Current antiretroviral guidelines recommend triple combination highly active antiretroviral therapy during pregnancy for the treatment of the mother and the prevention of mother-to-infant transmission of human immunodeficiency virus (HIV) (4
). During pregnancy, physiological changes occur that have the potential to change the pharmacokinetics of drugs (8
) and hence lead to decreased virologic efficacy. This study was initiated to determine the pharmacokinetics and safety of indinavir (IDV), coadministered with zidovudine (ZDV) and lamivudine (3TC), in HIV-infected women during pregnancy and postpartum.
Women with a singleton, uncomplicated pregnancy between 14 and 28 weeks of gestation according to menstrual and ultrasound dating were enrolled after signing institutional human subjects committee-approved informed consent statements. Exclusion criteria included prior treatment with a protease inhibitor of >3 weeks, nonnucleoside reverse transcriptase inhibitor use within 3 weeks, >3 weeks of prior 3TC usage unless taken for >3 months with a stable HIV RNA and CD4+
cell count, or intolerance to ZDV or 3TC. Women received 800 mg IDV three times daily according to standard instructions with ZDV (300 mg twice a day or 200 mg three times daily) and 150 mg 3TC twice daily through 12 weeks postpartum. During labor, IDV was temporarily discontinued, oral 3TC was continued, and ZDV was given intravenously (7
). Infants received orally 2 mg/kg 3TC twice daily and 2.6 mg/kg ZDV three times daily (or 2.0 mg/kg intravenously every 8 h) until 6 weeks. Infants did not receive IDV because of the concern for increases in indirect bilirubin concentrations.
Maternal and infant study visits for safety and efficacy evaluations occurred every 1 to 2 weeks antepartum, at delivery, at 6 and 12 weeks postpartum, and at 1, 6, 12, and 24 weeks of life, respectively. Adverse events were graded according to the National Institute of Allergy and Infectious Diseases Division of AIDS toxicity grading tables (1
). Maternal blood samples for IDV pharmacokinetics were obtained predose and at 0.5, 1.0, 1.5, 2, 4, 6, and 8 h postdose at 30 to 32 weeks of gestation and 6 weeks postpartum. Infant IDV plasma concentrations were obtained at birth and at 1, 2, 4, and 6 h of life. IDV plasma concentrations (as free base) were determined by a high-performance liquid chromatography-UV and, at later stages of the study, by the more sensitive high-performance liquid chromatography-mass spectrometry assay (University of California at San Diego PACTG Pharmacology Laboratory). For both assays, the inter- and intra-assay precision had a <13% coefficient of variation, and the accuracy had a <13% deviation from the expected. Pharmacokinetic parameters were estimated by noncompartmental analysis of the IDV plasma concentration-time profiles (WinNonLin 3.0; Pharsight, CA). With a sample size of 10 evaluable mothers, the study had approximately 80% power to show “equivalence” of maternal postpartum and antepartum area under the plasma concentration-time profiles (AUCs; defined as having a 90% confidence interval [CI] of the ratio of the geometric mean of antepartum to postpartum AUC in the range of 0.67 to 1.50).
Baseline characteristics of the 16 HIV-infected pregnant women and their infants and the adverse effects they experienced are provided in Table . Maternal adverse effects were similar in type and magnitude to those seen in nonpregnant individuals taking the same regimen (13
). Although two women had flank pain, renal ultrasound did not show kidney stones and there was no crystalluria. The major adverse event noted in infants was bone marrow suppression, previously described in infants receiving ZDV (7
). Of note, none of the women or their infants experienced grade 2 or higher episodes of indirect hyperbilirubinemia.
Summary of the baseline characteristics of women and infants enrolled in the study as well as their experience of adverse events
Among the 14 women who remained on study treatment through delivery and postpartum, the mean CD4 count increased from 358 cells/mm3 at baseline to 505 cells/mm3 at delivery (n = 12) and 605 cells/mm3 at 12 weeks postpartum (P < 0.001). Among the 12 women who had HIV-1 RNA evaluations at delivery, the number with levels less than 400 copies/ml increased from 4 (33%) at baseline to 10 (83%) at delivery.
Eleven women had complete, evaluable pharmacokinetic profiles (Table ). Except for time to reach maximum plasma concentration (Tmax), 95% confidence intervals for the geometric mean ratios (GMR) comparing postpartum to antepartum pharmacokinetic parameters did not encompass one, indicating that these parameters were significantly different postpartum versus antepartum. The postpartum pharmacokinetic parameters and their variability were comparable to those in nonpregnant women (AUC from time 0 to 8 h [AUC0-8], 1,124 ± 540 μg · min/ml; maximum concentration of drug in plasma, 8.37 ± 3.97 μg/ml; B.-Y. Nguyen, personal communication). Therefore, we conclude that the antepartum pharmacokinetics of IDV differ from those in nonpregnant HIV-1-infected women. Cord IDV plasma concentrations were evaluated from eight mother/infant pairs: they were below the assay limit of quantification for six, gave no detectable peak in one, and had a concentration of 39.9 ng/ml in one.
Geometric mean (95% CI), GMR, and arithmetic mean of the postpartum (6.3 ± 0.8 weeks) and antepartum (31.1 ± 2.4 weeks) indinavir pharmacokinetic parameters in HIV-1-infected women
The reduced exposure (AUC0-8
) to protease inhibitors during pregnancy observed here (Fig. ; Table ) and by others (6
) can be explained by several mechanisms. During pregnancy, the bioavailability (including the amount absorbed) of the drug may be lower, or the clearance of the drug may be increased. The oral clearance of IDV could be increased due to an increase in the free fraction of the drug in plasma due to pregnancy. However, this is unlikely, as IDV is not highly bound to plasma proteins (~60% protein bound). Therefore, its plasma protein binding would have to be reduced to almost zero during pregnancy to produce the change in oral clearance observed in our study. Such a change is physiologically unrealistic. Since IDV and the other protease inhibitors are predominately cleared by cytochrome P-450 (CYP) 3A4/5 metabolism (3
) and P-glycoprotein efflux (3
), their clearance through these pathways may be increased during pregnancy. Others have also suggested that intestinal and/or hepatic CYP3A activity is increased during pregnancy (10
). In the pregnant mouse, hepatic, but not intestinal, CYP3A activity and expression is increased, while hepatic and intestinal P-glycoprotein expression is unaffected (9
). Based on these studies, we propose that increased intestinal and/or hepatic activity of CYP3A activity is the most reasonable explanation for the enhanced oral clearance of protease inhibitors during pregnancy. Such an increase has significant ramifications for designing appropriate dosing regimens for pregnant women receiving narrow-therapeutic-window drugs cleared extensively by CYP3A. CYP3A enzymes clear more drugs on the market than any other drug-metabolizing enzyme. Therefore, we predict that dosing regimens of CYP3A-cleared drugs such as the anti-rejection drugs (e.g., cyclosporine), anti-epileptics (e.g., carbamazepine), cardiovascular agents (e.g., nifedipine), or antimicrobials (e.g., clarithromycin) will need to be adjusted upwards during pregnancy.
Area under the plasma concentration-time curves (AUC0-8) of indinavir in all but two women were lower antepartum versus postpartum. The horizontal solid line represents the arithmetic mean AUC0-8 values for the antepartum and postpartum periods.
Although viral suppression was documented in the majority of women who participated in this study, two women did not have viral suppression at delivery, one of which had very low exposure to IDV. There are significant concerns that utilizing dosing regimens of antiretrovirals yielding low plasma drug concentrations during pregnancy may be associated with lack of viral suppression as well as the development of resistance mutations (2
). It is possible that using standard doses of protease inhibitors for pregnant women could result in unnecessary cases of vertical transmission, with an additional concern for transmission of genotypically resistant virus to the infant. Antiretroviral agents utilized during pregnancy should be used in doses adequate to achieve plasma drug concentrations known to suppress the virus.
Since initiating this study, drug therapy of the HIV-infected pregnant woman has evolved (http://aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf
). The addition of low-dose ritonavir to IDV for dose “boosting” has been utilized in nonpregnant adults. Although the use of IDV during pregnancy is infrequent, based on these results, further study of the pharmacokinetics and safety of the combination of IDV plus ritonavir during pregnancy is warranted to provide an alternative agent for pregnant women intolerant of or not responding to other anti-HIV protease inhibitor regimens.