Session 17Oral Abstract Presentations HIV/SIV Vaccine Studies Session Day and Time: Wednesday 10 am - 12:30 pm Presentation Time: 12:00 Room: Ballroom B
84 What Constitutes Effectiveness for an HIV Vaccine that Ameliorates Viremia: Issues Involving Surrogate Endpoints in Efficacy Trials P. Gilbert*1, V. DeGruttola2, M. Hudgens1, S. Self1, S. Hammer3, L. Corey4 1Fred Hutchinson Cancer Res Ctr, Seattle, WA; 2Harvard Sch of Public Hlth, Boston, MA; 3Columbia Univ Coll of Physicians and Surgeons, New York, NY; and 4Univ of Washington, Seattle
Background: Initial HIV vaccines (vxs) are unlikely to prevent acquisition of HIV in all recipients. Moreover, several current vxs are designed to reduce viremia post-acquisition of infection. Efficacy trials of such vxs will evaluate vx effects on post-acquisition endpoints, including onset of antiretroviral therapy (ART) and surrogate endpoints (SEs) based on viral load and CD4 counts. Substantial and durable vx effects on certain post-acquisition endpoints should lead to vx licensure, for the vxs predicted effectiveness to ameliorate disease and abate the epidemic through reduced transmissibility; however, the key questions of how large, how durable, and which endpoints are unanswered. Two major challenges to resolving these questions are: 1) uncertainty about whether observed mid-term vx effects on SEs will translate into durable clinical vx benefits, and 2) accounting for ART of seroconverters.
Methods: Two (2) main SEs are proposed: preART viral load trajectories and a composite endpoint defined as either virologic failure (above a threshold) or ART initiation. Numerical simulations of a vx efficacy trial were conducted to identify effect sizes on the SEs that form a reliable basis for inferring clinically significant vx effects on progression and transmission. Rakai partners study data were used to model the vx effect on transmissibility predicted from a vx effect on the SEs, and MACS data were used to model the vx effect on progression predicted from a vx effect on the SEs.
Results: Substantial vx effects on SEs persisting > 1 yr post-infection are required for reliably inferring a significant vx impact on progression and transmissibility, e.g., the models showed that a vx effect to reduce the composite endpoint rate by > 50% over 1.5 yrs reliably predicts a > 2-fold vx reduction in transmissibility and a > 40% reduction in the rate of AIDS.
Conclusions: The composite endpoint can be used as a co-primary endpoint with HIV infection in efficacy trials, with minimum 50% vx effect reliably predicting clinical effectiveness. This endpoint usefully 1) has clear clinical interpretation; 2) can be analyzed using standard methods; 3) addresses ethical concerns by accommodating ART use; and 4) measures resource-saving capacity of the vx through its reductions in ART use. Phase 4 studies and long-term follow-up of Phase 3 seroconverters, with monitoring for vx resistance, are crucial for validating that vx effects on SEs lead to durable clinical vx effects.
