586 
An Updated Meta-analysis of Triple Combination Therapy in Antiretroviral-naive HIV-infected Adults
John Bartlett*1, M Fath2, R DeMasi3, J Quinn4, A Hermes2, and F Rousseau4
1Duke Univ, Durham, NC, USA; 2Abbott Labs, Abbott Park, IL, USA; 3Trimeris, Inc, Durham, NC, USA; and 4Gilead Sci, Foster City, CA, USA
Background: Triple
combination antiretroviral therapy (ART)
has dramatically improved prognosis for HIV-infected persons. With an
increasing number of treatment options for initial therapy, the optimal regimen
choice remains to be defined.
Methods: A systematic
overview was performed of clinical trials evaluating 3-drug ART in treatment-naïve HIV-infected adults. Triple
drug combinations included 2 nucleoside reverse transcriptase inhibitors (NRTI)
plus a protease inhibitor (PI), boosted PI (BPI), non-nucleoside RTI (NNRTI), or a third NRTI. The primary endpoint
for analysis was the percentage of subjects with plasma HIV RNA (RNA) < 50
copies/mL at week 48. Intent-to-treat (ITT, M=F) analyses were used for virologic
responses, and as-treated analyses for changes in CD4 cell counts. Pairwise comparisons of RNA and CD4 cell responses were
made between drug classes, and predictors were
analyzed using weighted multiple linear regressions.
Results: We included 64
clinical trials with 10,559 subjects evaluating 85 treatment arms. Median
baseline RNA and CD4 cells were 4.79 log10 copies/mL and 285/mm3. Overall 55% achieved RNA < 50
copies/mL at week 48. In contrast, a similar but
smaller study published in 2001 found that only 47% reached this important
threshold. Significantly greater percentages reached RNA < 50 copies/mL at week 48 in the BPI (64%) and NNRTI groups (63%) than
in the PI (44%) and NRTI (51%) groups, (p
< 0.05). Overall, mean CD4 cell increases at week 48 were 180 cells/mm3,
and increases were greatest with BPI (209 cells/mm3) compared to
NNRTI (180 cells/mm3, p = 0.003),
PI (178 cells/mm3, p = 0.002),
and NRTI (150 cells/mm3, p
< 0.001). Pill burden varied from 3 to 22 pills per day across the 85
treatment arms. In multiple regression analyses, pill burden was not
consistently associated with outcome, and the identified associations varied by
drug class, the endpoints analyzed, and modeling techniques.
Conclusions: Responses on ART in clinical trials for ART-naïve
patients have improved over time. A modest overall trend towards higher
response with lower pill burden was observed, but was inconsistent across drug
classes, endpoints, and methods of analysis, and some
analyses showed no correlation between pill burden and response. BPI- and
NNRTI-containing regimens provided superior virologic suppression compared to
PI- and NRTI-containing regimens, and BPI-containing regimens offered the
greatest increase in CD4 cells.
Keywords: Treatment-naive; Antiretroviral therapy; pill burden
