Background:  A recent study showed poor correlation of HIV RNA levels with subsequent CD4 decline. However, CD4 slopes were determined over a long period (upper quartile 4.6 years) and did not adjust for initial CD4 counts. Using data from the AIDS Link to Intravenous Experience (ALIVE) cohort and the Johns Hopkins HIV Clinical Cohort (JHHCC), this relationship and the predictive value of HIV RNA for clinical disease progression were investigated.

Methods:  A random-effects model was used to calculate CD4 slopes in semi-annual measurements from 197 seroconverters from ALIVE with data between 1988 and 1997. We determined the predictive value (R²) of the first HIV RNA value taken >1.5 years after the first HIV⁺ visit for subsequent CD4 slopes prior to clinical AIDS. Similarly short-term CD4 slopes within a specified time period (0.25, 0.5, 0.75, 1.0, 1.5, and 2 years) after initial HIV RNA measurement between 1996 and 2005 were determined for 392 patients free of clinical AIDS and not receiving any ART in the JHHCC. The R² for the first HIV RNA and CD4 count were determined for predicting subsequent CD4 slopes. Similar R² metrics were used to examine the predictive value of HIV RNA with time to clinical AIDS or death within the ALIVE data.

Results:  The R² of a single HIV RNA measurement for CD4 slope over a median of 3.18 years in ALIVE was minimal (R² <0.01) and for short-term CD4 slopes in the JHHCC data was <0.05. Initial CD4 explained a large proportion of CD4 slopes (p <0.001) especially over a short period of time (R² >0.76 for ≤0.75 years) even after adjusting for HIV RNA levels. The predictive value of a single HIV RNA measurement with time to clinical AIDS or death (n = 45; 23%) in ALIVE was substantially higher (R² = 0.32, p <0.002). Using time-updated HIV RNA values, the R² of HIV RNA was even stronger (0.61, p <0.002). After adjusting for time-varying CD4, the R² of HIV RNA were attenuated but remained high (0.19 and 0.42, respectively) and statistically significant from 0.01 (p ≤0.006).

Conclusions:  Our results confirm that HIV RNA has poor predictive value for CD4 slope alone and that any explanation for rate of CD4 slope should account for initial CD4. However, our analysis showing high predictive value for clinical disease events even after adjusting for CD4 levels reinforces the importance of HIV RNA as an independent marker of HIV disease progression.