The objective of this study was to assess the value of quantitative HIV-1 R
NA as a predictor for the short-term risk of developing AIDS-defining event
s in comparison with CD4 cell counts, A total of 1,028 samples from 324 pat
ients mere analysed, Median initial CD? cell counts and HIV-1 RNA mere 249
x 10(6)/l (range 0-1400 x 10(6)/l) and 4.5 log copies/ml (range: 2.3-6.4 lo
g copies/ml). CD4 cell counts and viral load (VL) values obtained the year
before a single AIDS-indicator disease were selected to define the risk of
developing that event. Cox regression models with CD4 cell counts and VL va
lues treated as time-dependent covariates were performed to analyse the ris
k for developing certain events. Receiver operating characteristic (ROC) cu
rves mere used to compare CD4 cell counts and VL values as predictive marke
rs for progression. During a median follow-up of 870 d (range 30-1381 d), 1
32 patients developed AIDS, Median log VL values during the year before the
event were 3.6 for non-progressors and 5.2 for those who developed AIDS (p
< 0.0001). Minimum log VL threshold values for developing diseases were 2.
3 for tuberculosis, 3.8 for Candida esophagitis, 4.4 for wasting syndrome,
4.5 for CMV disease and 4.7 for PCP, VL values mere not, however, a better
predictive marker for developing specific events than were CD4 cell counts,
Although we have identified VL thresholds for the risk of developing certa
in AIDS-indicator diseases, the indication for starting prophylactic regime
ns may still be based on CD4 cell counts.