Pharmacokinetics of indinavir in HIV-positive pregnant women

Since 1997, guidelines for the care of HIV-1-infected individuals stated that women should receive optimal antiretroviral therapy regardless of their reproductive status [1]. Highly active antiretroviral therapy (HAART) can effectively reduce viral load and limit the evolution of viral resistance. When used during pregnancy, HAART has the theoretical potential of reducing the risk of perinatal transmission [2]. However, information on the pharmacokinetics of these regimens during pregnancy is limited, especially concerning protease inhibitors (PI) such as indinavir (IDV). The need for ongoing clinical trials to evaluate the safety, tolerance, and pharmacokinetics of HAART in this patient population is clear [3]. Pregnancy produces unique physiological changes, including alterations in hepatic and placental drug metabolizing systems involving the cytochrome P450 isoenzymes, which affect drug disposition. Elevated levels of endogenous glucocorticoids, peaking during the third trimester, may induce CYP3A4 [4], the predominant enzyme responsible for the metabolism of IDV and other PI [5]. Our group further demonstrated the induction of CYP3A4 during pregnancy (unpublished data). Consequently, pregnancy may cause reduced PI plasma concentrations, resulting in sub-therapeutic exposure. We describe two cases of HIV-positive women followed from their second and third trimesters of pregnancy up to 3 months postpartum. Both patients were treated with HAART including IDV. The first case is a 35-year-old woman, stabilized with lamivudine (150 mg twice a day), stavudine (20 mg twice a day), and IDV (800 mg every 8 h) upon enrolment. The viral load and CD4 cell count at 19 weeks' estimated gestational age (EGA), one month before enrolment, were 65 copies/ml (ultrasensitive modification of Roche Amplicor HIV-1, Roche, Nutley, NJ, USA) and 924 cells/mm3, respectively. Her regimen remained unchanged during pregnancy and her viral load increased to 390 copies/ml at the third trimester (week 33), decreasing to 215 copies/ml postpartum. She delivered an HIV-uninfected infant in March 1999. Serial pharmokinetic plasma samples (0–8 h) were collected at 21, 33 and 38 weeks EGA, and 9 weeks postpartum to estimate IDV exposures. The urinary 6-beta-hydroxycortisol : cortisol ratio was also determined from urine collected pre-dosing to assess the activity of CYP3A [6], which catalyses the conversion of endogenous cortisol to 6-beta-hydroxycortisol. The second case was a 36-year-old woman enrolled at 32 weeks' EGA. She had been receiving zidovudine (300 mg twice a day), lamivudine (150 mg twice a day), and IDV (800 mg every 8 h) since February 1997. Her viral load was undetectable (< 500 copies/ml) at third trimester and postpartum visits. She delivered an HIV-uninfected infant in July 1998. Serial pharmacokinetic plasma samples were collected at week 34 and 12 weeks' postpartum. Liver and kidney function for both women remained normal throughout the observation period and no additional drugs known to affect CYP3A4 were received. The pharmacokinetic parameters of IDV and the urinary 6-beta-hydroxycortisol : cortisol ratio are summarized in Table 1. Plasma concentrations of IDV were determined by liquid chromatography tandem mass-spectrometry assay [7]. Urinary 6-beta-hydroxycortisol and cortisol were determined by high performance liquid chromatography assay [6], modified for application to samples from pregnant women. A marked difference was observed in plasma IDV area under the curve (AUC0−−8) between the third trimester and postpartum in both cases, despite an unchanged IDV dose. Although the AUC0−−8 from postpartum visits corresponded to reported values (19 300–42 098 nM × h) (CrixivanTM product information), AUC0−−8 at third trimester visits were reduced from the postpartum values by 63 and 86% in cases 1 and 2, respectively. Similar reductions were observed in the maximum plasma concentrations. The observed decrease in plasma concentrations of IDV could be attributed to the increased clearance and volume of distribution during the third trimester. The change in IDV clearance appeared to correspond to an increase in the urinary 6-beta-hydroxycortisol ratio, suggesting that CYP3A4 activity is enhanced during late pregnancy. Interestingly, a third case involving a women at 31 weeks' EGA on IDV (800 mg q12h) and ritonavir (200 mg q12h), a potent CYP inhibitor, produced an IDV AUC0−−12 five times that measured in women on IDV alone at similar weeks' EGA. Reduced absorption, placental drug passage, and protein binding changes may also contribute to lower plasma IDV concentrations.Table 1: Pharmacokinetic parameters of indinavir and urinary 6-beta-hydroxycortisol : cortisol ratio. These cases suggest that pregnant women may be exposed to sub-therapeutic drug levels during the later stages of pregnancy. This may require doses of PI higher than standard regimens or the use of dual PI including ritonavir. Although maternal viral load remained low during the course of our observations, long-term clinical outcomes remain uncertain. Anecdotal reports of viral breakthrough, while taking HAART, during the postpartum period remains a serious concern in new mothers. The striking pharmokinetic results of these two cases may provide a preliminary explanation for these reports. Further study is required to define the pharmokinetic changes during pregnancy to optimize maternal health and the long-term use of HAART. Sandra Hayashia Karen Beckermanb Masato Hommac Bradley W. Kosela Francesca T. Aweekaa