Extended duration of use of long-acting reversible contraceptives

Abstract

Objective

To document the efficacy and safety of etonogestrel subdermal contraceptive implants, levonorgestrel (LNG) intrauterine systems (IUS), and copper (Cu) intrauterine devices (IUDs) for birth control when extended beyond prescription limits.

Data sources

A systematic search of MEDLINE, Embase, and Global Health electronic databases.

Study selection

The search identified articles published between January 2000 and June 2023 on long-acting reversible contraceptive (LARC) methods and extended duration of use. Publications on LNG subdermal implants or the “frameless” IUD (not marketed in Canada), case reports, case series, animal studies, and publications not written in English or French were excluded.

Synthesis

Twenty-five publications were included. No pregnancies were reported during prolonged use of the etonogestrel-releasing implant for 1 or 2 additional years beyond the recommended limits (783 participants and 306 participants, respectively), nor with the use of the TCu380A IUD for years 8 to 11 (682 participants), 8 to 12 (356 participants), and 11 to 16 (228 participants). No pregnancies were reported with the LNG-IUS in 9 studies with follow-up from 6 to 15 years, while 2 pregnancies were reported during year 6 in 2 studies, and 4 pregnancies during year 7 in 3 studies. The risk of expulsion persisted during prolonged use of the TCu380A device and the 52-mg LNG-IUS. Other adverse events were infrequent.

Conclusion

Prolonged use of LARCs is safe and much more effective than short-acting contraceptive methods. Prolonged use of LARCs might be beneficial for individuals requiring contraception especially those with disrupted access to social and health care services or living in more remote areas.

Long-acting reversible contraceptive (LARC) methods, including nonhormonal copper (Cu) intrauterine devices (IUDs), levonorgestrel (LNG)-releasing intrauterine systems (IUS), and etonogestrel (ENG)-releasing subdermal contraceptive implants, are widely used around the world.¹ There are no recent Canadian data on LARC use, but a significant increase in IUD use (4.3% to 8.8%) was observed between 2006 and 2016 in sexually active Canadian women aged 15 to 50 years who reported using contraception (P<.001).² Available LARC methods are approved for limited durations of use: the ENG implant is approved for 3 years³; most Cu-IUDs, for 5 to 10 years^4,5; the 19.5-mg LNG-IUS for 5 years⁶; and the 52-mg LNG-IUS (in Canada, as of February 2024) for up to 8 years.⁷

Since it might be difficult for patients to access services and health care professionals,⁸ especially during periods of social disruption such as the COVID-19 pandemic,⁹ and because trauma might arise during insertion or removal of these devices (vascular or neurologic trauma for subdermal implants,^10,11 uterine perforation for IUDs,^12,13 etc), it might be desirable to use these devices for longer durations. As mentioned by Ali et al, extending the use of LARCs could reduce the number of insertion and withdrawal procedures, open up new consultation slots to other contraceptive seekers, and save millions of health care dollars.¹⁴

The objective of this systematic review was to document the efficacy and safety of ENG contraceptive implants, LNG-IUS, and Cu-IUDs for birth control when their duration of use is extended beyond the limits approved by Health Canada.

METHODS

Two library science experts searched MEDLINE, Embase, and the Global Health electronic database via the Ovid interface for studies published between January 2000 and June 2023, in either French or English, that identified the concepts of LARC and extended duration of use. The search strategy used terms describing the concepts of ENG implants, Cu-IUDs, LNG-IUS, and extended duration of use (Appendix 1, available from CFPlus^*). The reference lists of articles selected for extraction and quality assessment were searched for other relevant studies.

Study selection

Eligible study designs were systematic or narrative reviews, randomized controlled trials, cohort studies, case-control studies, descriptive studies, expert opinions, and guidelines. Publications on LNG implants or the “frameless” IUD (not marketed in Canada), case reports, case series, and animal studies were not retained. After excluding duplicates, all titles and abstracts were reviewed by 2 independent reviewers (H.A. and M.B.), and those meeting inclusion criteria were selected and read in their entirety. Any disagreements between the 2 reviewers were resolved by discussion, rechecking of the publication, and consultation with a third party (E.G.), if necessary. The research was documented in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocols.¹⁵ For each study, its objectives, design, location, population, inclusion and exclusion criteria, recruitment period, ethical aspects and consent, methodology, results, strengths and limitations, discussion points, and authors’ conclusions were documented by 2 data extractors (H.A. and M.B.). The following data were extracted for guidelines: generalities, efficacy, indications and targeted population, contraindications, noncontraceptive benefits, drug interactions, adverse effects, use according to weight or body mass index (BMI), initial assessment, interventions in case of difficulty, replacement, duration of use, and authors’ recommendations.

The quality of selected studies was assessed by 2 independent reviewers (H.A. and M.B.) using the JBI Critical Appraisal Tools,¹⁶ the CASP (Critical Appraisal Skills Programme) Systematic Review Checklist,¹⁷ and the SANRA (Scale for the Assessment of Narrative Review Articles) tool.¹⁸ Guidelines were assessed using the AGREE II (Appraisal of Guidelines for Research and Evaluation II) grid.¹⁹

In accordance with article 2.2 of the research ethics board of Laval University, Que, research was exempted from evaluation because it was based exclusively on information that was accessible to the public through a mechanism established by regulation and in the public domain, and because the persons concerned had no reasonable expectation of privacy.²⁰

SYNTHESIS

Results

The search yielded 1488 titles (Figure 1).¹⁵ Following exclusions, 22 publications (2 guidelines and 20 scientific articles) were retained. An additional 5 scientific articles were added from 1 author’s (E.G.) bibliographic database. All 27 publications were extracted and assessed for quality. Two articles were rejected because of low quality score: Rocca et al²¹ (score 17% with CASP Systematic Review Checklist) and Honorato et al²² (score 25% with the JBI checklist for analytical cross-sectional studies). Quality scores were greater than 85% for most of the selected articles and reviews.

Figure 1.

Flowchart for literature search and study selection based on PRISMA guidelines for systematic reviews

Efficacy and safety of the ENG contraceptive implant beyond the approved duration of use (Table 1).^14,23-32 No pregnancy occurred in 783 participants who extended implant use for 1 additional year^14,25,28 or in 306 participants who extended use for 2 additional years.^14,25 Two prospective studies^24,25 and 1 cross-sectional study²³ demonstrated median serum ENG levels above the theoretical minimum threshold for ovulation inhibition (90 pg/mL)²⁷ after 3, 4,^23-25 and 5 years²⁵ of implant use. A weak relationship between serum ENG and time since insertion was demonstrated in 1 study,²³ while no association with BMI was found in 3 studies.^23-25 The most frequent adverse events in years 4 and 5 were irregular bleeding, reported by 46.2% of users; headache, reported by 31.8% of users; and lower abdominal pain, reported by 17.9% of users.¹⁴

Table 1.

Summary of publications on prolonged duration of etonogestrel contraceptive implants: A) Scientific articles and B) guidelines.

A) SCIENTIFIC ARTICLES
AUTHORS, YEAR OF PUBLICATION, QUALITY SCORE	DESIGN	COUNTRY	OBJECTIVES	INCLUDED STUDIES OR DESCRIPTION OF SAMPLE	CHARACTERISTICS OF PARTICIPANTS	RESULTS
Morrell et al, 201423 Quality score=50%	Cross-sectional	USA	To determine whether BMI decreases serum ENG levels in ENG implant users, particularly when implant use reaches 3 y and beyond	Participants using the ENG implant for more than 1 y: BMI <25 kg/m2: n=10 BMI ≥25 and <30 kg/m2: n=19 BMI ≥30 kg/m2: n=23	Mean age: 28 y (SD 6 y) Median BMI in obese participants: 35 kg/m2 (range 31-56 kg/m2)	32% used the ENG implant for 3 to 4.4 y Median serum ENG levels comparable between BMI categories (P=.2): -BMI <25 kg/m2: 216 pg/mL (interquartile interval=214) -BMI ≥25 and <30 kg/m2: 288 pg/mL (interquartile interval=145) -BMI ≥30 kg/m2: 225 pg/mL (interquartile interval=111) Weak relationship between time since insertion and serum ENG level (P=.1). After adjustment for time since insertion: no association between BMI and serum ENG levels (P=.7) Weak relationship between time since insertion and serum ENG level (P=.1). After adjustment for time since insertion: no association between BMI and serum ENG levels (P=.7)
McNicholas et al, 201524 Quality score=100%	Prospective descriptive study	USA	To evaluate the efficacy of the ENG implant and the 52-mg LNG-IUS beyond the current USA FDA-approved duration (3 and 5 y, respectively)	237 ENG implant users: 123 used it for 4 y 34 used it for 5 y	Mean age 26.0 y (SD 5.4 y) Nulliparous and multiparous BMI: 16.6 to 53.2 kg/m2: BMI <25 kg/m2: 29% BMI ≥25 to <30 kg/m2: 25% BMI ≥30 kg/m2: 46%	Median duration of extended use: 12.5 additional mo (5.1 to 40.5 mo) No pregnancies recorded in years 4 and 5 Failure rate: 0 (97.5% CI 0 to 1.61) per 100 women-years Median serum ENG levels: -After 3 y: 188.8 pg/mL (63.8-802.6) (n=92) -After 4 y: 170.0 pg/mL (67.9-470.5) (n=47) -No difference between BMI groups after 3 y (P=.79) and after 4 y (P=.47)
Ali et al 201614 Quality score=90%	Comparative cohort study	WHO multicentre study: Brazil, Chile, Dominican Republic, Thailand, Turkey, Zimbabwe	To evaluate efficacy, side effects, and reasons for discontinuation following prolonged use of the ENG contraceptive implant for up to 5 y compared to LNG implant and Cu380A	390 ENG implant users who started prolonged use: n=311 for 4 y n=204 for 5 y	Mean age: 27.8 y (SD: 6.1 y; range: 18-43) Nulliparous and multiparous Mean BMI: 23.7 (3.8) kg/m2	Years 4 and 5: no pregnancies during 7060 woman-months of observation After 5 y: cumulative pregnancy rate among ENG implant users of 0.6 per 100 women-years (95% CI 0.2 to 1.8) Personal reasons were the most frequent reason for implant discontinuation in subsequent years of use (17.1%). Most frequent side effects in years 4 and 5: irregular bleeding (46.2%), headache (31.8%), and lower abdominal pain (17.9%). 12% of participants experienced heavy bleeding
McNicholas et al, 201725 Quality score=88%	Prospective descriptive study	USA	To evaluate the efficacy of the ENG implant and the 52-mg LNG-IUS beyond the current FDA-approved duration (3 and 5 y, respectively)	291 ENG implant users who began the EPIC study: 223 used it for >12 additional mo 102 used it for >24 additional mo	Age: 18-45 y 18-22 y: 30.9% 23-29 y: 47.4% 30-34 y: 11.3% 35-45 y: 10.3% Nulliparous and multiparous BMI: <25 kg/m2: 73 (25.1%) 25-30 kg/m2: 67 (23.0%) ≥30 kg/m2: 151 (51.9%)	Average duration of extended ENG implant use: 18.9 additional mo (range: 0-44.1 mo) 444 women-years of observation No pregnancies for a failure rate per 100 women-years: -during the 4th year of use: 0 (97.5% CI 0 to 1.48) -during the 5th year of use: 0 (97.5% CI 0 to 2.65) Median serum ENG levels: -after 3 y: 207.7 pg/mL (range 63.8-802.6 pg/mL) -after 4 y: 166.1 pg/mL (range 25.0-470.5 pg/mL) -after 5 y: 153.0 pg/mL (range 72.1-538.8 pg/mL) -no difference between BMI groups at the end of 3rd year (P=.10) or at the end of 5th year of use (P=.62) -difference according to BMI after 4 y of use (P=.04) (but median rates higher in 25-30 kg/m2 group than in other groups)
Goldstuck and Le, 201826 Quality score=38%	Systematic review	NA	To compare the controlled release of LNG and ENG from LNG-IUS and implants	1 study on prolonged use of the ENG implant: McNicholas et al, 201725	NA	Bennink, 199827: minimum serum ENG level to inhibit ovulation: >90 pg/mL McNicholas et al, 201725 (see summary above): average serum ENG levels in years 4 and 5: about 180 pg/mL
Thaxton and Lavelanet, 201928 Quality score=88%	Systematic review	Brazil, Chile, Dominican Republic, Hungary, Thailand, Turkey, Zimbabwe, USA, and China	To verify the contraceptive efficacy of progestin-only implants beyond currently approved durations of use	Ali et al, 201614; McNicholas et al, 201725; Yao and Du, 2003,29 n=51 (4 y); Zheng et al, 1999,30 n=151 (4 y); Kiriwat et al, 1998,31 n=47 (4 y)	NA	Ali et al, 2016,14 and McNicholas et al, 201725 (see summaries above) Yao and Du, 200329 (article in Chinese) (comparative cohort study): no pregnancies Zheng et al, 1999,30 and Kiriwat et al, 199831 (cohort studies): no pregnancies

B) GUIDELINES
AUTHORS, YEAR OF PUBLICATION, QUALITY SCORE	ORGANIZATION	COUNTRY	TITLE	SPECIFIC RECOMMENDATIONS AND LEVELS OF EVIDENCE FOR ENG CONTRACEPTIVE IMPLANT USE
Dethier et al, 202232 Quality score=86%	SFP	USA	SFP clinical recommendation: extended use of long-acting reversible contraception	Patients should be counselled that the contraceptive effectiveness of the 68-mg ENG implant is maintained beyond the current FDA-approved duration of 3 to 5 y (grade 1A) Extended use of the implant may be offered to patients of any BMI. However, given data are limited in patients with class 3 obesity and above (BMI >40 kg/m2); shared decision making and careful weighing of risks and benefits is warranted (grade 2B) Effectiveness may be reduced with extended implant use in patients taking CYP3A4 inducers; use of shared decision making and careful weighing of risks and benefits in these situations is warranted (grade 2C)

BMI—body mass index, Cu—copper, CYP3A4—cytochrome P450 3A4, ENG—etonogestrel, EPIC—Effectiveness the Prolonged Use of the IUD and Implant, FDA—Food and Drug Administration, IUS—intrauterine system, LNG—levonorgestrel, NA—not applicable, SFP—Society of Family Planning, USA—United States of America, WHO—World Health Organization.

In 2022, the Society of Family Planning (SFP)³² recommended informing patients that the contraceptive efficacy of the 68-mg ENG implant was maintained beyond the 3 to 5 years’ duration currently approved by the US Food and Drug Administration (grade 1A evidence).

Efficacy and safety of the Cu-IUD beyond the approved duration of use (Table 2).^32-39 The studies by Rowe et al from 1997 and 2016 reported no pregnancies in patients using the TCu380A IUD during, respectively, 1440 women-years of observation from 8 to 12 years after insertion (n=356) and 1342 women-years of observation from 7.2 to 11 years after insertion (n=682).^33,37 Combining data from the 1997 study by Rowe et al³³ and a 2005 study from Bahamondes et al (n=228),³⁴ Ti et al³⁸ calculated an overall pregnancy rate of 0.0 per 100 women-years (95% CI 0.0 to 0.8) for years 11 and 12. Sivin³⁵ combined data from the Population Council, the World Health Organization, and Bahamondes et al (2005)³⁴ and reported no pregnancies in 1886 women-years of observations from years 10 through 15.

Table 2.

Summary of publications on prolonged duration of Cu-IUDs: A) Scientific articles and B) guidelines.

A) SCIENTIFIC ARTICLES
AUTHORS, YEAR OF PUBLICATION, QUALITY SCORE	DESIGN	COUNTRY	OBJECTIVE	INCLUDED STUDIES OR DESCRIPTION OF SAMPLE	CHARACTERISTICS OF PARTICIPANTS	RESULTS
Rowe et al, 199733 Quality score=90%	Randomized controlled study	24 Chinese and non-Chinese centres	To compare the long-term efficacy (8, 10, and 12 y) of Cu-IUDs TCu380A and TCu220C	TCu380A: n=1396 insertions (172 completed 12 y) TCu220C: n=3277 insertions (341 completed 12 y)	Mean age: 28.7 y (SD 4.9 y) (range 16-40 y) Multiparous	At the end of the 12th year: Cumulative intrauterine pregnancy rates: 7.0 (SE 0.6) per 100 women for TCu220C and 1.9 (SE 0.5) per 100 women for TCu380A (P<.001). Cumulative ectopic pregnancy rates: 0.7 (SE 0.3) for TCu220C and 0.4 (SE 0.3) for TCu380A 7 pregnancies occurred with TCu220C between 8th and 12th year and none with TCu380A Annual pregnancy rates highest in early years of use and consistently higher for TCu220C than for TCu380A Cumulative withdrawal rates: 37.3 (SE 1.3) and 40.2 (SE 2.1) per 100 women for TCu220C and TCu380A, respectively Expulsion rate of 1.3 per 100 women-years for years 11 and 12 Withdrawal rate secondary to bleeding or pain: 4.6 per 100 women-years between years 10 and 12 No perforations or PID
Bahamondes et al, 200534 Quality score=75%	Descriptive study	Brazil	To evaluate the performance of TCu380A in participants using this IUD for more than 10 y and aged ≥35 y at the end of the 10th year of use	N = 228 Year 11: n=185 Year 14: n=31 Year 16: n=8	Mean (SD) age after 10 y of use: 38.8 (0.4) y Nulliparous and multiparous	Duration of follow-up after 10 y: 1 to 72 mo No pregnancies observed during the 366 women-years of observation. Pearl indices: -Months 1 to 12 (beyond 10 y): 0.0 (95% CI 0 to 1.99) -Months 1 to 24 (beyond 10 y): 0.0 (95% CI 0 to 1.32) Continuation rates in year 11 (n=185), year 14 (n=31), and year 16 (n=8): 67%, 34.3%, and 21.2%, respectively Number of dropouts: year 11=86, year 12=97, year 13=14, year 14=8, year 15=15, year 16=2 Expulsions: 2.1 per 100 and 1.1 per 100 participants for years 11 and 12, respectively; occurred mainly in year 11. One expulsion per year in years 13 to 16, no expulsions in year 15 Withdrawal rates due to bleeding or pain: 1.6 per 100 women-years in year 11, 1.1 in year 12, 2.5 in year 13, 3.4 in year 14, and nil in years 15 and 16
Sivin, 200735 Quality score=100%	Commentary and expert opinion	NA	To examine the extent of contraceptive need among participants aged ≥40 y, and the extent to which this need can be met when starting to use Cu-IUDs at 25 to 35 y	Bahamondes et al, 200534; study 799914: unpublished data from the comparative trial of the TCu220C and the T380A done by WHO (personal communication with T.M.M. Farley and S. Peregoudov, WHO, August 2004); Rowe et al, 199733; internal documented studies of the Population Council	NA	Studies by the Population Council, WHO, and by Bahamondes et al, 2005,34 based on 1886 women-years of observations, from year 10.0 through year 15.0, no pregnancies occurred during use of the TCu380A, nor with collared TCu220C The Population Council, in 2 clinics, during 70 women-years of observation (from 15.0 to 20.0 y after insertion), observed no pregnancies with TCu380Ag Unpublished WHO data (study 799914) showed that bleeding or pelvic pain was the most frequent reason for removal in the initial 2 y and in 5 of the last 6 y The Population Council study of TCu380Ag that went beyond 10 y showed little change in rates of termination for heavy menstrual bleeding between any 5-y period. Incidence of removals because of perimenopausal symptoms rose from 0 in the first decade to 2-3 per 100 women-years in each half of the second decade of continuous use In the WHO 15-y study, no terminations occurred because of PID in either the TCu380 or the TCu220C group between the end of the 10th year and the end of the 15th year. In the 20-y Population Council study, no participants left the study or reported a PID in the second decade, although removals were prescribed in a few participants when actinomycotic-like organisms were found upon repeated cervical smears
Kaneshiro and Aeby, 201036 Quality score=50%	Narrative review	NA	To describe the long-term safety, efficacy, and acceptability of the TCu-380A Cu-IUD	Rowe et al, 199733; Bahamondes et al, 200534; Sivin, 200735	NA	See summaries above
Rowe et al, 201637 Quality score=91%	Randomized controlled trial	Brazil, Chile, China, Mongolia, Slovenia, Thailand, Tunisia, Philippines	To compare the efficacy of TCu380A and 52-mg LNG-IUS over 7 y (with extended follow-up of 8 to 11 y)	TCu380A: n=1871 LNG-IUS: n = 1884	Mean age: 29.9 y (SD 4.95 y) Mostly multiparous (5 nulliparous in the TCu380A group)	Participants in the TCu380A group after 7 y of follow-up: n=682 No pregnancies occurred during the 1342 women-years of observation of TCu380A from 8 to 11 y of use Expulsion rate at 11 y: 9.6 per 100 women-years (95% CI 7.9 to 11.2) Cumulative probability of expulsion per 100 women-years: 5th year: 6.3 (SE=0.7); 6th year: 7.0 (SE=0.7); 7th year: 8.2 (SE=0.8); 8 to 11th year: 9.0 (SE=1.0)
Ti et al, 202038 Quality score=100%	Systematic review	NA	To determine the efficacy and safety of IUDs when used beyond the approved duration	Two studies on TCu380A: Rowe et al, 199733 (n=245) (12 y); Bahamondes et al, 200534 (n=228) (up to 16 y)	NA	Rowe et al, 1997,33 and Bahamondes et al, 200534: see summaries above Pooling these data, the overall pregnancy rate for years 11 and 12 was 0.0 per 100 women-years (95% CI 0.0 to 0.8) Pregnancy rates and adverse event rates during the first 2 y of extended use are comparable to published rates during approved use

B) GUIDELINES
AUTHORS, YEAR OF PUBLICATION, QUALITY SCORE	ORGANIZATION	COUNTRY	TITLE	SPECIFIC RECOMMENDATIONS AND LEVELS OF EVIDENCE FOR CU-IUD USE
Wu and Pickle, 201439 Quality score=71%	SFP	USA	Extended use of the IUD: a literature review and recommendations for clinical practice	Level A recommendations (based on consistent and good-quality patient-oriented evidence): Clinicians can recommend extended use of the following IUDs: -TCu380A: for 12 y (persons who are ≥35 y at the time of initial TCu380A IUD insertion can continue use until menopause with only a small, theoretical risk of pregnancy) -TCu220: for 12 y —Multiload Cu-375: for 10 y -Frameless IUD (330 mm2): for 9 y -Multiload Cu-250: for 4 y Level B recommendations (based on inconsistent or limited-quality patient-oriented evidence): Extended use of the Cu-IUD can be offered to overweight and obese persons
Dethier et al, 202232 Quality score=86%	SFP	USA	SFP clinical recommendation: extended use of long-acting reversible contraception	Patients should be counselled that the contraceptive effectiveness of the 380-mm2 Cu-IUD is maintained beyond the current FDA-approved duration of 10 to 12 y (grade 1B) For patients >30 y at time of insertion of the 380-mm2 Cu-IUD, extended use beyond 12 to 20 y, or to menopause, can be considered as a form of reversible, permanent contraception (grade 1C)

Cu—copper, FDA—Food and Drug Administration, IUD—intrauterine device, IUS—intrauterine system, LNG—levonorgestrel, NA—not applicable, PID—pelvic inflammatory disease, SE—standard error, SFP—Society of Family Planning, USA—United States of America, WHO—World Health Organization.

The expulsion rate was 1.3 per 100 women-years for years 11 and 12, and 0.7 per 100 women-years from 7.2 to 11 years, in Rowe et al (1997)³³ and Rowe et al (2016),³⁷ respectively. In Bahamondes et al (2005),³⁴ the expulsion rates were 2.1 and 1.1 per 100 women-years in years 11 and 12, and 3 expulsions occurred from years 13 to 16. In comparison, in Rowe et al (2016),³⁷ the cumulative expulsion rate was 3.49 per 100 women-years in the first year after insertion, and 7.30 per 100 women-years after 5 years. Bleeding or pain or both led to IUD discontinuation in 4.6 per 100 women-years between years 10 and 12, and 1.2 per 100 women-years from 7.2 to 11 years, in Rowe et al (1997)³³ and Rowe et al (2016),³⁷ respectively. In Bahamondes et al (2005),³⁴ these figures were 1.6, 1.1, 2.5, 3.4, and 0 per 100 women-years at years 11, 12, 13, 14, and 15 to 16, respectively. In Rowe et al (2016),³⁷ the cumulative rates for discontinuation due to heavy bleeding and pain were 2.42 and 1.72 per 100, respectively, 1 year after insertion, and 7.42 and 5.01 per 100, respectively, after 5 years. Other problems were found to be rare or nonexistent.^33-35,37

The 2014 and 2022 SFP guidelines^32,39 recommended prolonged use of the 380-mm² Cu-IUD (equivalent to TCu380A) for 12 years, regardless of the patient’s age at insertion. The 2022 SFP guideline³² recommended that the use of all 380-mm² Cu-IUDs be extended beyond 12 years, and up to 20 years or until menopause, for patients older than 30 years of age at the time of insertion.

Efficacy and safety of the LNG-IUS beyond the approved durations of use (Table 3).^{24-26,32,37-53} Prolonged use was reported only for 52-mg LNG-IUS. No pregnancies were reported in 9 studies at different durations of use, ranging from year 6^49,50 to years 6 and 7,^40,41 years 6 to 8,⁴⁴ year 8,^51,52 years 9 and 10,⁵³ years 6 to 11,³⁷ and years 6 to 15.⁴⁹ A total of 2 pregnancies were reported in year 6 in 2 studies^24,52 (n=625), and 4 pregnancies were reported in year 7 in 3 studies (n=1228).^25,51,52 In a 2020 meta-analysis of 4 studies^25,38,41,49 by Ti et al,³⁸ the cumulative pregnancy rate during years 6 and 7 was 0.02 per 100 women-years.

Table 3.

Summary of publications on prolonged duration of LNG-IUS: A) Scientific articles and B) guidelines.

A) SCIENTIFIC ARTICLES
AUTHORS, YEAR OF PUBLICATION, QUALITY SCORE	DESIGN	COUNTRY	OBJECTIVE	INCLUDED STUDIES OR DESCRIPTION OF SAMPLE	CHARACTERISTICS OF PARTICIPANTS	RESULTS
Sivin et al, 199140 Quality score=91%	2-y cohort study after 5-y randomized trial	USA, Brazil, Egypt, Singapore	To describe the events that occurred during the 7-y period following the insertion of either 52-mg LNG-IUS or TCu380A IUD from a trial undertaken by the Population Council	At the beginning of the 7th y of follow-up: 172 participants with a 52-mg LNG-IUS	Age: 18-38 y Multiparous	No pregnancies were reported among users of either IUD during the 6th or 7th y. Cumulative pregnancy rate for LNG-IUS over 7 y: 1.1 per 100; Pearl index: 0.18±0.07 per 100. No ectopic pregnancies in either group No expulsions in the LNG-IUS group in years 6 and 7. Cumulative expulsion rate: 11.7 per 100 Average annual discontinuation rates for adverse events, expulsion, or pregnancy lower in years 6 and 7 than before 7-y data showed an increase in hemoglobin of 1.44 g/dL in the LNG-IUS group compared with admission values In the 5 clinics where the LNG-IUS was studied, the 7-y continuation rate was 24.9 per 100
Rönnerdag and Odlind, 199941 Quality score=75%	Prospective descriptive study	Sweden	To evaluate, according to various health parameters, long-term effects of using the 52-mg LNG-IUS for more than 2, 5-y periods	Period 1 (mean duration 6.6 y): 1st LNG-IUS inserted in 300 participants for 5 y; 100 participants extended for 18 mo Period 2: 2nd LNG-IUS inserted in 82 of 100 participants for 5 y	Mean age: 30.3 (SD 4.9) y Multiparous	No pregnancies during the total follow-up period of 12.2 y (range: 9.8-13.3 y). At the end of the study, the mean age of participants was 44.7 y (range: 33.5-51.5 y) At the end of period 1: 26% of participants were in amenorrhea, 70% had regular light bleeding, and 4% had infrequent light bleeding. In the period following period 1 (ie, 5 to 6.5 y), only slight changes occurred in the bleeding pattern After approximately 12 y of follow-up: mean increase in hemoglobin: 1.35 g/dL (CI 1.1 g/dL to 1.6 g/dL); slight increase in systolic and diastolic blood pressure; body weight increased by 0.49 kg/y throughout the study period; 7 (8.5%) participants reached menopause with typical symptoms and elevated FSH level
Hidalgo et al, 200942 Quality score=88%	Prospective descriptive study	Brazil	To assess serum LNG and E2 levels and endometrial thickness every 6 mo between 7 and 8.5 y after insertion of 52-mg LNG-IUS	Subsample of 86 participants from a previous study (Diaz et al, 200043) out of 256 participants who received an LNG-IUS: 67 remaining after 7 y of use	Mean age: 33.4 y (SD 0.8 y) Mean BMI at 84 mo: 26.8 (range 18.3-47.3) kg/m2	Serum LNG levels (mean±SEM): First 2 mo after insertion (n=86): 253±27 pg/mL (range 86-760) At 7th y of use (n=67): 137±12 pg/mL (range 23-393) (P<.006) At 8th y of use (n=25): 132±18 pg/mL (range 51-505) After 8.5 y of use (n=2): 119±9 pg/mL (range 110-129) (nonsignificant compared with year 7) E2 levels (mean±SEM): -after 27-50 days (n=86): 99.2±11.4 pg/mL -after 7 y of use (n=67): 86.1±10.4 pg/mL -after 8 y of use (n=34): 103.0±20.9 pg/mL -after 8.5 y of use (n=8): 93.6±29.8 pg/mL Endometrial thickness (mean±SEM): -at 7 y of use: 2.79±0.1 mm -at 8 y of use: 3.75±0.3 mm -at 8.5 y of use: 3.76±0.5 mm Amenorrhea: 42% at year 7 of use, 37.5% at 8.5 y Participants with low LNG values (<100 pg/mL): E2 levels remained within the range observed in the follicular phase of the normal menstrual cycle. Mean endometrial thickness increased with statistical significance from 2.7 mm at 7 y to 4.1 mm at 8 y of use A weak correlation was found between high BMI (kg/m2), high weight, and low serum LNG levels
Seeber et al, 201244 Quality score=100%	Retrospective descriptive study	Austria	To determine the extent and evolution over time of LNG plasma distribution in patients with LNG-IUS	113 participants among which 35 used LNG-IUS ≥6 y	Mean age: 42.3 (SD 8.2) (range 18.3-55.4) y; mean BMI: 24.3 (SD 4.3) (range 17.8-38.5) kg/m2	Mean±SD plasma LNG level in all participants: 147±59 pg/mL (range 63-361 pg/mL). Negative correlation between plasma LNG level and duration of LNG-IUS use: r=−0.353, P<.001 Mean±SD LNG levels: -1st y of use: 191±71 pg/mL -2nd and 3rd y: 157±68 and 134±41 pg/mL, respectively -4th and 5th y: 150±47 and 141±59 pg/mL, respectively -6th and 7th y: not significantly different from years 2 to 5 -8th y: 117±59 pg/mL LNG-IUS group ≤5 y: 157±62 pg/mL vs prolonged-use group: 126±44 pg/mL (P=.014) Negative correlation between BMI and plasma LNG levels (r=−0.239, P=.012). Mean LNG level: significantly lower in participants with BMI ≥25 kg/m2 compared with BMI <25 kg/m2 (P=.045). Mean LNG level according to BMI (kg/m2): BMI <20 (n = 13)=165±57 pg/mL; BMI 20-24.9 (n=57)=152±59 pg/mL; BMI 25-29.9 (n=28)=141±64 pg/mL; BMI ≥30 (n=11)=119±43 pg/mL No significant difference between plasma LNG levels and vaginal bleeding (P=.684) No pregnancies in either group
McNicholas et al, 201524 Quality score=100%	Prospective descriptive study	USA	To evaluate the efficacy of the ENG implant and the 52-mg LNG-IUS beyond the current USA FDA-approved duration (3 and 5 y, respectively)	263 users of 52-mg LNG-IUS (197.7 women-years of use); >1 additional year of use: n=108	Mean age: 32.2 y (SD 5.9 y) BMI (kg/m2): <25: 31.9% 25-30: 22.8% ≥30: 45.3%	1 pregnancy during follow-up period (<1 mo after approved duration due to partially expelled IUS) for a median duration of extended use of 11.7 additional mo (range: 4.7-36.2 mo) Estimated failure rate: 0.51 (95% CI 0.01 to 2.82) per 100 women-years Sensitivity analysis excluding participants older than 35 y at the time of recruitment showed that the estimated failure rate per 100 women-years beyond the FDA-approved duration was 0.79 (95% CI 0.02 to 4.39) for the LNG-IUS
Rowe et al, 201637 Quality score=96%	Randomized controlled trial	Brazil, Chile, China, Mongolia, Slovenia, Thailand, Tunisia, Philippines	To compare rates of unintended pregnancy, method continuation, and reasons for withdrawal in participants using the 52-mg LNG-IUS or TCu380A over a 10-y period	Out of 3836 eligible participants: 1884 with LNG-IUS (7903 women-years of observation)	Mean age: 29.8 y (SD 5.12 y) Nulliparous (4 participants) and multiparous	398 participants remained in the LNG-IUS group at the 8th year of observation Cumulative pregnancy rate at the end of year 7: 0.53 (SE=0.21) per 100 women-years and no pregnancies during the 681 women-years of LNG-IUS use from 8 to 11 y 104 expulsions of LNG-IUS (7-y cumulative rate: 8.18 (SE=0.84) and 1 perforation For each of the 7 y, the cumulative rate of method discontinuation due to bleeding problems was higher in the LNG-IUS group (due to amenorrhea) than in the TCu380A group (39.2% vs 13.8%). The cumulative amenorrhea rate at 7 y was 26.1% for the LNG-IUS vs 0.5% for the TCu380A
McNicholas et al, 201725 Quality score=88%	Prospective descriptive study	USA	To evaluate the efficacy of the ENG implant and the 52-mg LNG-IUS in participants using the method beyond the current USA FDA-approved duration of 3 and 5 y	496 users of 52-mg LNG-IUS: >12 additional mo: n=347; >24 additional mo: n=160	Age, y: 18-22: 1.4% 23-29: 37.5% 30-34: 28.2% 35-45: 32.9% BMI (kg/m2): <25: 32.8% 25-30: 21.5% ≥30: 45.7% Nulliparous (29.6%) and multiparous	Mean duration of LNG-IUS extension: 19.3 additional months (range 02.-41.1 mo) for 696.9 women-years of use 2 pregnancies during the follow-up Failure rate in the 6th year of use: 0.25 (95% CI 0.04 to 1.42) per 100 women-years Failure rate in the 7th year of use: 0.43 (95% CI 0.08 to 2.39) per 100 women-years
Goldstuck and Le, 201826 Quality score=38%	Systematic review	NA	To compare the controlled release of LNG and ENG from LNG-IUS and implants	Data from studies of Bayer Pharmaceuticals, 201745; Xiao et al, 199046; Goldstuck, 201747; McNicholas et al, 201725; Hidalgo et al, 200942	NA	Serum levels of LNG (mean±SD) at 1 mo after insertion of 52-mg LNG-IUS: 250 pg/mL (range 150-220) Serum levels of LNG (mean±SD) at 1 y: 180±66 pg/mL (range 120-200) Serum levels of LNG (mean±SD) at 7 y: 137±12 pg/mL (range 23-393) Initial serum levels of about 200 pg/mL are around the threshold for ovulation inhibition which is why up to 40% of users experience ovulation inhibition initially. Contraceptive efficacy well maintained at lower serum levels because intrauterine release rates of LNG as low as 3 μg/d are enough to produce cervical mucus and endometrial changes that are antifertility
Creinin et al, 201848 Quality score=64%	Cohort study (abstract)	USA	To evaluate LNG plasma concentrations for up to 7 y in nonobese and obese participants using a 52-mg LNG-IUS	40 participants in a planned substudy (19 obese, 21 nonobese) who had samples taken between insertion and 84 mo (7 y)	Age: 16-45 y Mean BMI: ≥30 kg/m2: 38.6±5.8 (30.4-49.2) kg/m2 <30 kg/m2: 24.6±2.9 (18.9-29.5) kg/m2	Peak plasma concentrations of LNG: -<30 kg/m2: 603 pg/mL at week 1, 492 pg/mL at week 14, and <300 pg/mL at 1 y -≥30 kg/m2: <300 pg/mL at day 30 and thereafter From 3 to 84 mo after insertion, LNG concentrations were 21% to 41% lower in the ≥30 kg/m2 group (P<.01 for all mo)
Bahamondes et al, 201849 Quality score=100%	Retrospective descriptive study	Brazil	To evaluate the contraceptive efficacy and clinical performance of a 52-mg LNG-IUS in a cohort of participants who have used the LNG-IUS beyond 60 mo of its approved lifespan (up to 15 y of use)	776 participants who used the LNG-IUS beyond 60 mo: 5-6 y: n=61 7-8 y: n=96 9-10 y: n=31 11-12 y: n = 17 ≥13 y: n = 16	Mean (SD) age: 32.0 (SD 0.24) y Nulliparous and multiparous	Median duration of use of 73 mo (6.1 y) (61-184 mo) (5.1-15.3 y) No pregnancies reported Cumulative rates per 100 women of discontinuation (calculated by period after month 60 of initiation of use) due to: -Expulsion: at 5-6 y, 0.3 (0-0.7); 7-8 y, 0.6 (0-1.4); 9-10 y, 0; 11-12 y, 1.1 (0-2.4); ≥13 y, 2.1 (0-4.4) -Bleeding or pain: at 5-6 y, 0.2 (0-0.6); 7-8 y, 2.3 (0-4.6); 9-10 y, 0; 11-12 y, 0; ≥13 y, 0 -Menopause: at 5-6 y, 0.4 (0.1-0.7); 7-8 y, 0.8 (0.3-1.3); 9-10 y, 0; 11-12 y, 1.8 (0.8-2.7); ≥13 y, 5.7 (2.0-9.4) -Infection: at 5-6 y, 3.2 (0.7-5.6); all other periods, 0
Westhoff et al, 202050 Quality score=86%	Prospective descriptive study	USA	To report data supporting contraceptive use of a 52-mg LNG-IUS for ≥6 y	1751 participants: 612 participants (5 y of use); 321 participants (6 y of use); 703, 402, 191, and 122 participants in the safety population had 5, 6, 7, and 8 y of use, respectively	Mean (SD) age: 27.3 (SD 5.7) y: 16-35 y: n=1600 36-45 y: n=151 Mean (SD) BMI (kg/m2): 26.9 (SD 6.8) (BMI ≥30: 25.1% and BMI ≥40: 5.3%) Nulliparous (57.7%) and multiparous	No pregnancies in year 6; Pearl index for year 6 (n=229; 5091 cycles) was 0.00 (95% CI 0.00 to 0.94); cumulative Pearl index from 0 to 6 y: 0.18 per 100 women-years (95% CI 0.08 to 0.33); life table pregnancy rate: 0.87 (95% CI 0.44 to 1.70) Most frequent adverse effects over ≥6 y: bacterial vulvovaginal infections (n=305, 17.4%), yeast infections (n=291, 16.6%), or urinary tract infections (n=296, 16.9%) 14 (0.8%) diagnostics of pelvic infection: 2 after year 4, 1 in year 6, and 1 in year 8 Amenorrhea within 90 d after 6th year in 141/349 (40%) Most frequent reasons for discontinuation: adverse event (n=329, 19.2%), desire of pregnancy (n=265, 15.5%), loss to follow-up or withdrawal of consent (n=259, 15.1%), relocation away from a study site (n=111, 6.5%); in the 36-45 y age group, 45 of 122 (36.9%) discontinued due to mandatory exit after 8 y Partial or complete expulsion in 68 participants (4.0%), with 2 partial expulsions per year in years 6 and 7; 40 (2.3%) discontinuations due to bleeding, including 1 in year 6 and none thereafter
Ti et al, 202038 Quality score=100%	Systematic review	NA	To determine the efficacy and safety of IUDs when used beyond the approved duration	4 studies on LNG-IUS (total: 2098 participants): Rowe et al, 201637 Bahamondes et al, 201849 McNicholas et al, 201725 Rönnerdag and Odlind, 199941	NA	See summary above No pregnancies in 3 of 4 studies during 1-3 y of extended use37,41,49 and 2 pregnancies reported in the 4th study25; pregnancy rate for years 6 and 7 combined: 0.02 per 100 women-years (95% CI 0.00 to 0.29) Expulsions: Rowe et al, 201637: expulsion rates of 0.7 and 1.2 per 100 participants in years 6 and 7, respectively; Bahamondes et al, 201849: expulsion rate 0.3 (95% CI 0.0 to 0.7) per 100 participants in years 6-7 Menstrual-related adverse events: Rowe et al 201637: discontinuation rate due to bleeding in 3.7 per 100 in year 6 and 3.1 per 100 in year 7, and due to pain in 0.4 per 100 in years 6 and 7. Bahamondes et al 201849: discontinuation rate of 0.2 per 100 participants due to bleeding or pain in years 6 and 7 Pregnancy rates and rates of adverse events or discontinuation due to side effects during the 2 y of extended use comparable to published rates during the approved duration of use
Creinin et al, 202251 Quality score=100%	Prospective descriptive study	USA	To evaluate the efficacy and safety of the 52-mg LNG-IUS over a period of continuous use of up to 8 y	1714 participants who had a 52-mg LNG-IUS inserted: 576 started year 7 478 completed year 7 343 completed year 8	At the start of year 7: 499 participants ≤35 y and 465 participants aged ≤39 y Mean age: 34.4±6.3 y (19.3% ≥40 y) Mean BMI: 27.1±6.9 kg/m2 BMI ≥30 kg/m2: 25.5% Nulliparous (58.3%) and multiparous	11 pregnancies within 8 y, including 7 (64%) ectopic for a cumulative pregnancy rate of 1.23 per 100 women-years (0.69-2.51). 2 pregnancies in year 7 (1 ectopic), and 0 in year 8. In year 7, Pearl index according to age at the start of year 7: ≤35 y: 0.46 (95% CI 0.06 to 1.66) per 100 women-years, similar to that of ≤39 y: Pearl index=0.49 per 100 (95% CI 0.06 to 1.78) 71 expulsions (4.1%), most frequently in multiparous participants (47 of 71), including 3 in year 7 and 2 in year 8 44 withdrawals (2.6%) of LNG-IUS due to unacceptable bleeding over 8 y (5 [0.8%] in years 6, 7, and 8) Amenorrhea rate in years 7 and 8 was 39%, with rates ranging from 37% to 42% in years 3 to 8 16 cases of PID (0.9%): 1 in year 7 and 1 in year 8 Premature discontinuation most often for participants wishing to become pregnant (n=315, 18.4%) or for reasons specific to the study (n=572, 33.4%), and not for complications associated with LNG-IUS use or the desire to use another contraceptive method
Jensen et al, 202252 Quality score=100%	Prospective descriptive study	USA	To evaluate the efficacy and safety of the 52-mg LNG-IUS during prolonged use (6-8 y)	501 participants enrolled in the study: 362 (72.3%) started year 6; 243 started year 8; 223 completed year 8	Mean age: 29.4 (SD 3.1) (range 18-35) y Nulliparous (47.2%) and multiparous BMI 27.9 kg/m2 (SD 6.9)	26 participants (7.2%) were lost to follow-up during the study period 2 pregnancies occurred in years 6 to 8, both with 52-mg LNG-IUS in situ: year 6: 1 pregnancy of undetermined location, spontaneously resolved; year 7: 1 ectopic pregnancy (nulliparous) resolved with methotrexate after IUS removal Pearl index for years 6, 7, and 8: 0.34 (95% CI 0.01 to 1.88), 0.40 (95% CI 0.01 to 2.25), and 0.00 (95% CI 0.00 to 1.90), respectively; Kaplan-Meier 3-y cumulative failure rate for years 6, 7, and 8 was 0.68 per 100 (95% CI 0.17 to 2.71) 4 perforations: 2 cases diagnosed when LNG-IUS was removed at the end of the study; 3 of 4 cases with ultrasound diagnosis of myometrial embedding and 1 case of laparoscopic removal 5 expulsions (1.4%). No cases of PID Discontinuation: 3% due to irregular bleeding, 2.2% due to pain. Most frequent reason for withdrawal: desire for pregnancy in 12.2% (44 of 362) of participants
Chen et al, 202353 Quality score=100%	Prospective descriptive study	USA	To evaluate the efficacy and safety of the 52-mg LNG-IUS over a period of continuous use lasting more than 8 y (8-10 y)	1714 participants who had inserted a 52-mg LNG-IUS: 339 (>8 y) 166 (>8.5 y)	Mean age at the start of year 9: 35.3±6.2 y (74.9% <40 y); nulliparous (56.3%) and multiparous; BMI ≥30 kg/m2: 25.3%	No pregnancies among 2062 and 966 cycles of exposure in years 9 and 10, respectively 1 partial expulsion at 19 days of use in year 9. No perforation No pelvic infections or serious adverse events related to the LNG-IUS in years 9 and 10 Absence of bleeding or spotting reported by 33 (39.8%) and 28 (36.4%) participants in years 9 and 10, respectively. No discontinuation due to bleeding during the 2-y period
Jensen et al, 202354 Quality score=100%	Prospective descriptive study	USA	To estimate in vivo LNG release and systemic exposure to LNG and SHBG for 52-mg LNG-IUS until the end of 8 y of use and to compare with 19.5 mg LNG-IUS and 13.5 mg LNG-IUS during their respective approved periods of use	501 participants enrolled in the study: 362 (72.3%) started year 6 and 223 (44.5%) completed year 8	Age: 18-35 y; mean age: 29.4 y (SE=3.1) Nulliparous (47.2%) and multiparous Mean BMI: 27.9 kg/m2 (SE=6.9)	Total plasmatic concentrations of LNG (geometric mean [CV%]) decreased from 127 ng/L (41.5%) at the beginning of year 6 to 83.4 ng/L (49.6%) at the end of year 8 Geometric mean (CV%) of SHBG decreased from 42.3 nmol/L (50.2%) to 38.6 nmol/L (55.5%) over the 3-y study period LNG concentrations of participants with body weight ≤55 kg (n=24) in years 5 to 8: 27% to 43% higher than concentrations for participants weighing >55 kg in 52-mg LNG-IUS from 300 participants whose IUDs were removed between 5 and 8 y after insertion: median residual LNG content fell from about 25 mg at the start of year 6 to about 13.5 mg at the end of year 8 (annual decline of about 20%, or a total of 46%) Estimated plasma LNG concentrations (CV%) for 52-mg LNG-IUS: 100 ng/L (39.9%) after 8 y, similar to those for 19.5 mg LNG-IUS (84.8 ng/L [39.9%]) after 5 y, and higher than those for 13.5 mg LNG-IUS (58.1 ng/L [40.8%]) after 3 y Estimated geometric mean of unbound LNG (CV%): decrease from 2.82 ng/L (24.2%) at 15 d to 1.91 ng/L (24.1%) at 5 y and to 1.56 ng/L (24.8%) at 8 y Typical release rates estimated in vivo for the 52-mg LNG-IUS decreased from 21.6 μg/d 15 d after insertion to 10.7 μg/d at 5 y and 7.04 μg/d at 8 y. At the end of year 8, the LNG release rate of 52 mg LNG-IUS (7.04 μg/d) was similar to that of the 19.5-mg LNG-IUS after 5 y (about 7.6 μg/d) and about 28% higher than that of the 13.5 mg LNG-IUS after 3 y (about 5.5 μg/d)

B) GUIDELINES
AUTHORS, YEAR OF PUBLICATION, QUALITY SCORE	ORGANIZATION	COUNTRY	TITLE	SPECIFIC RECOMMENDATIONS AND LEVELS OF EVIDENCE FOR LNG-IUS USE
Wu and Pickle, 201439 Quality score=71%	SFP	USA	Extended use of the IUD: a literature review and recommendations for clinical practice	Level A recommendation (based on consistent and good-quality patient-oriented evidence): Clinicians should counsel that extended IUD use is currently an off-label practice, but likely highly effective among parous persons who are at least 25 y at the time of IUD insertion Clinicians can recommend extended use of 52-mg LNG-IUS for 7 y Level B recommendations (based on inconsistent or limited-quality patient-oriented evidence): Extended use of the 52-mg LNG-IUS can be offered to overweight and obese persons
Dethier et al, 202232 Quality score=86%	SFP	USA	SFP clinical recommendation: extended use of long-acting reversible contraception	Patients should be counselled that the contraceptive effectiveness of the 52-mg LNG-IUS is maintained beyond the current FDA-approved duration of 7 to 8 y (grade 1B) Patients should be counselled that the contraceptive effectiveness with extended use of IUDs is independent of BMI (grade 1B) Since no studies describe pharmacokinetics or contraceptive efficacy beyond the current FDA-approved windows of 3 y for the 13.5-mg LNG-IUS or 5 y for the 19.5-mg LNG-IUS, we cannot draw definitive conclusions on the extended use of these IUDs Patients taking CYP3A4-inducing medications may be offered extended use of IUDs (grade 1C)

BMI—body mass index, Cu—copper, CV—coefficient of variance, CYP3A4—cytochrome P450 3A4, E2—estradiol, ENG—etonogestrel, FDA—Food and Drug Administration, FSH—follicle-stimulating hormone, IUD—intrauterine device, IUS—intrauterine system, LNG—levonorgestrel, NA—not applicable, PID—pelvic inflammatory disease, SE—standard error, SEM—standard error of the mean, SFP—Society of Family Planning, SHBG—sex hormone—binding globulin, USA—United States of America.

Expulsions occurred rarely per additional year in several descriptive studies.^40,50-53 Rowe et al (2016)³⁷ observed an expulsion rate of 0.7, 1.2, and 0.8 per 100 women-years, in years 6, 7, and from years 8 to 11, respectively. In Bahamondes et al 2018,⁴⁹ the cumulative rates of discontinuation (per 100 women-years) due to expulsions were 0.3, 0.6, 0, 1.1, and 2.1 for years 5 to 6, 7 to 8, 9 to 10, 11 to 12, and 13 to 14, respectively.

Regarding menstrual-related adverse events, most studies reported discontinuation rates of less than 5% during each additional year of use.^37,40,50-53 In Bahamondes et al (2018),⁴⁹ cumulative rates (per 100 women-years) of discontinuation due to bleeding or pain were 0.2, 2.3, and 0 for years 5 to 6, 7 to 8, and thereafter, respectively. Other adverse events occurred in very low frequencies in all studies.

Prolonged use of the 52-mg LNG-IUS for 7 years³⁹ or 8 years³² is recommended by the SFP, which also states that contraceptive effectiveness with extended use of 52-mg LNG-IUS is independent of BMI,^32,39 and that patients taking cytochrome P450 3A4 inducers might be offered extended use of this IUS.³²

DISCUSSION

The evidence reported no pregnancies with prolonged ENG implant use for 1 year^14,25,28 and 2 years,^14,25 with median serum ENG levels remaining high enough to inhibit ovulation during these additional years and not associated with BMI.^23-25

For the TCu380A Cu-IUD, 3 large studies observed participants for prolonged use: 1 during years 8 to 11 in 682 women,³⁷ 1 during years 8 to 12 in 356 participants,³³ and 1 during years 11 to 16 in 228 participants.³⁴ These studies, mainly carried out in multiparous participants who were under 30 years of age at insertion, did not report any pregnancy.

For the 52-mg LNG-IUS, only a minimal number of pregnancies were reported between 6 and 15 years of use in nulliparous and multiparous participants. Although a statistically significant negative correlation was observed between plasma LNG concentrations and duration of use,⁴⁴ these concentrations^26,54 and in vivo release rates for 52-mg LNG-IUS⁵⁴ after 7 or 8 years of use were similar to or higher than for 19.5-mg and 13.5-mg LNG-IUS at the end of their approved durations of use (5 and 3 years, respectively). Levonorgestrel levels, although lower in individuals with higher BMI or weight, did not appear to affect effectiveness.^42,44,48,54

This reassuring evidence indicate that the long-term efficacy of the ENG implant, TCu380A IUDs, and the 52-mg LNG-IUS are comparable to their short-term efficacy and superior to those of short-acting reversible methods.⁵⁵ Society of Family Planning guidelines^32,39 recommend that the ENG implant be used for 5 years, the TCu380A for 12 years, and the 52-mg LNG-IUS for 8 years in all participants. The SFP also recommends that all 380-mm² Cu-IUDs be used for 12 to 20 years, or until menopause if inserted after the age of 30.

Limitations

Despite the quality of the studies, none compared the efficacy of prolonged use of LARCs with the efficacy of LARCs replaced after each approved duration of use in a randomized trial. Low pregnancy rates might be explained by the fact that most of the participants who were followed in long-term studies might have low fertility and that their fertility declines with age. The selection of participants recruited in long-term studies, often healthy multiparous individuals, might not be representative of the general population and limits the generalizability of the results. Pharmacokinetic data, obtained on very small samples, do not provide a clear understanding of how much of the efficacy of the LNG-IUS is attributable to LNG or to the shape and size of the IUS. Data on the long-term efficacy of the LNG-IUS and the ENG implant according to BMI, or in relation to other medications, are still limited. High upfront cost of LARCs and the individual economic situations of persons interested in these methods might be factors limiting their use. Further studies are needed to clarify these aspects.

Life-table statistics expressed in terms of “women-years” are reported from original studies. However, we acknowledge that these data are relevant to transgender men and other gender-diverse people with a uterus who desire contraception.

Conclusion

The evidence and several guidelines support the extended use of LARCs. Such an extension can be very useful in standard living conditions and in situations of social and health care service disruption, or for populations living in remote areas. With the help of their health care providers, individuals should discuss the benefits versus the risks associated with LARC use beyond the limits currently recommended for birth control.