavir (+AZT/ 3TC) regimen (80% vs 65%). The study authors estimated that if a person remained fully adherent on the efavirenz regimen after being <50 copies/ml for 2 years they were likely to remain <50 copies/ml for at least 6 years, and if they were <50 copies/ml for 3 years the preliminary data suggests they could remain undetectable for 10 years. In addition, this study reported data suggesting that baseline CD4 count does predict response to therapy. The data suggests that if a person starts therapy when CD4s are >350 with a potent regimen that is capable of fully suppressing viral load to <50 copies/ml the person is more likely to have undetectable viral load 2 years out than if they started with cd4s <350. This data is preliminary but suggests that the higher the CD4 count when starting therapy the more likely a person will sustain viral load <50 copies/ml. Of course, this assumes the patient is fully adherent. And this study took place in treatment-naïve patients. Also important, the follow-up on this study is only 2-3 years. So, after 5 or more years it may become more clear that baseline CD4 count matters in response to therapy, and of course we could see that baseline CD4 doesn't matter. Another important finding reported by the study authors. They found that the increase in CD4s from therapy was proportional to the CD4 count when therapy was started. The lower a patient's cd4 count was when starting therapy the less of an increase in CD4 count the patient achieved. Read the details below for a better understanding of the findings.

The Study

Levy from DuPont reported a 2 year follow-up analysis on efavirenz from the 006 study which compared EFV+AZT/ 3TC to IDV+AZT/ 3TC and EFV+IDV (without nukes). They compared the time to virologic failure between the two regimens, and left out any patient who discontinued, and only included confirmed viral loads >50 copies/ml. I assume this analysis is in part aimed at addressing the initial criticisms of the study that it was an open-label study and people randomized to IDV arms could withdraw from study.

This is a retrospective analysis of the 006 study which enrolled 1266 treatment naïve patients. Two-year follow-up is available on all 1266 patients & 3-year follow-up data is available on 764 patients. Study authors said this study objective was to assess whether evolving recommendations for the timing of HAART are supported by study 006.

For this study they used an analysis called: say Time-To-Virologic Failure (TVF). AIDS defining events & discontinuations from the study for any reason are censored and not considered treatment failures, & thus not included in this analysis. Patients who discontinued from the study when viral load was >50 copies/ml are virologic failures at all subsequent times.

RESULTS
After 2 years on study:

• In EFV+AZT/ 3TC arm, 80% still had <50 copies/ml (n=422)
• In IDV+AZT/ 3TC arm, 65% still had <50 copies/ml (n=415)
• In EFV+IDV arm, 60% still had <50 copies/ml (n=429)

During the first 48 weeks of the study:

• In EFV+AZT/ 3TC--5% never achieved <50 copies/ml
• In IDV+AZT/ 3TC--12% never achieved <50 copies/ml
• In EFV+IDV arm--14% never achieved <50 copies/ml
• 8% had a confirmed viral load rebound to >50 copies/ml in EFV+AZT/ 3TC
• 15% in IDV+AZT/ 3TC arm had confirmed VL rebound to >50 copies/ml
• 19% in the EFV+IDV arm rebounded to >50 copies/ml

During weeks 49-96:

• an additional 7% rebounded >50 copies/ml in EFV+AZT/ 3TC arm
• additional 8% rebounded in IDV+AZT/ 3TC arm
• additional 7% rebounded in EFV+IDV arm

Levy reported on a more standard analysis Time-To-Treatment Failure (TTF), which they defined as the time until entry into the study until (see below). Levy said "treatment failures by this analysis include premature discontinuations and reversible AIDS defining events which do not necessarily limit future treatment options".

• a confirmed viral load >50 copies/ml
• an AIDS defining event (including CD4 <200 and herpes simplex virus infection)
• discontinuation from study for any reason
• never responded (I.e., never <50 copies/ml, TTF=0 days)

By this definition, 56% in EFV+AZT/ 3TC were still <50 copies/ml after 96 weeks, vs 40% in the IDV+AZT/ 3TC arm, and 36% in the EFV+IDV arm. But this analysis includes discontinuation for any reason which would include people who dropped out because they didn't want to be on IDV.

Levy estimated that if viral failure rates in subsequent years remain at 7-8% per years, the median time to failure with EFV+AZT/ 3TC will exceed 6 years.

In a preliminary look at the 3 year TVF analysis which only includes 754 study participants (about 250 in each arm), 80% were still <50 copies/ml in the EFV+AZT/ 3TC arm, and about 60% in the other two arms.

TVF 3-Year Follow-up
They did an analysis of data for subjects enrolled for more than 3 years and estimated the virologic failure in the 3rd year (96-144 weeks):

• EFV+AZT/ 3TC 3rd year failure rate = 3.5%
• EFV+IDV = 2.5%
• IDV+AZT/ 3TC = 12%

Based on this data, they said that if annual virologic failures were reduced to 3.5% in the EFV+AZT/ 3TC arm after the first two years, the median time to virologic failure would be about 10 years.

TVF by Baseline CD4; When To Begin Therapy
Levy showed data demonstrating baseline CD4s and viral loads were relatively similar between the 3 treatment groups. This data also showed that the viral loads were relatively equal between the 3 treatment groups within cd4 categories. For example, for people with cd4s 200-350, median cd4 was about 270 and VL was about 4.8 log in all 3 treatment groups.

Time-To-Virologic Failure By Baseline CD4: All Treatment Groups Combined
When pooling all three treatment arms, the data suggested a difference in response to treatment based on cd4 count at baseline. When looking at the data following this paragraph it appears as though the responses from people in the EFV/AZT+3TC arm were driving the data. This suggests that when looking at large retrospective databases like EuroSida there are differences between regimens & other differences (such as differences between individuals) that are not reflected in the final data & the final conclusions. Additionally, in this dataset as well as in others like EuroSida, possibly after longer follow-up the differences in response between CD4 groups would be greater. Levy showed a graph demonstrating that at week 96:

• About 78% with baseline CD4s 350-500 (n=302) had <50 copies/ml
• About 76% with baseline CD4s >500 (n=236) had <50 copies/ml
• About 64% with baseline cd4s 200-350 (n=345) had <50 copies/ml
• About 56% when cd4s were <200 (n=331) 56% had <50 copies/ml.

Levy said TVF was longer with EFV+AZT/ 3TC than with other two regimens across all cd4 groups. He cautioned that analyses within cd4 count subgroups were not prospectively planned and were not adequately powered to demonstrate differences within each group. Therefore, they concluded that there is an advantage to starting therapy when the CD4 count is >200. And there may be a small advantage when CD4 is >350.

Levy concluded that with EFV/AZT/ 3TC virologic failure was independent of baseline CD4. With EFV/IDV & with IDV/AZT/ 3TC virologic failure was higher if treatment was initiated when cd4 count is <200 (suggesting there was no difference between cd4 groups >200). With IDV/AZT/ 3TC, trend suggests that virologic failure rates may be higher when treatment is initiated with the cd4 count between 200-350 than if started when cd4 count is between 350-500.

Achieved CD4 Counts Based on Baseline CD4 Count: baseline CD4s matter, different findings than EuroSida data
This data suggests that cd4 increase is blunted when you start therapy with a lower CD4 count. This is opposite of the data from European databases presented by Cozzi & Phillips at Glasgow which shows CD4 & VL responses were the same if therapy started when CD4s were >200.

In the EFV+AZT/ 3TC treatment group, for patients who started therapy with CD4s >500 (n=68), CD4s at week 96 were 850. For patients who started with CD4s 350-500 (n=103), CD4s at week 96 were 600. For patients who started therapy with CD4s 200-350, they had 450 CD4s at week 96. And for patients who started therapy at <200 they achieved CD4s at week 96 of about 270. These findings were similar for the other two treatment groups.

Levy says patients who start treatment with lower cd4 count never achieve the same cd4 counts as patients who start treatment with higher cd4 counts, and there was no difference in cd4 count responses between treatment groups.

For patients who started therapy with <200 cd4s, 30% in the EFV/AZT/ 3TC, 28% in the EFV/IDV arm, and 41% in the IDV/AZT/ 3TC arm still had <200 cd4s at week 96.

Patients who started study when cd4s were 200-500 had rates of cd4s <200 of 0-4%.