Background:  Recently it was reported that plasma HIV-1 RNA explained <10% of the variability in short-term CD4 cell decline. This result led the researchers to question the importance of viremia in AIDS pathogenesis and to conclude that HIV-1 RNA has “limited clinical value.” To address the validity of these assertions, we analyzed the predictive value of HIV-1 RNA, CD4 cell counts, and CD38 relative fluorescence intensity (RFI) on CD8 T cells for time to AIDS progression and CD4 cell decline.

Methods:  HIV-1 RNA (real time polymerase chain reaction [RT-PCR], or bDNA equivalent) and CD4 cell counts were measured among 1679 HIV-1⁺ men enrolled in the Multicenter AIDS Cohort Study (MACS). CD38 RFI on CD8 T cells was measured in 214 of these participants. AIDS (1993 definition) developed by 1991 in 651 of the 1679 participants and in 77 of the 214 with CD38 RFI data. CD4 cell slope (>2 measurements) was estimated using data from 1985 to mid-1988. Variation around CD4 regression lines was assessed by the root mean square error (RMSE). Coefficients of determination (R²) for censored survival data were derived from log normal regressions and were used to quantify the percentage variability in time to AIDS explained by markers.

Results:  Overall, a single HIV-1 RNA measurement was the best predictor, explaining 45% of the variability in time to AIDS, compared with only 20% for CD4 cell count. The 2 markers combined explained 51%. Among the 214 participants, the percentage of variability in time to AIDS explained by HIV-1 RNA, CD4 cell count, and CD38 RFI were 52%, 11%, and 46%, respectively. CD38 RFI was strongly correlated with HIV-1 RNA (r = 0.63; p <0.001). HIV-1 RNA and CD38 RFI jointly explained 57%, and adding CD4 cell counts had no effect. By contrast, HIV-1 RNA, CD4 cell counts, and CD38 RFI explained very little of variability in CD4 cell decline: 4.5%, 9.8%, and 1.4%, respectively. The variance around individual CD4 regression lines was very high; the mean RSME (121 cells/µL) substantially exceeded the average CD4 cell slope (–71.5 cells/µL/year).

Conclusions:  A single HIV-1 RNA measurement was the best predictor of time to AIDS, explaining almost half of the variability in this clinically relevant outcome. High variance around CD4 regression lines likely accounts for the inability of any marker to explain CD4 cell decline. The predictive value of HIV-1 RNA for time to AIDS reaffirms the central role of viremia in AIDS pathogenesis and strongly supports the use of HIV-1 RNA for patient management.