Objective: To determine the effectiveness of antiretroviral therapy in controlling cerebrospinal fluid (CSF) HIV-1 replication and to assess factors related to virological response in advanced patients.
Design: A cross-sectional and longitudinal study.
Methods: Consecutive paired CSF and plasma samples from HIV-1-infected patients were collected before starting or changing highly active antiretroviral therapy (HAART).
Results: In the cross-sectional analysis 75 patients were included, 55 (73%) with neurological disease, 28 (37%) naive for antiretroviral agents. A significant correlation between plasma and CSF levels at baseline was observed only in antiretroviral-experienced patients. The absence of neurological disease, lower plasma HIV-1 load and a previous exposure to indinavir were all associated with a baseline CSF HIV-1-RNA level less than 80 copies/ml at multivariate analysis. In 29 patients included in the longitudinal study a significant reduction in CSF HIV-1 RNA was observed. Plasma HIV-1-RNA change, CSF HIV-1-RNA level at baseline, overall months of antiretroviral treatment and the magnitude of difference between plasma and CSF HIV-1-RNA levels were all correlated to CSF HIV-1-RNA change during treatment. A significant difference in the magnitude of CSF HIV-1-RNA reduction was observed according to naive status and to the use of three or more drugs penetrating the blood-brain barrier.
Conclusion: HAART effectively reduces HIV-1 replication in CSF. A variable response to antiretroviral therapy was observed in CSF, reflecting a different compartmentalization of infection during treatment. Naive status and the use of CNS-penetrating drugs substantially enhance antiviral response. A negative interaction between virological response and the duration of antiretroviral treatment suggests long-term selection of drug-resistant CSF HIV-1 strains.