If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Extended use of levonorgestrel-releasing intrauterine system (LNG-IUS) 52 mg: A population pharmacokinetic approach to estimate in vivo levonorgestrel release rates and systemic exposure including comparison with two other LNG-IUSs
To characterize performance of levonorgestrel-releasing intrauterine system (LNG-IUS) 52mg (Mirena) over 8 years of use and facilitate comparisons with LNG-IUS 19.5mg and LNG-IUS 13.5mg.
Study design
We estimated in vivo levonorgestrel (LNG) release rates and LNG plasma/serum concentrations for LNG-IUS 52mg up to 8 years of use with a population pharmacokinetic (popPK) approach using data from the Mirena Extension Trial (MET) and earlier clinical trials. We compared these with previously published release rates and exposure data for LNG-IUS 19.5mg and 13.5mg. Our 8-year popPK and release models were developed based on measured plasma/serum LNG and sex hormone-binding globulin concentrations and residual LNG content from removed LNG-IUS 52mg devices.
Results
Model-based estimated LNG release rates for LNG-IUS 52mg decreased from ∼21 µg/d after insertion to ∼7.0 µg/d after 8 years, similar to LNG-IUS 19.5mg after 5 years (7.6 µg/d) and higher than LNG-IUS 13.5mg after 3 years (5.5 µg/d). Model-based estimated and measured plasma/serum LNG concentrations showed satisfactory agreement. Average model-based estimated LNG concentrations after 8 years of LNG-IUS 52mg use (100 ng/L [coefficient of variance 39.9%]) were similar to LNG-IUS 19.5mg after 5 years (84.8 ng/L [39.9%]) and higher than LNG-IUS 13.5mg after 3 years (58.1 ng/L [40.8%]).
Conclusions
The 8-year release and popPK models provide reliable in vivo LNG release rates and concentration estimates, respectively, facilitating direct comparisons between the 3 studied LNG-IUSs. LNG release rates from LNG-IUS 52mg at 8 years are similar to LNG-IUS 19.5mg at 5 years and higher than LNG-IUS 13.5mg at 3 years.
Levonorgestrel release from intrauterine system reservoirs declines with duration of use in a predictable way, and in relation to the initial load. As release rates and plasma concentrations of levonorgestrel may influence endometrial and systemic side effects, these data may assist clinical decision-making.
1. Introduction
Levonorgestrel-releasing intrauterine systems (LNG-IUSs) are highly effective contraceptive methods, with over 30 years of use and a significant global impact on women's health [
]. Levonorgestrel (LNG), a progestin of the 19-nortestosterone class, is the active ingredient in LNG-IUS and is released directly into the uterine cavity resulting in progestogenic effects including thickened cervical mucus and suppressed endometrial maturation [
A number of LNG-IUS options with different reservoir contents and release rates are available, including 3 produced by Bayer AG: LNG-IUS 13.5 mg (Skyla/Jaydess; initial release rate [IRR] 8 μg/d) [
A randomized, phase II study describing the efficacy, bleeding profile, and safety of two low-dose levonorgestrel-releasing intrauterine contraceptive systems and Mirena.
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
]. A standardized comparison of LNG release and exposure over their respective approved usage periods may support informed, patient-centered decision-making regarding product choice [
American college of obstetricians and gynecologists’ committee on health care for underserved women, contraceptive equity expert work group, and committee on ethics, patient-centered contraceptive counseling. committee statement no.1.
The population pharmacokinetic (popPK) approach provides a uniform method of comparison among different pharmacologic studies and thus yields a comprehensive understanding of pharmacokinetics (PK), including drug exposure and relevant covariates, across different products [
]. We previously developed comprehensive popPK and release models estimating LNG and sex hormone-binding globulin (SHBG) systemic exposure and in vivo LNG release for the 3 Bayer-produced intrauterine systems (IUSs) (including LNG-IUS 52 mg up to 5 years), subdermal implants, and oral contraceptives [
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
The Mirena Extension Trial (MET) was designed to evaluate the safety, efficacy, and PK of extended Mirena use from beyond 5 years up to the end of 8 years. In the present analysis, we aimed to estimate in vivo LNG release and systemic exposure of LNG (total and unbound) and SHBG for LNG-IUS 52 mg up to the end of 8 years of use, and to facilitate comparisons with LNG-IUS 19.5 mg and LNG-IUS 13.5 mg over their respective approved periods of use, using a popPK approach with data from the MET and earlier clinical trials of all 3 IUSs.
2. Materials and methods
2.1 Study design
The MET (NCT02985541) was a multicenter, single-arm study enrolling premenopausal women aged 18–35 years who had used LNG-IUS 52 mg for the last 4.5–5 years and were willing to continue use of the device for up to 8 years. The MET was conducted in 54 US centers between December 22, 2016, and May 28, 2021. The primary study outcome was contraceptive efficacy as measured by the Pearl Index and failure rate as measured by the Kaplan–Meier method. Local ethics committees/institutional review boards approved all studies included in this paper, and all participants provided written informed consent.
2.2 MET data collection
We collected blood samples to determine LNG and SHBG [
] concentrations from all women at baseline (∼5 years after insertion), study end (∼8 years after insertion) or at premature study discontinuation (before the LNG-IUS 52 mg device was removed), and randomly at 2 of 4 visits within the study period. We collected removed devices to determine residual LNG content. We also weighed study participants annually and at screening to monitor for adverse effects and to accommodate for the known effect of body weight on the PK of LNG in obese patients, including in those using LNG-IUS [
We determined LNG concentrations in plasma using validated liquid chromatography-tandem mass spectrometry methods and SHBG concentrations in serum by a target-dependent dissociation-enhanced lanthanide fluorescence immunoassay method. We analyzed residual LNG content in used LNG-IUS 52 mg devices (removed at study end or prematurely when women discontinued the study) [
]. Residual LNG was extracted from the elastomer material of removed IUS devices using dichloromethane and quantified by liquid chromatography on a reversed-phase column using external calibration [
] (see Supplementary Methods for assay details and method validations).
2.4 PopPK approach
We built upon our previously developed comprehensive popPK and release models to estimate LNG (total and unbound) and SHBG exposure as well as in vivo LNG release rates. The previously developed models were based on measured LNG (plasma/serum) and SHBG (serum) concentrations and residual LNG content from removed LNG-IUS 52 mg, 19.5 mg, and 13.5 mg devices from multiple clinical studies [
A randomized, phase II study describing the efficacy, bleeding profile, and safety of two low-dose levonorgestrel-releasing intrauterine contraceptive systems and Mirena.
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
Pharmacokinetics of two low-dose levonorgestrel-releasing intrauterine systems and effects on ovulation rate and cervical function: pooled analyses of phase II and III studies.
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
]. The present popPK model and associated code are described in the Supplementary Materials. We refer to “serum” and “plasma” interchangeably throughout the remainder of the present paper [
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
We applied and refined the 5-year comprehensive popPK and release models in a stepwise manner up to 8 years of LNG-IUS 52 mg use, using data from the MET and earlier trials, in order to estimate in vivo LNG (total and unbound) and SHBG concentrations (8-year popPK model) and typical in vivo LNG release rates (8-year release model). We then compared these findings with previously published data for LNG-IUS 19.5 mg (over 5 years) and LNG-IUS 13.5 mg (over 3 years).
In the 8-year popPK model, we described the release of LNG from the device by a combination of a zero-order, a first-order, and a time-dependent first-order release. We used a 2-compartment model to describe the biphasic LNG disposition considering LNG binding to albumin (assumed to be constant) and to SHBG. We described the change in endogenous SHBG concentrations in serum over time using a turnover model assuming a constant, zero-order endogenous SHBG synthesis rate and a first-order elimination rate.
SHBG synthesis was assumed to be inhibited by LNG where the effect of the LNG concentration on the SHBG synthesis rate was linear and delayed. Moreover, it was assumed that only unbound LNG could be eliminated from serum and distributed to peripheral tissues. Thus, SHBG was needed to estimate unbound LNG concentrations, which were expected to reflect effective exposure. We included interindividual variability and body weight on the apparent clearance of LNG and on the SHBG baseline concentration. We also calculated LNG concentrations for 2 subsets of women weighing ≤55 kg and >55 kg.
The 8-year release model contained only the release processes of the popPK model, i.e., a combination of a zero-order, a first-order, and a time-dependent first-order release, without the link between LNG and SHBG concentrations.
The popPK analysis was performed by means of nonlinear mixed-effects modeling using NONMEM software (version 7.5.0; Icon Development Solutions, Ellicott City, MD, USA) in combination with PsN (version 5.0.0) and Gfortran (version 9.3.0) as a compiler. Diagnostic graphics, exploratory analyses, and postprocessing of NONMEM output were performed using R (version 4.0.3; The R Foundation for Statistical Computing, Vienna, Austria) and Rstudio (version 1.3.1093, Rstudio Inc., Boston, MA, USA). Further details of the final 8-year popPK and release models are given in the Supplementary Methods.
3. Results
Below, we present measured device residual LNG content, plasma LNG and SHBG concentrations, model-based estimated LNG release rates, and model-based estimated LNG and SHBG concentrations.
3.1 Participants
Of 501 healthy premenopausal women initially enrolled in the MET, 362 started year 6 and 223/362 (64.1%) completed year 8. Efficacy and safety findings are published elsewhere [
]. Summary statistics of the MET and other included study participants are given in Supplementary Table 1 and Supplementary Results.
3.2 MET measured results
We collected LNG-IUS 52 mg devices from 300 women who had their device removed between 5 and 8 years after insertion. The median residual LNG content declined from ∼25 mg at the start of year 6 to ∼13.5 mg at the end of year 8 (a ∼20% annual decline, or a total of 46% during the 3-year extended use period) (Fig. 1A).
Fig. 1Visual Predictive Check of (A) model-based predicted median residual LNG content compared with measured residual content, and of the model-based predicted concentrations compared with measured geometric mean concentrations of (B) LNG and (C) SHBG over 8 years of LNG-IUS 52 mg use (in the MET and two earlier clinical trials).
For LNG-IUS 52 mg users, measured plasma LNG geometric mean (coefficient of variance [CV%]) concentrations reduced from 127 ng/L (41.5%) at the start of year 6 to 83.4 ng/L (49.6%) at the end of year 8 (Supplementary Table 2; Fig. 1B). The geometric mean (CV%) of measured SHBG remained relatively stable from 42.3 nmol/L (50.2%) to 38.6 nmol/L (55.5%) over the 3 study years (Supplementary Table 2; Fig. 1C).
A subgroup analysis of women with body weight ≤55 kg (n=24) showed that measured LNG concentrations during years 5–8 were 27% to 43% higher than concentrations for women weighing >55 kg and 25% to 39% higher when compared with all women in the study (Supplementary Fig. 1).
3.3 Model-based estimated LNG release rates
The 8-year LNG-IUS 52 mg release model adequately described the ex vivo residual content measurements, as shown in Fig. 1B. Model-based estimated typical in vivo LNG release rates for LNG-IUS 52 mg decreased from 21.6 µg/d at 15 days after insertion to 10.7 µg/d after 5 years and 7.04 µg/d after 8 years. The estimated average release rate for LNG-IUS 52 mg was 20 µg/d over the first year, 15 µg/d over the first 5 years, and 13 µg/d over the 8 years of LNG-IUS 52 mg use.
Comparing LNG-IUS 52 mg with lower-dose LNG-IUSs, at the end of year 8, the LNG release rate (7.04 µg/d) was similar to LNG-IUS 19.5 mg after 5 years (∼7.6 µg/d) and ∼28% higher than that of LNG-IUS 13.5 mg after 3 years (∼5.5 µg/d) (Fig. 2, Table 1).
Fig. 2Model-based estimated typical in vivo LNG release rates for LNG-IUS 52 mg compared with LNG-IUS 19.5 mg and LNG-IUS 13.5 mg at representative time points from 15 days to 8 years after insertion.
Table 1Overview of model-based estimated typical in vivo release rates, LNG (total and unbound) concentrations, and SHBG concentrations for LNG-IUS 52 mg compared with LNG-IUS 19.5 mg and LNG-IUS 13.5 mg
Model-based estimated typical in vivo release rates (µg/d)
Model-based estimated total and unbound (in italics) LNG concentrations (ng/L)
Typical estimated in vivo release rates estimated based on 8-year release model, and total and unbound (in italics) LNG and SHBG concentrations in plasma/serum estimated based on the 8-year popPK model for LNG-IUS 52 mg. 8-year release model: All women with a valid residual content measurement were included. Calculations of the 90% CI are based on 1000 simulations, considering the covariance matrix of the structural parameter estimates. 8-year popPK model: Only women with at least one valid LNG concentration (above lower limit of quantitation) were included.
Typical estimated in vivo release rates estimated based on 8-year release model, and total and unbound (in italics) LNG and SHBG concentrations in plasma/serum estimated based on the 8-year popPK model for LNG-IUS 52 mg. 8-year release model: All women with a valid residual content measurement were included. Calculations of the 90% CI are based on 1000 simulations, considering the covariance matrix of the structural parameter estimates. 8-year popPK model: Only women with at least one valid LNG concentration (above lower limit of quantitation) were included.
Typical estimated in vivo release rates estimated based on 8-year release model, and total and unbound (in italics) LNG and SHBG concentrations in plasma/serum estimated based on the 8-year popPK model for LNG-IUS 52 mg. 8-year release model: All women with a valid residual content measurement were included. Calculations of the 90% CI are based on 1000 simulations, considering the covariance matrix of the structural parameter estimates. 8-year popPK model: Only women with at least one valid LNG concentration (above lower limit of quantitation) were included.
Baseline concentration of SHBG for LNG-IUS 52 mg was 52 nmol/L; for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg, baseline concentrations of SHBG were both 51.5 nmol/L.
Baseline concentration of SHBG for LNG-IUS 52 mg was 52 nmol/L; for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg, baseline concentrations of SHBG were both 51.5 nmol/L.
Baseline concentration of SHBG for LNG-IUS 52 mg was 52 nmol/L; for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg, baseline concentrations of SHBG were both 51.5 nmol/L.
Due to the open ends of the hormone elastomer core for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg, the initial LNG release (in vitro and in vivo) is faster and more variable and stabilizes only from day 24 onward. Therefore, initial release for LNG-IUS 13.5 mg and LNG-IUS 19.5 mg were only estimated from day 24 onward.
Estimation beyond start of data collection for women participating in the Mirena Extension Trial.
12.3
7.78
Beyond approved period of use
361
131 [38.0]
673
87.0 [40.9]
–
42.6 [44.0]
48.3 [42.4]
Beyond approved period of use
[12.3–12.3]
[7.69–7.86]
2.04 [24.1]
1.29 [24.3]
5 y
10.7
7.59
361
123 [38.2]
230
84.8 [39.9]
–
42.9 [44.3]
49.0 [41.7]
[10.7–10.7]
[7.48–7.71]
1.91 [24.1]
1.26 [24.0]
6 y
9.31
Beyond approved period of use
330
114 [38.7]
–
–
42.7 [44.4]
Beyond approved period of use
[9.30–9.31]
1.78 [24.5]
7 y
8.1
275
106 [39.2]
–
–
42.5 [45.4]
[8.09–8.10]
1.65 [24.7]
8 y
7.04
224
100 [39.9]
–
–
43.2 [46.2]
[7.04–7.05]
1.56 [24.8]
Average over first y
20.1
12.1
8.04
[20.0–20.2]
[12.0–12.1]
[8.00–8.08]
Average over 3 y
17.6
9.83
6.44
[17.5–17.6]
[9.78–9.86]
[6.42–6.46]
Average over 5 y
15.5
9.02
Beyond approved period of use
[15.4–15.5]
[8.98–9.04]
Average over 6 y
14.6
Beyond approved period of use
[14.5–14.6]
Average over 7 y
13.7
[13.7–13.8]
Average over 8 y
13
[12.9–13.0]
1 year = 365 days.
CI, confidence interval; CV%, coefficient of variance; Geo, geometric; LNG, levonorgestrel; LNG-IUS, levonorgestrel-releasing intrauterine system; N/A, not available; popPK, population pharmacokinetic; SHBG, sex hormone-binding globulin.
a Typical estimated in vivo release rates estimated based on 8-year release model, and total and unbound (in italics) LNG and SHBG concentrations in plasma/serum estimated based on the 8-year popPK model for LNG-IUS 52 mg. 8-year release model: All women with a valid residual content measurement were included. Calculations of the 90% CI are based on 1000 simulations, considering the covariance matrix of the structural parameter estimates. 8-year popPK model: Only women with at least one valid LNG concentration (above lower limit of quantitation) were included.
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
c Baseline concentration of SHBG for LNG-IUS 52 mg was 52 nmol/L; for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg, baseline concentrations of SHBG were both 51.5 nmol/L.
d Concentrations were estimated for women who had the intrauterine system still in place.
e The CI is narrow because only one parameter of the release model was re-estimated with low uncertainty.
f Estimation beyond start of data collection for women participating in the Mirena Extension Trial.
g Due to the open ends of the hormone elastomer core for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg, the initial LNG release (in vitro and in vivo) is faster and more variable and stabilizes only from day 24 onward. Therefore, initial release for LNG-IUS 13.5 mg and LNG-IUS 19.5 mg were only estimated from day 24 onward.
3.4 Model-based estimated LNG and SHBG concentrations
The 8-year LNG-IUS 52 mg popPK model was qualified using Visual Predictive Checks comparing predicted and measured LNG and SHBG concentrations (Fig. 1B and 1C).
Using data from the MET and previous LNG-IUS 52 mg studies, the model-based estimated total LNG geometric mean (CV%) concentration was 181 ng/L (38.3%) at 15 days after device insertion, falling to 123 ng/L (38.2%) at 5 years and to 100 ng/L (39.9%) at 8 years (Table 1; Fig. 3A). The estimated unbound LNG geometric mean also showed a gradual decline from 2.82 ng/L (CV%: 24.2%) at 15 days to 1.91 ng/L (24.1%) at 5 years and 1.56 ng/L (24.8%) at 8 years (Table 1; Fig. 3B). The previously estimated fraction of unbound LNG was about 1.6% of the total LNG concentration, which was confirmed by the popPK analysis of MET data [
Fig. 3Model-based estimated LNG-IUS 52 mg plasma/serum concentration time curves for (A) total LNG, (B) unbound LNG, and (C) SHBG, compared with LNG-IUS 19.5 mg and LNG-IUS 13.5 mg at representative time points from 15 days to 8 years after insertion.
The model adequately described the LNG concentrations, with an average deviation in geometric mean concentrations of 5% for years 5–7 and 17% for year 8 (Supplementary Table 2). Similarly, we found satisfactory agreement between measured and estimated SHBG concentrations. The concentration ranges of measured and model-based estimated LNG and SHBG concentrations overlapped widely at all assessed time points (Supplementary Table 2).
When comparing all 3 LNG-IUSs using data from the MET and previous studies, both the estimated total and unbound LNG concentration curves showed a steady decrease, and both showed high variability (Fig. 3A and 3B, respectively). After 8 years of use, estimated LNG plasma concentrations for LNG-IUS 52 mg (100 ng/L [CV%: 39.9%]) were similar to LNG-IUS 19.5 mg (84.8 ng/L [39.9%]) after 5 years, and higher than for LNG-IUS 13.5 mg (58.1 ng/L [40.8%]) at 3 years.
Estimated serum SHBG concentrations declined by ∼20% from a typical baseline concentration of 52.0 nmol/L after LNG-IUS 52 mg insertion to 41.2 nmol/L at 2 months, remaining relatively stable thereafter, with a slight increase to 43.2 nmol/L by the end of 8 years (Table 1; Fig. 3C). Estimated SHBG levels for LNG-IUS 52 mg were slightly lower than for both lower-dose LNG-IUSs. The estimated SHBG baseline concentrations for LNG-IUS 19.5 mg and LNG-IUS 13.5 mg were 51.5 nmol/L, and the initial decrease during the first 2 months after insertion was 13% and 10%, respectively.
The previously identified effect of body weight on the apparent clearance of LNG and on the SHBG baseline concentration was confirmed by the popPK analysis.
4. Discussion
This analysis describes the performance of LNG-IUS 52 mg over an 8-year duration of use and allows comparisons with lower-dose shorter-duration LNG-IUSs. Our popPK approach delivers a reliable estimation of in vivo LNG release and exposure from LNG-IUS 52 mg over the complete 8-year period of use. Our LNG-IUS 52 mg release rates in years 0–5 were similar to previously reported release rates using a different methodology [
]. After 8 years, both the model-based estimated release rate and estimated LNG concentration for LNG-IUS 52 mg were similar to LNG-IUS 19.5 mg after 5 years and higher than LNG-IUS 13.5 mg after 3 years. In comparison with other LNG-containing contraceptives, systemic exposures for all 3 examined LNG-IUSs at the end of their approved use periods remain considerably lower than average steady-state exposures for a combined oral contraceptive (1676 ng/L), progestin-only pill (327 ng/L), or implant (259 ng/L at 5 years) [
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
Women value control of their contraceptive choices, informed by comprehensive information about side effects, treatment options, and patient-centered counseling [
American college of obstetricians and gynecologists’ committee on health care for underserved women, contraceptive equity expert work group, and committee on ethics, patient-centered contraceptive counseling. committee statement no.1.
]. These factors are particularly important in enhancing choice and contraceptive uptake, including in marginalized groups with low contraceptive use [
]. One factor important to women is bleeding profile (including amenorrhea), which plays a role in counseling, patient choice and satisfaction, and continuation [
]. In a separate 3-year comparison of all 3 LNG-IUS devices, bleeding profiles and amenorrhea rates were similar, although total bleeding and spotting days were higher with decreasing LNG dose [
]. Notably, rates of dysmenorrhea, vaginal hemorrhage, and total bleeding and spotting days were lower for LNG-IUS 52 mg over years 6 to 8 compared to the lower dose LNG-IUSs at the end of their respective approved periods of use [
A randomized, phase II study describing the efficacy, bleeding profile, and safety of two low-dose levonorgestrel-releasing intrauterine contraceptive systems and Mirena.
]. While use of all 3 devices improves bleeding patterns overall, amenorrhea rates in other studies were higher in LNG-IUS 52 mg (18% at the start of year 6 increasing to 34% at the final reference period in year 8 [
A randomized, phase II study describing the efficacy, bleeding profile, and safety of two low-dose levonorgestrel-releasing intrauterine contraceptive systems and Mirena.
]). While LNG-IUS 52 mg is approved for use as a contraceptive in women with or without heavy menstrual bleeding, many women and clinicians consider using LNG-IUS for its non-contraceptive benefits, including the management of menstrual symptoms in young women and for endometrial protection during hormone therapy [
Our findings of higher LNG release and exposure estimates in LNG-IUS 52 mg compared with those of the lower dose LNG-IUS 13.5 mg, taken together with established safety data, may inform clinicians and patients in discussions about the risks and benefits of the different LNG-IUSs available with different approved durations and LNG exposures. The recent confirmation of ongoing high contraceptive efficacy and safety [
] of LNG-IUS 52 mg use up to 8 years, as well as US approval of its extended use, expands treatment choices for patients. Moreover, as levonorgestrel release rates and plasma concentrations may influence local and systemic effects, these data may assist in formulating future research hypotheses.
The finding that the release rate of LNG-IUS 52 mg at 8 years is higher than that of LNG-IUS 13.5 mg at 3 years raises the question of possible extrapolation beyond the data. For example, could the model predict at what duration of use the release rate for LNG-IUS 52 mg drops below the efficacious release rate of LNG-IUS 13.5 mg at 3 years? While an interesting hypothesis, our model-based approach should not be used to predict beyond current data sets for several reasons. At some point the LNG-IUS reservoirs will approach depletion, which may change the shape of the LNG release (e.g., leading to a sudden drop in LNG release), which cannot be reliably predicted by the model without underlying data. As no PK, residual content, safety, or efficacy data were collected beyond the study duration, we do not recommend LNG-IUS 52 mg use beyond the currently approved 8 years. Future research on extended use, including the collection of PK, residual content, efficacy, and safety data, may help answer this question.
While there was, overall, satisfactory agreement between measured and model-based estimated concentrations, the difference toward the end of year 8 may be attributed to the model's slight overestimation of LNG concentrations at that time and a different selection of women with differing average body weight at that time point (N=198 measured; N=224 model-based estimated).
Model parameter estimates for all fixed effects were similar to the ones estimated for the 5-year LNG-IUS 52 mg comprehensive popPK and release models [
], with overlapping confidence intervals. As expected, the random effects were estimated to be slightly lower in the 8-year popPK analysis than in the previous models, since the previous models included not only LNG-IUS 52 mg but also LNG-IUS 19.5 mg and LNG-IUS 13.5 mg.
Finally, the 8-year popPK analysis using MET data confirmed the inverse relationship between body weight and LNG concentrations (due to the positive correlation with LNG clearance and the negative correlation with SHBG baseline concentration), in line with existing evidence [
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
Influence of changes in the concentration of sex hormone-binding globulin in human serum on the protein binding of the contraceptive steroids levonorgestrel, 3-keto-desogestrel and gestodene.
]; accordingly, changes in the SHBG concentration result in an increase (at higher SHBG concentrations) or in a decrease (at lower SHBG concentrations) of the total LNG concentration.
Strengths of the popPK approach include: the robustness of the popPK methodology, confirmed by the wide overlap of concentration ranges of measured and estimated LNG and SHBG concentrations for all assessed time points; the stepwise demonstration of model suitability from years 6 to 8 of use; the ability to estimate in vivo release rates; the ability to estimate both population and individual PK profiles when only sparse sampling is available; and the model-based estimation of unbound LNG, which took the interaction between SHBG and LNG into account.
There are several study limitations. Firstly, our findings may represent group effects and may not be applicable to all LNG-IUS users or indications for use. Secondly, these data represent systemic LNG exposure, and do not provide an understanding of local endometrial and cervical LNG concentrations. Thirdly, without supporting data, the popPK model should not be used for addressing theoretical questions such as estimating release rates beyond the available data sets.
To conclude, the 8-year popPK and release models deliver reliable LNG exposure and release rate estimates for up to 8 years of LNG-IUS 52 mg use based on a broad data set. Our PK results support the clinical findings of efficacy and safety of LNG-IUS 52 mg for 6 to 8 years of use [
], and furthermore confirm constant LNG release over this extended period in the whole population. In vivo LNG release rates during extended use of LNG-IUS 52 mg from 6 to 8 years are at least on the same level as for the lower dose LNG-IUS devices at the end of their approved periods of use. Systemic LNG exposure of LNG-IUS 52 mg during years 6 to 8 is similar to other LNG-IUS devices and considerably lower than implanted and oral LNG-containing contraceptives.
Data availability statement
Availability of the data underlying this publication will be determined later according to Bayer's commitment to the EFPIA/PhRMA “Principles for responsible clinical trial data sharing”. This pertains to scope, time point, and process of data access.
As such, Bayer commits to sharing upon request from qualified scientific and medical researchers, patient-level clinical trial data, study-level clinical trial data, and protocols from clinical trials in patients for medicines and indications approved in the United States (US) and European Union (EU) as necessary for conducting legitimate research. This applies to data on new medicines and indications that have been approved by the EU and US regulatory agencies on or after January 01, 2014.
Interested researchers can use www.clinicalstudydatarequest.com to request access to anonymized patient-level data and supporting documents from clinical studies to conduct further research that can help advance medical science or improve patient care. Information on the Bayer criteria for listing studies and other relevant information is provided in the Study sponsors section of the portal.
Data access to anonymized patient-level data, protocols, and clinical study reports will be granted after approval by an independent scientific review panel. Bayer is not involved in the decisions made by the independent review panel. Bayer will take all necessary measures to ensure that patient privacy is safeguarded.
Acknowledgments
The authors would like to thank all of the participants and research staff for their kind contributions. Highfield, Oxford, UK (funded by Bayer AG) provided editorial assistance in the preparation of this manuscript. The authors would like to acknowledge Stefan Zeiser (LAP&P Consultants BV, Leiden, Netherlands) for his contributions to the 8-year population pharmacokinetic analysis.
A randomized, phase II study describing the efficacy, bleeding profile, and safety of two low-dose levonorgestrel-releasing intrauterine contraceptive systems and Mirena.
Comparative pharmacokinetic analysis of levonorgestrel-releasing intrauterine systems and levonorgestrel-containing contraceptives with oral or subdermal administration route.
American college of obstetricians and gynecologists’ committee on health care for underserved women, contraceptive equity expert work group, and committee on ethics, patient-centered contraceptive counseling. committee statement no.1.
Pharmacokinetics of two low-dose levonorgestrel-releasing intrauterine systems and effects on ovulation rate and cervical function: pooled analyses of phase II and III studies.
Influence of changes in the concentration of sex hormone-binding globulin in human serum on the protein binding of the contraceptive steroids levonorgestrel, 3-keto-desogestrel and gestodene.
Conflict of interest: Dr. Jensen has received payments for consulting from Bayer Healthcare, Evofem Biosciences, Mayne Pharma, Merck, Sebela Pharmaceuticals, and TherapeuticsMD. Oregon Health & Science University (OHSU) has received research support from AbbVie, Bayer Healthcare, Daré Bioscience, Estetra SPRL, Medicines360, Merck, and Sebela Pharmaceuticals. These companies and organizations may have a commercial or financial interest in the results of this research and technology. These potential conflicts of interest have been reviewed and managed by OHSU. Dr. Reinecke is an employee of Bayer AB. Dr. Post is an employee of LAP&P Consultants BV, an organization receiving payments for modeling and simulation services and consulting for the pharmaceutical industry, and the 8-year population pharmacokinetic analysis was funded by Bayer AG. Dr. Lukkari-Lax is an employee of Bayer Oy. Dr. Hofmann is an employee of Bayer AG.
Funding: The study was funded by Bayer AG, Berlin, Germany. Bayer AG is responsible for the study design, data collection and analysis, writing, decision to submit the article for publication, and overall conduct of the study.
Clinical trial registration: ClinicalTrials.gov, number NCT02985541.
00007-0&id=gr1.jpg){kind=link}
00007-0&id=gr2.jpg){kind=link}
00007-0&id=gr3.jpg){kind=link}