Study medication

The etonogestrel contraceptive rod is an implantable device consisting of 68 mg of etonogestrel as the active ingredient with an average release rate of 60-70 μg/d in weeks 5-6, decreasing to approximately 35-45 μg/d by year 1, 30-40 μg/d by year 2, and then to 25-30 μg/d at the end of the third year. The bioavailability remains constant and close to 100%, and the elimination half-life of the parent compound is around 25 hours. Etonogestrel is mainly bound to albumin, which is not affected by changes in body’s serum estrogen levels. There is no serum accumulation; the steady decline in concentration is due mainly to the slightly declining release rate over time.

Study procedures

This observational pharmacokinetic study was conducted at the University of Chicago Medical Center in collaboration with the University of Washington from June 2008 through September 2009. Institutional review board (IRB) approval was obtained at both institutions for the duration of the study. An IRB-approved flier was developed by study staff for participant recruitment. Fliers were posted at the University of Chicago Obstetrics and Gynecology outpatient medical clinics and in various community-based locations on the south side of Chicago, IL. An IRB-approved online advertisement was also posted to Craigslist and on the University of Chicago Intranet World Wide Web site. Participants were all self-identified. Interested participants then spoke with a study staff member via telephone to discuss study protocols and complete a basic eligibility screening. All eligible women were invited to a screening and enrollment visit. Nineteen women were screened and of those, 17 remained interested and eligible. Informed consent was then obtained from all participants in accordance with standard human subjects protection guidelines. Two cohorts of women were enrolled: women with BMI ≥30 (obese women) and a smaller cohort of women with BMI <25 (normal-weight women). Normal-weight women were enrolled to compare their pharmacokinetic parameters to those of normal-weight historical controls, aiding in the validation of the etonogestrel assay developed for this study. Women aged 18-45 years were included in the study if they were in general good health, met BMI criteria, were premenopausal with a uterus and at least 1 ovary, were willing to comply with the study protocol and visit schedule, were weight stable/not interested in gaining or losing weight during the study period, and had regular menses with monthly moliminal symptoms over the year prior to participation. Women were excluded if they had any contraindication to use of the implant, were breast-feeding, had a known or suspected pregnancy, were planning a pregnancy within 12 months, or had a delivery or abortion within 4 months of device insertion. Other exclusion criteria included: abnormal genital bleeding, hypersensitivity or allergy to any components of the implant, or a history of polycystic ovarian syndrome, thromboembolic issues, liver disease, diabetes, glucose abnormality, >1 cardiovascular risk factor, or known medical contraindications. Women were excluded if they had used a cytochrome P450 3A4 inducer within 2 months, a cytochrome P450 3A4 inhibitor within 2 weeks, an investigational drug within 2 months, or injectable contraception within 6 months of the start of the trial medication.

Consistent with prior studies, we used a washout period of 14 days for women who were current combined hormonal contraceptive users. Participants were advised to use a backup nonhormonal form of contraception or abstain from intercourse throughout the washout period.

All women who met study criteria were asked to participate. After a urine pregnancy test was confirmed negative, height and weight were measured, BMI was calculated, and vital signs were obtained, participants underwent standard insertion of the implant. All insertions were performed by a single physician who had completed company-required Implanon (Organon, Whitehouse Station, NJ) insertion training. Insertion took place during days 1-5 of the menstrual cycle. If deviating from the preferred timing of device insertion, insertion took place only if the participant had abstained from intercourse since the last menstrual period.

As peak serum concentration is usually achieved between 96-144 hours after the insertion of the device in normal-weight women (range, 24–263 hours), blood samples were obtained at 50-hour intervals until 300 hours (6 blood draws over 13 days) postinsertion for plasma etonogestrel concentration determination. Two additional samples were obtained at 3 and 6 months postinsertion in the obese cohort to assess pharmacokinetics after the postdistributive phase. At all postinsertion visits, we measured blood pressure and weight, calculated BMI, asked about bleeding changes, evaluated the implant site, and described ease of device palpation.

Obese participants were asked to respond to a staff-administered survey regarding acceptability at the 300-hour, and the 3- and 6-month postinsertion visits. For participants requesting implant removal, the reason for discontinuation was documented. All participants were given the option of keeping the device for the full 3 years of use.

Laboratory procedures for developing etonogestrel assay

Serum and plasma samples were collected by venipuncture at the University of Chicago Medical Center in regular plasma tubes (green top; heparinized tubes). The vials of blood were centrifuged at 5000 g for 10-15 minutes and plasma extracted into microcentrifuge tubes. Samples were stored at our site at −80°C and then bulk-shipped in dry ice to the University of Washington School of Pharmacy pharmacokinetic laboratory, where samples remained frozen at −80°C until the assays were performed.

Total plasma concentrations of etonogestrel were determined using a modification of an extraction, derivatization, and liquid chromatography-tandem mass spectrometry protocol as described by Kalhorn et al for the measurement of testosterone. Briefly, 50 μL methanol containing testosterone-16,16,17-d ₃ (Sigma Aldridge, St. Louis, MO) as an internal standard was added to screw-capped tubes and the methanol evaporated under nitrogen gas. In all, 100 μL of plasma samples were added. Samples were extracted twice with 5 mL of 80:20 hexane:ethyl acetate and dried under nitrogen gas. To form the oxime derivatives of etonogestrel and testosterone for liquid chromatography-mass spectrometry detection, the samples were reconstituted with 100 μL of 0.1 mol/L hydroxylamine hydrochloride (in 50:50 methanol:water) and transferred to autosampler vials with conical glass inserts. Samples were heated at 60°C for 1 hour. A final volume of 20 μL was injected onto a Waters Acquity UPLC coupled with a Micromass Premiere-XE tandem quadrupole mass spectrometer operated in the ES+ mode equipped with a Waters BEH C8 50 × 2.1 mm column (Waters Corp., Milford, MA). Column flow was 0.3 mL/min with a gradient starting at 95% water and 5% acetonitrile, changing to 10% water and 90% acetonitrile at 3 minutes. The column was rinsed at 100% acetonitrile and returned to 95% water and 5% acetonitrile at 4.2 minutes. Run time was 5.5 minutes. The oximes of 3D-testosterone eluted at 2.5 minutes, and etonogestrel at 2.8 minutes. Daughter ions (m/z) 112.1 and 124.1 of the precursor ion (m/z) 307.2 were monitored for the 3D-testosterone internal standard using a cone voltage of 50 and collision energy of 30 eV. The daughter ion (m/z) 124.1 of the precursor ion (m/z) 340.2 was monitored for etonogestrel using a cone voltage of 40 and collision energy of 30 eV. The range of quantitation for this assay was 100-2000 pg/mL. Quality control specimens were run each day. Intraday coefficients of variation were 4.9%, 3.7%, and 5.8% and interday coefficients of variation were 12.5%, 11.1%, and 11.0% for 500, 1000, and 1500 pg/mL levels, respectively.

Sample size

For this descriptive study, we calculated our sample size to allow for the characterization of the pharmacokinetic curve for etonogestrel concentrations in a previously unstudied population (ie, obese women). Similar published pharmacokinetic studies were of comparable size. Due to the existence of ample historical data on normal-weight women, we did not power this study to produce a robust pharmacokinetic curve for the normal-weight cohort or make statistical comparisons between the normal-weight and obese participants.

Determination of pharmacokinetic parameters

Serial plasma etonogestrel concentrations were plotted in a concentration-time curve based on the actual plasma etonogestrel concentration from each individual (the data are expressed as mean ± SD). The following pharmacokinetics parameters of etonogestrel were obtained directly from the measured results: peak plasma concentration (C _max ) and time to achieve C _max . The area under the plasma concentration-time curve in the first 6 months was calculated using the trapezoidal rule in both groups. In the control group, the plasma etonogestrel concentrations at 3 and 6 months were modeled based on each patient’s plasma concentrations up to 300 hours by linear regression analysis using data points obtained after C _max under the assumption that etonogestrel elimination follows first-order kinetics. Based on data from the manufacturer as well as a previously published pharmacokinetic study, the elimination phase of etonogestrel in the implant fits characteristics of first-order kinetics. Additionally, the systemic clearance of the etonogestrel in the first 2 years after implantation remains around 7.5 L/h, further suggesting first-order kinetics. Therefore, linear regression analysis of the log concentrations in the descending phase was used to estimate a rate elimination constant and extrapolate concentrations >6 months after implantation.

Statistical analysis

Descriptive statistics for baseline demographic and survey data were used to characterize the study population with frequencies and medians provided as appropriate. Statistical analyses were performed using STATA 9 (StataCorp LP, College Station, TX) and SigmaPlot (Systat Software, Chicago, IL). Concentration-time curve was created by using SigmaPlot (v.11.1).