Combination antiretroviral therapy with ritonavir and saquinavir has established potent and durable activity on plasma viremia. CNS HIV infection may be sequestered from drug therapy that does not penetrate the blood-brain barrier. Penetration of these protease inhibitors into the cerebrospinal fluid (CSF) and CSF HIV RNA levels on such therapy has not been well described.
Design/methods: In a cross-sectional study, 28 HIV1-infected study subjects were evaluated either before initiation of or before maximal response to ritonavir-saquinavir therapy, during maximal plasma virologic response, and after virologic failure. Simultaneous samples of plasma and cerebrospinal fluid were obtained from 24 study subjects to measure HIV RNA and protease inhibitor levels.
Results: Across the treatment groups, a strong correlation was found between plasma and CSF HIV RNA levels (r = 0.870; p < .001). In each study subject with plasma HIV RNA levels below assay limit (80 copies/ml), the CSF HIV RNA level was also below the limit of quantitation. Low levels of saquinavir (<2 ng/ml) and ritonavir (<25 ng/ml) in the CSF were observed, with a CSF:plasma drug concentration ratio of < or = 0.005 (0.5%) in all study subjects evaluated (n = 11). The plasma:CSF HIV RNA ratio was high before or early in treatment (median, 38; interquartile range [IQR], 13,97), but low (median, 0.29; IQR, 0.17, 7.5) in those failing therapy (group C, p < .001).
Conclusions: CSF ritonavir and saquinavir levels are consistent with the estimated known fraction of unbound drug in plasma (<2%). Across these treatment response groups, suppression of plasma viremia can predict low CSF HIV RNA levels. This correlation may represent HIV RNA transport and equilibrium between CSF and plasma, or it may represent CNS anti-HIV activity of protease inhibitors. The low drug levels and inverted ratio of HIV RNA in the CSF compared with plasma early in plasma virologic breakthrough suggests CSF virologic failure may contribute to failure of plasma virologic response.