November 19, 1999
The penetration of antiretroviral agents into CSF is a potentially important determinant of drug activity against HIV-1 in the central nervous system (CNS). Because of the blood-brain barrier, most drugs do not enter freely into the CSF from plasma. Several factors influence the rate at which drugs penetrate the CSF, including size, charge, fat solubility, and the extent of binding to serum proteins. Because measuring drug concentrations in the CSF requires performing a lumbar puncture (LP, or "spinal tap"), relatively little information is available on the CSF pharmacokinetics of most antiretroviral agents. Previous studies have shown that relying on simultaneous measurement of drug concentrations in plasma and CSF can give misleading results, since peak CSF concentrations might occur later than peak plasma concentrations. Haas and colleagues have applied a novel method of analysis in which an indwelling catheter is placed in the epidural space to allow continuous sampling of CSF over several days. This approach uses techniques similar to those used in delivering spinal anesthesia or intrathecal chemotherapy.
In this study, Clough et al reported the pharmacokinetics of 3TC in CSF and plasma at steady-state [Abstract 330]. Four treatment-naïve patients initiated therapy with a 3TC-containing regimen using standard doses of 3TC (15 mg q12 hours). After four days, patients underwent frequent sampling of plasma and CSF over a 12-hour period. Peak 3TC concentrations(Cmax) were 6.1 mM and 0.37 mM in the plasma and CSF, respectively, giving a ratio CSF:plasma ratio of 8%. It is interesting to note that the Cmax of 3TC in CSF occurred approximately 3 hours after the plasma Cmax. The ratio of the 3TC area-under-the-curve (AUC) for CSF vs plasma was approximately 15%.
Similar data were generated for stavudine (d4T). The Cmax for d4T was 3.05 mM in plasma and 0.32 mM in CSF, for a CSF:plasma ratio of approximately 10%. Despite this difference in peak concentrations, CSF d4T concentrations actually exceeded plasma d4T concentrations for more than half the dosing interval. As with 3TC, the time to peak d4T concentration in the CSF was longer than the time to peak plasma concentration (3.5 hours vs 1 hour), suggesting a delay in entry of d4T into the CSF [Abstract 340].
Haas et al. also studied the steady-state PK of indinavir (IDV) in the CSF in eight patients who had achieved complete suppression of plasma HIV-1 RNA [Abstract 338]. Although peak CSF concentrations of IDV were only 6% of peak plasma concentrations (366 nM vs 5.45 mM, respectively), the estimated mean trough (Cmin) of IDV concentration in CSF was greater than the plasma trough concentration (93 nM vs 130 nM, respectively) and exceeded the IC95 for IDV (100 nM) throughout the 8-hour dosing interval in 5/8 subjects.
Another application of the frequent sampling technique is determining the kinetics of virus production in the CSF. Using four of the same patients as in the PK studies, Haas et al. characterized estimated the half-life and clearance of HIV-1 in CSF. Baseline HIV-1 RNA levels in CSF were slightly lower than in plasma (3.35-4.68 log10 copies/mL vs 4..63-4.87 log10 copies/mL, respectively). The estimated half-life of CSF virus ranged from 1.3-4.0 days, as compared to 1.1-1.9 days for plasma virus. Based on these data the investigators estimated that the rate of virus entry into the CSF varied from 74,000 to more than 2 million per day. These experiments could not determine, however, whether virus found in the CSF was produced locally in the central nervous system or crossed into the CSF from peripheral blood and lymphoid tissues.
What is the clinical significance of these results? The data generated by Haas's group demonstrate the feasibility of conducting form PK analysis of drug penetration into the CSF, and show that there is a lag of serveral hours between the time to peak plasma and CSF concentrations. These differences in time to peak concentration can have a substantial effect on calculating the CSF:plasma ratio of a drug. In addition, this system provides the potential for more detailed analysis of viral dynamics in the CSF, which may provide further insights into the best strategies for controlling this important reservoir HIV-1.