Human immunodeficiency virus type 1 RNA level and CD4 count as prognostic markers and surrogate end points: a meta-analysis
A Babiker, J Bartlett, A Breckenridge, Gary L Collins, R Coombs, D Cooper, T Creagh, A Cross, M Daniels, J Darbyshire, D Dawson, V DeGruttola, R DeMasi, R Dolin, J Eron, M Fischl, S Grossberg, J Hamilton, S Hammer, P Hartigan, William K Henry, A Hill, M Hughes, J Kahn, C Katlama, D Katzenstein, S Kim, D Mildvan, J Montaner, M Moore, Jim Neaton, W O'Brien, H Ribaudo, D Richman, M Saag, M Salgo, L Saravolatz, R Schooley, M Seligmann, S Staszewski, L Struthers, C Tierney, A Tsiatis, S Welles, MD Hughes
2000-08-10
Abstract:Objective To evaluate treatment-mediated changes in HIV-1 RNA and CD4 count as prognostic markers and surrogate end points for disease progression (AIDS/death). Methods Data from 13,045 subjects in all 16 randomized trials comparing nucleoside analogue reverse transcriptase inhibitors and having HIV-1 RNA measurements at 24 weeks were obtained. A total of 3146 subjects had HIV-1 RNA and CD4 count determinations at 24 weeks after starting treatment. Results At Week 24, the percentage of subjects experiencing an HIV-1 RNA decrease of> 1 log 10 copies/ml or a CD4 count increase of> 33% was similar (22% vs 25%). Changes in both markers at Week 24 were significant independent predictors of AIDS/death: Across trials, the average reduction in hazard was 51% per 1 log 10 HIV-1 RNA copies/ml decrease (95% confidence interval: 41%, 59%) and 20% per 33% CD4 count increase (17%, 24%). In univariate analyses, the hazard ratio for AIDS/death in randomized treatment comparisons was significantly associated with differences between treatments in mean area under the curve of HIV-1 RNA changes to Weeks 8 and 24 (AUCMB) and mean CD4 change at Week 24, but, in multivariate analysis, only mean CD4 change was significant. Conclusions Change in HIV-1 RNA, particularly using AUCMB, and in CD4 count should be measured to aid patient management and evaluation of treatment activity in clinical trials. However, short-term changes in these markers are imperfect as surrogate end points for long-term clinical outcome because two randomized treatment comparisons may show similar differences between …