613   Factors Influencing Cerebrospinal Fluid Virological Response to Antiretroviral Drugs in Advanced HIV-1-Infected Patients.

A. Antinori*, M. L. Giancola, G. Liuzzi, S. Grisetti, F. Soldani, V. Tozzi, S. Calcaterra, F. Forbici, E. Girardi, M. Capobianchi, C. F. Perno, and G. Ippolito.
Natl. Inst. for Infectious Diseases, Lazzaro Spallanzani IRCCS, Rome, Italy.

Background:The objective of this study was to determine the effectiveness of HAART in treating cerebrospinal fluid (CSF) HIV-1 infection and to assess factors related to virological response.

Methods:We conducted a prospective study on consecutive HIV-1 patients with paired CSF and plasma samples collected. Subgroup longitudinal observation was performed. HIV-1 RNA concentrations were detected by Nuc­lisense HIV-1 QT assay.

Results:75 patients were enrolled, 55 (73%) with neurological diseases. The median CD4 count was 131 (IQR 47—317) cells/mmc. 28 (37%) pts. were naive for antiretrovirals. Median plasma and CSF HIV-1 RNA was 5.0 (4.2—5.8) and 3.5 (1.8—4.7) log₁₀cl, respectively. A significant correlation between plasma and CSF levels at baseline was observed in experienced (r = 0.483; P = 0.0005) but not in naive (r = 0.262; P = 0.17) samples. Neurological picture (OR 0.10; 95% CI 0.02—0.53), log₁₀copies of plasma viremia (OR 0.32; 0.17—0.59) and indinavir exposure (OR 6.62; 1.44—30.37) resulted associated with baseline CSF HIV-1 RNA level <80 copies. The longitudinal study included 29 patients with a reduction of CSF HIV-1 RNA between baseline and 12 weeks (mean -1.07 log₁₀c; 95% CI 0.42—1.71; P = 0.002). Plasma HIV-1 RNA change (r = 0.636; P = 0.0002), CSF HIV-1 RNA log₁₀c at baseline (r = -0.720; P < 0.0001), months of HAART (r = 0.643; P = 0.0001). and magnitude of difference between plasma and CSF HIV-1 RNA (r = 0.509; P = 0.004) were correlated to reduction of CSF HIV-1 RNA change during treatment. Differences in magnitude of CSF HIV-1 RNA reduction were observed according with naive status (-2.25 vs -0.23; P = 0.001) and use of >3 drugs penetrating BBB after starting or changing therapy (-1.73 vs +0.25; P = 0.01). But only naive status was independently associated with an increased probability of >1.0 log₁₀CSF reduction during therapy (OR 32.5; 95% CI 3.12—337.8).

Conclusions:Variable response to antiretroviral therapy was observed in CSF with a different compartmentalization of CSF infection. Naive status and use of CNS-penetrating drugs were required for effective viral reduction. Negative interaction with duration of antiretroviral treatment suggested long-term selection of drug-resistant CSF HIV-1 strains.