: In naive HIV-1 infected patients who start a highly active antiretroviral therapy (HAART), the relationship between time-updated CD4+ cell count, HIV RNA, and clinical progression (new AIDS-defining event or death) is incompletely understood. A 2-step statistical approach was adopted: first, modeling the evolution of the 2 markers taking into account left-censoring of HIV RNA and, second, studying their respective effect on clinical progression. The study sample consisted in 551 previously untreated patients of the Aquitaine Cohort who started their first HAART regimen between 1996 and 2000. During a median follow-up of 33 months, 46 patients experienced a new AIDS-defining diagnosis and 23 died. In multivariate survival analysis, time-updated CD4+ cell count (hazard ratio [HR] = 1,92 for 100 cells/mm3 lower, P < 10-4) and HIV RNA (HR = 1.30 for 1 log10 copies/mL higher, P = 0.04) on continuous scale were associated with clinical progression. When analyzing the effect of updated biomarkers using usual thresholds, the association with clinical progression was weaker for CD4+ but still significant (P = 0.007) whereas it remained only significant for updated HIV RNA above 4 log10 copies/mL (P = 0.01). The prognostic information of updated HIV RNA adjusted on updated CD4+ is significant but depends on how the markers are taken into account. Clinical decisions and interpretation of clinical trial results must weigh the signification of each of these 2 biomarkers.
(C) 2003 Lippincott Williams & Wilkins, Inc.