Antiretroviral drug exposure in the female genital tract: implications for oral pre- and post-exposure prophylaxis

Study design and population

HIV-infected, non-pregnant adult women were enrolled in a non-blinded, observational, open-label prospective, pharmacokinetic study to assess first dose and steady-state antiretroviral drug exposure in cervicovaginal fluid (CVF) relative to blood plasma (BP) from November 2002 to May 2006. Subjects were recruited from the Infectious Diseases Clinic at the University of North Carolina at Chapel Hill. Exclusion criteria included the following: ≤ 17 years of age; active bacterial, fungal, or opportunistic GT or systemic infection at the time of enrollment; hysterectomy; pregnancy; hematocrit less than 30%; unable to abstain from sexual intercourse or douching for 48 h prior to study visits; unable to complete a dosing card; and an estimated drug adherence rate of <80%, as determined by dosing card and patient self-report.

The antiretroviral drug regimen was selected by the patient’s primary HIV provider. Women could be antiretroviral naive and initiating antiretroviral drugs, changing a failing regimen, or re-initiating antiretroviral drugs after a therapy interruption. All patients gave informed consent prior to study participation. This protocol was approved by the Biomedical Institutional Review Board of the University of North Carolina at Chapel Hill.

Before being enrolled into the study, patients underwent a routine pelvic examination, physical examination, extensive sexually transmitted infection (STI) screening, and routine safety laboratory tests, including a complete blood count, serum chemistries, liver function tests, and serum pregnancy testing. STI screening included testing for chlamydia and gonorrhea by cervical polymerase chain reaction (PCR), testing for trichomonas by urine culture and PCR [11], and testing for syphilis by serum rapid plasma reagin (RPR) with confirmatory treponemal testing if positive. STI testing was performed in the University of North Carolina Sexually Transmitted Diseases Cooperative Research Center Microbiology Core Lab or the McLendon Laboratories of the UNC Hospitals. Any patient with a STI at screening was referred to her primary provider for treatment, and then offered re-screening after successfully completing STI therapy.

Study visits

Following acceptable screening results, patients were admitted for an overnight stay in the Verne S. Caviness General Clinical Research Center. The first dose of the new antiretroviral drug regimen was observed, and six paired BP and CVF samples were obtained (‘first dose’ conditions). At least 3 weeks later, after continuing the regimen (‘steady-state’ conditions), patients returned to the research center whereby the first-dose BP and CVF sampling strategy was repeated after an observed dose. Patients were also followed monthly for an additional 6 months, with BP and GT samples obtained at a single time points 2 to 24 h postdose.

Sample collection and processing

Under both first dose and steady-state conditions, CVF and BP samples were obtained at 0 (immediately predose), 2, 4, 6, 12, and 24 h after observed doses of the antiretroviral drug regimen. All CVF samples were obtained via direct aspiration with a volumetric aspiration device. Subjects were instructed to remain supine for approximately 10 min prior to collection in order to pool CVF in the posterior fornix of the vagina for optimal collection volume. Samples at 0 and 24 h were obtained during a pelvic examination, and the remaining samples at 2, 4, 6, and 12 h post-dose were self-collected. CVF volumes typically ranged from 0.1 to 0.4 ml per sample. A previously conducted pilot study [12] demonstrated this method to be a reproducible and robust approach to specimen collection. To minimize any HIV RNA and drug concentration variability, all visits were scheduled in the 7 to 10 days following the end of menses.

After collection, directly aspirated CVF was transferred to labeled cryovials and stored at −80°C until analysis. Whole blood was obtained using K₃EDTA-containing collection tubes (Becton Dickinson Diagnostics, Franklin Lakes, New Jersey, USA) and centrifuged at 1300 g (2600 rpm) at 4°C for 15 min. The resulting plasma was aliquoted into labeled cryovials and stored at −80°C until analysis. Drug concentrations in BP were measured using validated high performance liquid chromatography (HPLC)/UV methods [13–15], and concentrations in CVF were quantified using a validated HPLC-mass spectrometry (MS)/MS method [16].

Briefly, CVF concentrations were measured using a simultaneous assay for 17 antiretroviral drugs. After thawing, samples were centrifuged and the resultant supernatant underwent solid phase extraction using BOND ELUT C-18 columns (Varian, Harbor City, California, USA) as previously described [15]. Cimetidine (in acetate buffer, pH 5.0) was used as internal standard, and was applied directly to the conditioned column prior to CVF introduction. A Shimadzu solvent delivery system (Columbia, Maryland, USA) and a LEAP HTC Pal thermostatted (6°C) autosampler (Carrboro, North Carolina, USA) connected to an Applied Biosystems API4000 triple quandruple mass spectrometer and Turbospray ion source (Applied Biosystems, Foster City, California, USA) with an Aquasil C18 column (Thermo-Electron, San Jose, California, USA) was used for the analysis. Multiple reaction monitoring and positive-to-negative polarity switching were used. Assay sensitivity was 1 ng/ml for abacavir (ABC); 5 ng/ml for lamivudine (3TC), zidovudine (ZDV), emtricitabine (FTC), lopinavir (LPV), atazanavir(ATV) and efavirenz (EFV); and 10 ng/ml for tenofovir (TDF), didanosine (ddI), stavudine (d4T) and ritonavir (RTV). Overall assay precision was 2.0–14.3 CV%, and accuracy was 88–113%. Recovery for the drugs studied ranged from 80% for RTV and LPV to 99% for 3TC, ddI and ABC.

All analytical work was performed by the UNC Center for AIDS Research (CFAR) Clinical Pharmacology and Analytical Chemistry Core, which participates in quarterly national and international external proficiency testing [17,18]. These results consistently demonstrate high levels of accuracy and precision for our antiretroviral assays.

Data analysis methods

Pharmacokinetic parameters, including the area under the time–concentration curve (AUC_0−τ), were estimated for both CVF and BP using WinNonlin (version 4.0.1, Pharsight, Inc. Mountain View, California, USA). For these computations, concentration measurements below the lower limit of detection were imputed as zero and those below the lower limit of quantitation (LLQ) were imputed as ½ LLQ. For each antiretroviral agent, the GT: BP AUC_0−τ ratios were calculated, and multiplied by 100 to represent penetration of drug into the GT relative to BP [19]. Descriptive statistical methods, particularly medians and the interquartile range (IQR), were used in the primary analyses of these AUC ratios. The 95% confidence intervals of the median AUC ratios for each drug at each visit were calculated using Intercooled STATA Release 8.0 (Stata Corporation, College Station, Texas, USA).

Demographics

Demographic data for the 27 women enrolled are presented in Table 1. The study population was predominantly African American (70%), middle-aged (median: 35 years; IQR: 31–42 years), with a BP HIV RNA of 4.7 (IQR, 4.0–5.0) log₁₀ copies/ml and a CD4+T-cell count of 307 (220–372) cells/μl at study entry. As most subjects were antiretroviral experienced (67%), the provider-selected regimens varied widely. Seventy percent (19/27) of the regimens contained 3TC. TDF was found in 56% of the regimens, ABC in 30%, and ZDV in 24%. Thirty-three percent of the women were on EFV-containing therapy, and 52% were on a PI-containing regimen (57% ATV/r, 36% LPV/r, and 7% ATV); one woman (4%) received EFV with LPV/r. Two individuals (7%) were on triple nucleoside therapy, and one (4%) received nevirapine. (Nevirapine data not presented.)

Antiretroviral pharmacokinetics in blood plasma and genital tract

Using the direct cervicovaginal aspirate sampling method, all drugs were detected in the GT 2–4 h after the first dose, and reached peak concentrations for the dosing interval by 4–6 h. Figure 1a–g depict individual pharmacokinetic profiles of selected antiretroviral drugs over their respective dosing intervals in BP and GTafter a first dose and at steady state. These drugs were selected as achieving the highest concentrations in their respective drug classes, using doses and/or dosing regimens most commonly prescribed.

Nucleoside/tide reverse transcriptase inhibitors

Figure 1a depicts extracellular ZDV concentrations over a 12 h dosing interval. After a first dose, ZDV concentrations were detectable in both BP and GT secretions 2 h after a dose. Although BP concentrations steadily declined from that time point, GT concentrations continued to increase and reached a maximum concentration at 6 h post-dose. At steady state, concentrations were detectable in the GT, but not in BP, prior to dosing. The GTand BP concentrations were similar 4 h postdose; after this point, GT concentrations declined minimally, while BP concentrations declined significantly. An average extracellular BP concentration (C_av) previously suggested for efficacy [20,21], is indicated in the figure as a black horizontal line. GT concentrations were at or near this target for a longer period of time than BP concentrations.

Lamivudine exposure after once-daily dosing can be seen in Fig. 1b. Similar results were seen for subjects receiving twice daily dosing (not shown). GT concentrations were approximately 1-log higher than BP concentrations throughout the dosing interval after a single dose and under steady-state conditions. The GT concentrations exceeded the suggested average extracellular target concentration (C_av) previously suggested for efficacy [20,21]. Similarly, GT concentrations of FTC were higher than BP concentrations from 6–24 h after a single dose, and remained higher than BP concentrations during the entire dosing interval under steady-state conditions (Fig. 1c). For TDF (Fig. 1d), the GT concentrations were at or above BP concentrations 4 h after a dose, and remained higher throughout the dosing interval.

Protease inhibitors

Pharmacokinetic profiles for ATV and LPV are shown in Fig. 1e and f. In contrast to the NRTIs, ATV GT concentrations (when given as 300 mg ATV/100 mg RTV once daily) were lower than BP throughout the dosing interval both after first dose and at steady-state. These concentrations still, however, exceeded a suggested BP trough concentration of 150 ng/ml for efficacy against wild-type virus [22]: the mean (±SD) ATV trough concentration in the GT after a first dose was 417 ± 818 ng/ml, and at steady state was 169 ± 148 ng/ml. The GT concentrations for LPV were also lower than BP at both first dose and steady-state. A suggested LPV trough concentration for efficacy of 1000 ng/ml is depicted in Fig. 1f [22]. The GT concentrations approached this target over a dosing interval, and briefly exceed this concentration at the 4-h time point (mean ± SD = 1522 ± 1659 ng/ml). At the end of the dosing interval, mean ±SD GT concentrations were 437 ± 631 ng/ml after a single dose and 354 ± 384 ng/ml at steady-state.

Nonnucleoside reverse transcriptase inhibitors

The summary EFV pharmacokinetic profiles are shown in Fig. 1g. The BP and GTexposures are clearly distinct. Whereas EFV was detected in the GT 2 h after a single dose, concentrations were approximately 10 ng/ml: 2 logs lower than BP. The suggested target trough concentration for EFV is 1000 ng/ml [22].

Pharmacokinetic parameters

Table 2 lists steady-state pharmacokinetic parameter estimates in BP and GT for each drug. BP concentrations at the end of the dosing interval (Cτ) and the AUC over 24 h (AUC_0–24h) for each drug were within the expected range. On average, GT time to maximal concentration (T_max) was delayed in comparison with BP. All drugs were, however, detectable in GT secretions within 4 h of antiretroviral drug ingestion.

Table 3 lists median (IQR) GT: BP AUC ratios in rank order from the highest exposure to lowest exposure after single and multiple doses of antiretroviral drugs. A ratio equal to 100% indicates drug exposure in the GT that is equivalent to BP. Therefore, ratios greater than 100% indicate greater GT drug exposures than BP, and those less than 100% indicate lower GT exposures than BP.

ZDV, 3TC, and FTC achieved greater GTexposures than BP both after a single dose and under steady-state conditions. The median GT exposures for these drugs ranged from 111 to 371% that of BP after a single dose and 235 to 411% after multiple doses. TDF GT exposure at first dose achieved a median of 135%, and 75% of BP at steady state. In contrast, ATV, stavudine (d4T), EFV, and RTVachieved lower GTexposures than BP, ranging from 0.5 to 21% that of BP after a first dose, and 0.4 to 26% after multiple doses.