Background: The ability to predict long-term virologic and immunologic responses from measurements made during the first few days of antiretroviral therapy is of critical importance in order to prevent the emergence of drug-resistant HIV strains and to avoid unnecessary exposure to drugs. Currently, the efficacy of antiretroviral therapy is evaluated by observing the changes in levels of viremia many weeks after initiation of the therapy.

Methods: We analyzed the kinetics of plasma HIV-1 RNA, CD4 T lymphocytes and drug plasma levels in 41 HIV-infected children treated at four different dose levels of the protease inhibitor ritonavir. This dose-finding phase I/II study was performed by the HIV&AIDS Malignancy Branch, National Cancer Institute,Bethesda, MD; University of South Florida, St. Petersburg, FL;University of Florida College of Medicine, Gainesville, FL.

Results: Patients exhibited different types of responses as measured by the characteristic patterns of plasma HIV RNA concentration changes, and were divided into “good” responders (with undetectable or very low levels of virus at week 12) and “poor” responders (with high levels of plasma virus RNA). By fitting the kinetics of the initial response with mathematical models, the statistical analysis of their correlations and prediction of response by a cross-validation multivariate pattern recognition method, we were able to predict a relatively long-term (twelve weeks) response based on very short-term (one week or shorter) measurements. A combination of five parameters (two virus decay rate constants, the drug concentration in the blood plasma, the pre-treatment levels of plasma virus RNA and CD4 cell counts) contained sufficient information to predict the type of response and possibly clinical outcome in 40 of 41 pediatric patients treated with the HIV-1 protease inhibitor ritonavir.

Conclusions: Although performed during sub-optimal treatment conditions(monotherapy), these results offer an early indication of individual long-term responses suggesting that treatment could be optimized without waiting for clinical, virologic or immunologic deterioration.