Effect of ketoconazole on ritonavir and saquinavir concentrations in plasma and cerebrospinal fluid from patients infected with human immunodeficiency virus

Aim: Our aim was to evaluate the effect of ketoconazole on ritonavir and sa quinavir plasma and cerebrospinal fluid (CSP) concentrations. Methods: Twelve patients who were human immunodeficiency virus-seropositive and who were receiving 400 mg of ritonavir and 400 mg of saquinavir twice daily completed a nonfasted, two-period, two-group, longitudinal pharmacoki netic study. Blood samples were collected over the daytime 12-hour dosing i nterval of the protease inhibitors at baseline (period I, day 0) and after 10 days of coadministration of 200 mg (n = 6) or 400 mg (n = 6) of ketocona zole once daily (period 2, day 10). One set of paired CSF and blood samples was collected between 4 and 5 hours after the dose on both days. Results: Ketoconazole significantly increased area under the plasma concent ration-time curve, plasma concentration at 12 hours after the dose, and hal f-life of ritonavir by 29% (95% confidence interval (CI), 13%-46%), 62% (95 % CI, 37%-92%), and 31% (95% CI, 13%-51%), respectively. Similar increases of 37% (95% CI, 4%-81%), 94% (95% CI, 41%-167%), and 38% (95% CI, 15%-66%), respectively, were observed for these parameters for saquinavir. Ketoconaz ole significantly elevated ritonavir CSP concentration by 178% (95% CI, 59% -385%), from 2.4 to 6.6 ng/mL, with no change in paired unbound plasma leve l (26 ng/mL); this led to a commensurate 181% increase (95% CI, 47%-437%) i n CSF/plasma unbound ratio. All pharmacokinetic changes were unrelated to k etoconazole dose or plasma exposures. Corresponding changes for saquinavir CSF pharmacokinetics were insignificant (P > .06); saquinavir CSP levels we re unmeasurable in 7 patients (<0.2 ng/mL). Conclusions: The disproportionate increase in CSF compared with plasma conc entrations of ritonavir is consistent with ketoconazole inhibiting both dru g efflux from CSF and systemic clearance.