icon_folder.gif   Conference Reports for NATAP  
 
  ICAAC
Interscience Conference on Antimicrobal Agents and Chemotherapy
December 16-19, 2005 Washington DC
Back grey_arrow_rt.gif
 
 
 
African-Americans Have Faster Viral failure in 5095
 
 
  Reported by Jules Levin
ICAAC Dec 17, 2005 Wash DC
 
ACTG A5095:
"Phase III, randomized, Double-Blind Comparison of 3 Protease Inhibitor sparing Regimens for Initial Treatment of HIV-1 Infection"
FINAL RESULTS
 
Roy Gulick (Weill Medical College of Cornell University) presented this talk. This report will focus on the emphasis of his talk, which was the comparison of the 3 & 4 drug efavrienz containing regimens in this three-arm study. The comparison of the third arm in this study, which was discontinued by the DSMB (Data Safety Monitoring Board) due to not performing as well as the 3-drug efavirenz, was reported on previously (Gulick et al, NEJM 2004;350:1850-61). Gulick summarized the findings: in treatment-naive subjects, there was no difference between AZT/3TC + EFV and AZT/3TC/ABC +EFV over about 3 years of followup with regards to initial viral response, rates & times to viral failure, CD4 response, grade 3-4 adverse events, self-reported adherence, and resistance at viral failure. RACE/ETHNICITY EFFECT: Blacks (but not Hispanics) had a higher rate of virologic failure than whites. This could be explained by an interaction between race and adherence - blacks who self-reported non-adherence at week 12 on EFV-based regimens had a shorter time to virologic failure compared to other groups. Gulick said further genetic and adherence analyses are in progress. (see discussion below).
 
During the Q&A period following the talk a number of questions were and included a comment that perhaps the pattern of reporting non-adherence was different. However, the discussion ended without any better understanding of the reported finding which I think left everyone in the audience perplexed.
 
The two main points of my report are: (1) there was no difference in efficacy found between the 3 and 4 drug EFV regimens, and you can read the data reported immediately below; (2) perhaps of more interest is the difference in Time To Virologic Failure between whites and blacks, which is reported below.
 
THREE STUDY REGIMENS
--AZT/3TC/ABC (3 nukes)
--AZT/3TC+EFV (control arm)
--AZT/3TC/ABC + EFV (4 drugs)
 
2 EFV-Based Regimens
 
Study subjects
N=765
81% men
41% white, 35% black, 21% Hispanic, 2% other
11% with IDU history
Mean baseline:
-HIV RNA: 4.86 log (72,000 copiesml)
57% <100,000 c/ml
43% >100,000 c/ml
CD4: 240
MEDIAN FOLLOWUP: 144 weeks (about 3 years)
 
Disposition
765 randomized
595 (78%) completed the study
154 (20%) prematurely discontinued study f/u
16 (2%) died
--no difference in premature study d/c between study groups (p=0.56)
758 started randomized study rx (99%)
463 (61%) were taking initial rx at study completion
108 (14%) d/c study rx, but continued f/u
103 (14%) d/c study rx and study f/u
--no difference in initial treatment d/c between groups (p=0.45)
 
Drug Substitutions
Protocol permitting substitutions:
--69 (9%) substituted d4T for AZT
--25 (7%) substituted ddI for ABC
--75 (10%) substituted NVP for EFV
 
RESULTS
 
Time To Virologic Response
(confirmed HIV- RNA <200 c/ml)

The curves of viral response were virtually overlapping through the first 24 weeks & the entire study followup reported with both groups achieving 90% < 200 c/ml at week 24. There were no obvious differences between the regimens, Gulick commented.
 
Primary Efficacy Endpoint
 
193 (25%) of subjects reached protocol defined virologic failure
--99 (26%) of 382 on AZT/3TC + EFV
--94 (25%) of 383 on AZT/3TC/ABC + EFV
 
NO STATISTICAL DIFFERENCE BETWEEN 2 ARMS
p=0.73, log rank test
HR 0.95; 97.5% CI=[0.69, 1.31]
 
The Kaplan Meier Curve also demonstrated no differences in Time To first Virolog Failure. Median f/u was 144 weeks (p=0.73; no difference).
 
When repeating this analysis with the subgroup of patients with baseline VL of >100,000 c/ml there was no significant difference (p=0.87; HR=1.03 [0.64, 1.67]).
 
MULTIVARIATE ANALYSIS
When looking at various covariates and their association with virologic failure, only 2 factors showed significance: black ethnicity & HCV antibody positive.
 
Gulick noted: Significant associations were seen overall with Race/ethnicity and this was specifically due to that black non-hispanic patients had a shorter time to virologic failure than white non-hispanic patients with a Hazard Ratio of 1.67 and this reached a high level of significance. No other race/ethnicity differences were observed. Patients with HCV+ antibody at baseline also had a shorter time to virologic failure with a Hazard Ratio of 1.57.
 

covariate-1.gif

Proportion of subjects with HIV RNA <200 and <50 cps/ml (ITT)
 
No difference in response rates were seen, with patients in both groups acheiving 85-90% <200 c/ml durable through 144 weeks of followup; and 80-85% of patients reducing VL to <50 c/ml and durable through followup.
 
CD4 Cell Responses
 
There were also no differences in CD4 count response between the two arms. By week 144 CD4 increases were about 300.
 
Adverse Events
 
Signs/symptoms:
--grade 3: 137 (18%)
--grade 4: 22 (3%)
no difference between rx groups (p=0.25)
 
lab abnormalities:
--grade 3: 178 (23%)
--grade 4: 153 (20%)
no difference between rx groups (p=0.90)
 
Suspected hypersensitivity reactions (HSF)
--37 (10%) on AZT/3TC/ABC + EFV
--28 (7%) on AZT/3TC + EFV (+ABC placebo)
 
Resistance Results
Of the 193 patients who experienced virologic failure, at baseline 89% had wild-type virus at baseline and 9% had drug resistance; genotypic testing was performed retrospectively. Of these 9%: 3% (n=5) had NNRTI resistance; 2% (n=4) had NRTI+NNRTI resistance; 2% (n=4) had NNRTI+PI resistance; 1 patient had 3TC only resistance; 1 patient had NRTI resistance; 1 patient had NRTI+PI resistance; 2 patients had NRTI+NNRTI+PI resistance. Sequence was not available for 3 patients.
 
Resistance At Time of Viral Failure for Patients Who Did Not Have Pre-existing Resistance (n=172)
Genotype available for 142 (83%).
No sequence was attempted on 24 patients because VL<500. 6 patients samples could nor=t be sequenced.
 
At the time of viral failure the most common type finding was wild type (n=37 in 3-drug, n=34 in 4-drug regimens).
 
NNRTI resistance alone: 15 patients in 3-drug & 22 in 4-drug regimen.
M184I/V + NNRTI: 13 in 3-drug & 7 patients in 4-drug regimen.
M184I/V + NRTI + NNRTI: 1 in 3-drug 7 3 in 4-drug regimen
M184I/V only: 4 in 3-drug 7 2 in 4-drug.
 
Gulick reported there were NO SIGNIFICANT differences in resistance (p=0.31)
 
Self-Reported Adherence
was not different between the 2 groups. About 85% of patients in both study groups at all times points evaluated in the study reported never missing a dose (based on 4-day recall; in an as-treated, missing=ignored analysis)
 
Race/Ethnicity Differences
 
In order to investigate the finding that blacks had a shorter time to virologic failure than whites, a post-hoc secondary analysis was performed.
 
Findings:
 
Viral response rates were good in all groups, but there appeared to be differences:
At week 144, the proportion with VL <50 c/ml was 80% (blacks), 91% (Hispanic),
89% (white).
 
The most common reason for d/c was viral failure at 27%.
 
Blacks (but not Latinos) compared to whites had:
--significantly less baseline drug resistance
--significantly shorter time to both initial regimen discontinuation and to grade 3-4 adverse events
--no difference in self-reported adherence (based on 4-day recall)
--significantly more resistance (M184I/V + K103N) at viral failure
--similar results were found when either this 2 or the original 3 study arms were compared.
 
Time To First Virologic Failure
(by race and week 12 self-reported adherence)

This graph was perhaps the most interesting but perplexing of the entire talk.
Patients were divided by ethnicity and adherent or non-adherent;
White (n=235/32)
Black (n=176/50)
Hispanic (n=111/25)
 
So, there were 6 lines on the graph representing each of 6 patient groups, and the finding was that: blacks who reported non-adherence at week 12 had a significantly shorter time to virologic failure than all other 5 patient groups.