7th IAS Conference on HIV Pathogenesis, Treatment and Prevention (IAS 2013)


WEAB0104 - Oral Abstract Session

Efficacy of initial antiretroviral therapy: a meta-analysis of 40,124 adults with up to 144 weeks' follow-up

Presented by Frederick J. Lee (Australia).

F.J. Lee1, J. Amin2, A. Carr1

1Clinical Research Program, St. Vincent's Centre for Applied Medical Research, Sydney, Australia, 2The Kirby Institute, University of New South Wales, Sydney, Australia

Background: Meta-analysis of efficacy outcomes stratified by follow-up duration, baseline viral load (VL) and by DHHS 2012-'Preferred'/'Alternative' regimens have not been performed.
Methods: We evaluated initial ART studies in adults reporting intent-to-treat efficacy data (VL< 50 copies/mL) for ≥48 weeks. Variables abstracted: design (phase, placebo); eligibility (genotype, CD4 count, AIDS); ART (drugs, daily pills/doses); participants (baseline CD4/VL, demographics); and year commenced. Stratified and study size-weighted efficacy outcomes were calculated. Variables were analysed for association with efficacy in multivariable, size-weighted linear regression models using backwards, stepwise selection.
Results: We identified 114 studies (216 arms; 40,124 subjects): mean follow-up 82 weeks (28% for 96 weeks; 12% for 144 weeks); 76% men; 65% white; CD4 248 cells/mm3; VL log 4.9. ART (mean 6 pills/day in 2 doses) comprised a NRTI backbone (97%) with a NNRTI (49%), boosted PI (24%), unboosted PI (14%) or INSTI (5%). 30% of subjects were on 'Preferred' or 'Alternative' regimens (mean 3 and 6 pills/day, respectively). 45% reported efficacy stratified by VL (≥/< 100,000 copies/mL).

 Anti-viral efficacy, mean % (SD)Causes of ART cessation, mean %(SD)
 All studiesDHHS 2012-'Preferred'DHHS 2012 -'Alternative'All subjects with VL≥100,000All subjects with VL<100,000Withdrew / lost to follow-upAdverse eventVirological failureDeath
Overall60 (16)75 (8)65 (7)62 (15)70 (15)11 (7)8 (6)3 (4)2 (2)
Week 4866 (16)81 (5)71 (9)65 (13)74 (13)8 (5)7 (5)2 (3)1 (2)
Week 9660 (16)73 (11)66 (10)59 (16)63 (17)14 (8)8 (5)5 (5)2 (3)
Week 14452 (18)70 (4)63 (3)no data available15 (9)10 (6)5 (3)3 (4)
[Efficacy and cessation of initial ART]

Mean overall difference between 'Preferred' and 'Alternative' ART was 10% (95%CI:7.64-15.4, p< 0.0001) and between viral load strata was 8% (95%CI:6.00-10.90, p< 0.0001,). On multivariable analysis, no variable independently predicted greater efficacy within 'Preferred' regimens. Fewer pills/day predicted greater efficacy (coefficient 1.43, p=0.0008) within 'Preferred'/'Alternative' regimens. INSTI as the third drug independently predicted greater efficacy in subjects with VL≥100,000, adjusted for the year study commenced (reference NNRTI, coefficient 16.49, p=0.001).
Conclusion: Initial ART efficacy is mostly determined in white men and is suboptimal over 144 weeks, even with 'Preferred' regimens, particularly for VL≥100,000. The similar difference in efficacy between regimen types and between VL strata suggests the same consideration be given to initiating ART while VL< 100,000 as to regimen selection.

Back to the Programme-at-a-Glance