Abstract
Study Objective:
There is inconsistent data regarding hormonal contraception and weight. Weight concerns may deter teens from using highly effective contraception such as the etonogestrel subdermal implant (ENG). There is little literature about weight gain and adolescent ENG use, most studies involve adult women. The purpose of this study is to evaluate weight/BMI change in adolescent and young adult ENG users compared to non-users.
Design:
Retrospective chart review of 197 ENG users and age, race, BMI, and follow-up time-matched controls.
Setting:
Adolescent medicine clinic.
Participants:
Individuals who had been using ENG for ≥ 6 months were eligible. Control group of non-ENG users who had been seen during the same period was identified to compare weight/BMI over time. Cases were matched to controls on age, BMI, race.
Main outcome measures:
Electronic medical records were reviewed for weight/BMI change and ENG side effects. Study designed to have 80% power to detect a 2 kg weight difference between cases and controls.
Results:
Participant mean age was 17±2 years. Mean follow-up was 24.5±9.3 months. 43/197 ENG users removed implant early; 3/43 (6.3%) patients cited weight gain as the primary reason for removal. Mean weight change for ENG users was +3.6±7.8 kg vs +3.1±5.9 kg for controls (p=0.43); mean BMI change was +1.3±2.9 in cases vs. +1.0±2.3 in controls (p=0.204). Overall regression analyses found no group differences among cases and controls.
Conclusions:
Long-term ENG use does not lead to significant weight gain in this sample of adolescent and young adult women. This study supports the statement that ENGs are an effective and weight neutral option.
Keywords: contraceptive implant, weight, adolescents, contraception, etonogestrel
Introduction
Multiple national organizations including the American Academy of Pediatrics and the Society for Adolescent Health and Medicine recommend long-acting reversible contraception (LARC) as first-line contraception for adolescents.1 When choosing among LARC methods, adolescents more often opt for the contraceptive implant versus an intra-uterine device (IUD), as compared to adults.2 Etonogestrel subdermal implants (ENGs) are a form of LARC, a progestin only birth control with a contraceptive efficacy of 99.9%, effective for at least three years.
There are many factors which affect contraceptive method decision-making which have been well described in the literature (cost, reliability, side effects, patient adherence, provider recommendations, peer/family influence). Concern about weight gain can deter the initiation and/or continuation of contraceptive methods, particularly for adolescent and young adult women.3,4 One of the possible side effects of depot medroxyprogesterone acetate (DMPA, an injectable progestin only contraceptive method) is weight gain, with an average 8 pound weight increase over two years of use.5 Previous studies have found inconsistent data regarding other progestin only birth control methods (including ENG) and weight gain.3,6 In a 2006 retrospective study of ENG users, 5% of study participants discontinued the method with the complaint of weight gain, but most of those participants were adult users.7 In a 2013 study of young adult LARC users (more than 50% were between 21 and 30 years of age), almost 20% of study participants discontinued ENG use with complaints of weight gain.8 A sub-study of the CHOICE project analyzed weight changes in young adult women using progestin-only contraceptive methods. The prospective cohort included 427 young adult women with a mean age of 24.4 years and found that ENG users experienced a 2.1 kg weight gain as compared to 0.2 kg for non-hormonal copper IUD users.9 Adjusted analyses found no difference in weight gain amongst all progestin contraceptive methods as compared to the copper IUD. This study did find that black race was associated with significant weight gain (1.3 kg) amongst all users of progestin only contraceptive methods (ENG, DMPA, and Levonorgestrel IUD),8 but there was no separate analysis of contraceptive method type. A more recent study (2018) found a small but statistically significant weight increase in adult (18–44 years) ENG users as compared to users of the copper IUD.10 In nearly all of the above referenced studies, participants were adult women whose physical growth was complete. There is little information about weight gain and ENG use in adolescents; this is the reason for our study.
Thus the purpose of this study is to evaluate weight and BMI changes in adolescent and young adult users of ENG compared to non-users. We hypothesized that adolescent and young adult women who use the contraceptive implant for more than 6 months will have a measurable change in weight and/or BMI compared to controls. Since one of the reported side effects of DMPA is weight gain, we hypothesized that adolescents who had used this contraceptive method prior to ENG insertion would be at increased risk for weight/BMI change.
Materials and Methods
Participants
This is a retrospective chart review of 197 adolescent and young adult women who chose ENG insertion at an adolescent medicine clinic in Nashville, TN. This is a primary care and sub-specialty clinic that routinely provides contraceptive counseling and management options. Nursing and clinical staff enter adolescents who chose LARC methods in a database panel which is housed within the institutional electronic medical record system. This database includes patient name, medical record number, and ENG insertion date and removal (if applicable). Based on this database, we reviewed medical chart data for all adolescent females ages 12–22 who received the ENG implant between August 2011 and August 2016 in this clinic. All adolescents who had the ENG implant in place for ≥ 6 months were eligible and selected as cases, and then followed for the duration of ENG use. ENG users with previous pregnancy were not excluded but this data was abstracted from their chart and used in analysis.
A normal, natural part of adolescent pubertal development involves linear and weight growth. Therefore, distinguishing between normal and abnormal weight gain can be difficult. To aid in analysis, a group of control patients who had been seen at the same clinical site for a well visit during the same time frame (2011–2016) was identified to compare weight change/trajectory over time.
Case-control matching
The following objective weight variables were collected for cases: weight at insertion (time 0), weight at last documented follow-up clinic visit (or removal, time 1), BMI at insertion (time 0), BMI at last documented follow-up clinic visit (or removal, time 1), and BMI percentile at insertion (time 0) and at last documented follow-up clinic visit (or removal, time 1). As per office triage procedures at this clinical site, weight and height were measured at each visit and entered into the electronic medical record by a licensed medical assistant. BMI is calculated by the electronic medical record based on the visit height and weight. BMI percentile (for those less than 20 years of age) was also abstracted from electronic medical record chart review. Post-insertion office visit notes and phone call records were reviewed to obtain objective weight and BMI values, patient reported side effects, and reasons for early ENG discontinuation (if applicable).
All adolescent females who had been seen at the same clinic site for an annual well physical examination visit during the study period (2011–2016) were eligible to be controls. Cases were matched to controls with respect to age, BMI (+/− 1 kg/m2) and race (based on demographic data at the time of ENG insertion). Given the potential weight gain association with DMPA, controls with current or previous DMPA use were excluded. Controls with a recent (as defined as < 1 year ago) or current pregnancy were also excluded. Controls were ineligible if they had a previous or current ENG use. Those utilizing oral contraceptive pills, transdermal patch or vaginal ring were not excluded as these contraceptive methods are thought to be weight neutral. Since this study was based on chart review data, we were unable to screen for other medical conditions at the time of inclusion. Based on electronic medical record review, we identified 1999 possible control subjects. From this sample, a single control was matched to a single case using the following criteria: case age at the time of ENG insertion (within 6 months), race, and follow-up. Follow-up was defined by case duration of ENG use. For cases, months since ENG insertion was obtained from the electronic database and we identified controls who had been seen in the clinical setting and had a weight documented on a similar time frame (± 1 month). As an example, if a 16 year old African-American, non-Hispanic female case with a BMI of 17 kg/m2 had ENG in place for 20 months, we identified a 16 year old African-American non-Hispanic control with a similar BMI of 17 kg/m2 (±1) who had been seen in the clinic for an annual well visit within the study period (time 0), and then had a follow-up visit 20 months later (time 1). Both authors were responsible for matching controls and followed the same procedure.
The following objective weight variables were collected for controls: weight and BMI at time 0 (including BMI percentile if ≤ 20 years), weight and BMI/BMI percentile at time 1 (which was based on months since ENG insertion for the age-matched case). As was done with case subjects, weight and height were measured at each clinical visit and entered into the electronic medical record by a licensed medical assistant. BMI percentile was abstracted from electronic medical record during chart review. Additional demographic and patient characteristics obtained in data abstraction for both cases and controls included: race, insurance status, sexual history, current and previous contraceptive use (when applicable), and current or prior pregnancy.
Statistical Analysis
Based on previously reported literature, a BMI change of 1 and a weight change of 2 kg over time is considered clinically significant.11 Therefore, this study was sufficiently powered to provide 80% power to detect a change of (1) 1 in BMI between the 2 groups (assuming a common standard deviation of 5); (2) 2 kg in weight (assuming a standard deviation of 10). We calculated the mean and median weight and BMI change over time in our sample. Continuous variables were compared between matched cases and controls using Wilcoxon signed rank test. Binary variables were compared between groups using McNemar’s chi-squared test (of note, race was treated as a binary variable: black vs other). The primary outcomes were weight at time 1, weight change, BMI at time 1, and BMI change. We fit separate linear mixed-effects model to assess the differences between cases and controls. Other covariates in the model included weight at time of insertion, height, duration of use, race, and prior pregnancy. We also fit multiple linear model for weight and BMI among cases including weight or BMI at time of insertion, height (for weight model), age, duration of use, race, use of DMPA, and prior pregnancy as the independent variables. Data analysis was conducted using statistical software R version 3.3.0.12 This project met criteria for expedited review by the Institutional Review Board.
Results
The mean age of participants in both cases and controls was 17±2 years and almost 70% identified as non-Hispanic black (see table 1). The median BMI percentile at study onset for both cases and controls was 86, which is considered overweight. We found significant differences between the cases and controls at baseline (time 0) for insurance status, with cases being more likely to have public insurance (p<0.001). There was also a significant difference in sexual activity and use of hormonal contraceptive pills. Cases were more likely to be sexually active (p<0.001), and more likely to have previously used oral contraceptive pills (p=0.04). In addition, cases were significantly more likely than controls to have previously used DMPA and ENG. Given that DMPA users were excluded from the control group and that some cases were repeat ENG users, these differences were expected.
Table 1:
Demographics and Clinical Characteristics by Group.
| ENG users (n=197) | Control (n=197) | P-value | |
|---|---|---|---|
| Mean Age, years (N±SD) | 17 ± 2 | 17 ± 2 | NA |
| Mean Weight at Time 0, kg (N±SD) | 69 ± 19 | 69 ± 19 | 0.265 |
| Mean BMI at Time 0, kg/m2 (N±SD) | 26.3 ± 6.8 | 26.3 ± 6.8 | NA |
| Mean BMI percentile at Time 0 (N±SD) | 73 ± 29 (n=180) | 73 ± 29 (n=183) | 0.41 |
| Mean Weight at Time 1, kg (N±SD) | 72 ± 21 | 72 ± 20 | 0.764 |
| Mean BMI at Time 1, kg/m2 (N±SD) | 27.6 ± 7.6 | 27.3 ± 7.3 | 0.207 |
| Mean BMI percentile at Time 1 (N±SD) | 75 ± 28 (N=138) | 72 ± 31 (N=138) | 0.104 |
| Race | NA | ||
| Non-Hispanic White | 51 (26%) | 52 (26%) | |
| Non-Hispanic Black | 136 (69%) | 137 (70%) | |
| White Hispanic | 8 (4%) | 6 (3%) | |
| Asian | 1 (1%) | 1 (1%) | |
| Black Hispanic | 1 (1%) | 1 (1%) | |
| Previous DMPA usage (YES) | 51 (26%) | 0 (0%) | <0.001 |
| Insurance status | |||
| Private | 37 (19%) | 78 (40%) | <0.001 |
| Public | 160 (81%) | 119 (60%) | <0.001 |
| Prior Pregnancy (YES) | 19 (10%) | 12 (6%) | 0.265 |
| Prior Birth (YES) | 12 (55%) | 4 (33%) | 1 |
| Sexually Active (YES) | 153 (78%) | 73 (37%) | < 0.001 |
| Previous Oral Contraceptive Pill Use (YES) | 99 (50%) | 80 (41%) | 0.04 |
| Previous Transdermal Contraceptive Use (YES) | 6 (3%) | 4 (2%) | 0.752 |
| Previous Nuvaring® Use (YES) | 8 (4%) | 5 (3%) | 0.579 |
| Previous Progestin-Only Pill Use (YES) | 3 (2%) | 2 (1%) | 1 |
| Previous ENG Use (YES) | 7 (4%) | 0 (0%) | 0.023 |
| Previous IUD use (YES) | 4 (2%) | 4 (2%) | 1 |
| Mean Duration of ENG use, months (N±SD) | 24.5 ± 9.3 | ||
| Mean Weight change Time 0 to Time 1, kg (N±SD) | +3.6 ± 7.8 | +3.1 ± 5.9 | 0.43 |
| Mean BMI change Time 0 to Time 1 (N±SD) | +1.3 ± 2.9 | +1.0 ± 2.3 | 0.204 |
• ENG=etonogestrel subdermal implant
• SD=standard deviation
• P-values not provided for matching variables (age, race, BMI)
• Time 0=date of ESI insertion (cases); date matched within 1 month of insertion (controls)
• Time 1= last documented clinic visit or date of ESI removal (cases); date matched within 1 month of last documented clinic visit (controls)
The mean duration of ENG use was 24.5±9.3 months with 21.8% of users (43/197) discontinuing the method early (prior to 3 years). Among those with early removal, 6.3% (n=3) cited weight gain as the primary reason for removal. Overall mean (±SD) weight gain for ENG users, 3.6±7.8 kg, was not statistically different than that for controls, 3.1±5.9 kg (p=0.43). The mean BMI increase for ENG users was 1.3±2.9 compared to 1.0±2.3 for controls (p=0.204).
Based on linear models, cases had an increase in weight that was 0.55 kg more than controls (95% CI −0.84–1.95; p=0.4286) with same baseline weight, height, race, and prior pregnancy status. We utilized the same linear mixed-effects model to compare BMI between cases and controls. Cases had a 0.26 higher change in BMI as compared to controls (95% CI - 0.27–0.79; p=0.3254) with same baseline BMI, current height, race, and prior pregnancy status. These findings are summarized graphically in Figure 1. Additional sub-analyses of cases did not find any statistically significant change in weight or BMI, controlling for DMPA use. Overall regression analyses found no group differences between cases and controls, when adjusting for age, length of ENG use, race, prior DMPA use, or prior pregnancy.
Figure 1.
Box-whisker plot of weight (kg) and BMI (kg/m2) change by group. The bold solid line represents median value, and the upper and lower box margins represent the 75th and 25th quartiles respectively.
*Current = Time 1 as referenced in the text.
Discussion
This study is one of the first to examine weight and BMI changes in a large sample of exclusively adolescent and young adult women with long term use of the etonogestrel contraceptive implant. It is also the first study to look at weight and BMI changes in a younger patient population who can be expected to have some developmentally normal changes in weight and BMI. This study demonstrates that long term ENG use does not lead to significant weight change. This lack of effect persists even when adjusted for multiple potentially confounding factors, such as previous DMPA use, pregnancy, baseline weight/BMI and race.
Weight gain is a reported side effect in 5–22% of ENG users and perceived weight gain has been cited as a reason for hormonal method discontinuation.9,13 However only a small minority of patients actually discontinue ENG use due to weight changes.9,14 While this current study did not look at subjective patient assessment of weight gain, we did objectively collect information about weight and BMI in adolescent and young adult ENG users. The international multi-site randomized trial by Bahamondes in 2018 found that adult ENG users gained a mean of 3.0 kg over 36 months of use, compared to 1.1 kg in non-hormonal IUD users.10 We found a similar weight increase over time in our adolescent ENG users, with a mean increase of 3.6±7.8 kg (over a mean of 24.5±9.3 months). This weight change was in an adolescent and young adult sample that is expected to gain weight as part of normal adolescent growth and development. Since we do not have a large number of non-hormonal copper IUD users in our clinic, we used BMI and age-matched population controls for study comparison. Thus, our case-control study revealed that weight gain in ENG users was not statistically different when compared to weight gain in non-users. The previously referenced study by Bahamondes showed a 1.1 kg annual increase in weight per year over the duration of ENG device use, but this was thought to have little clinical significance and similar to normal annual adult weight trends.10 While 3.6 kg in our adolescent and young adult sample may seem clinically significant, it was not statistically significant when compared with the 3.1 kg weight gain in non-ENG users. Our study results are also consistent with the 2017 CHOICE project prospective data which found no differences in BMI, body weight or composition among 149 young adult LARC users (levonorgestrel IUD, non-hormonal copper IUD, or ENG) over 12 months of continuous use.14 Modesto et al. recently reported a 4.1 kg increase in weight in adult ENG users (mean age 30.4 years) after 12 months of continuous ENG use.15 Comparatively, we report a 3.6 kg increase in weight over a longer duration (mean 24.5 months) and in a younger population (mean age 17 years). We were unable to stratify by baseline status (underweight, normal weight, overweight, obese) which may have been helpful as Bahamondes et al. found that heavier adult women tended to gain more weight compared to their thinner counterparts.10 This may have been particularly relevant for our study as the median BMI percentile for ENG users and controls at time 0 was at the 86th percentile, which meets the criteria for overweight. Therefore, stratifying by baseline BMI and BMI percentile may be a suggested area for further study. Our study did, however, look at baseline weight as a confounding variable in weight/BMI change. No association was found in our study participants.
In general, adolescent and young adult users of ENGs are satisfied with their contraceptive method.8,16,17 Despite the anecdotal reports of weight gain in progestin only methods, only 6.3% of our study patients cited weight gain as the primary reason for removal.
This study has several strengths including a large, racially diverse study population and the inclusion of age and BMI matched controls in data analysis. We also controlled for a number of variables that may affect weight and BMI change in an adolescent and young adult population (pregnancy and previous injectable contraceptive use). We also used objective measurement of body weight at two distinct time points, not subjective reports of weight gain. However, there are also some study limitations. This retrospective study was conducted via chart review and retrospective data may be inaccurate or incomplete. There was also no standard follow-up or scheduled post-insertion appointments and therefore data points used in analysis were based on the available chart review data. Matched controls were selected at time of data collection, not at time of ENG insertion as study design did not allow for control matching at the study onset. We also did not control for diet, exercise or metabolic conditions that may affect weight. However, this is true for both ENG users and controls. There were also certain characteristics for which there was a significant difference between cases and controls. Cases were more likely to be sexually active, to have used oral contraceptive pills and to have public insurance. We cannot accurately assess or predict what role sexual activity or insurance played in weight changes or behaviors that might contribute to changes in weight. Oral contraceptive pills are thought to be weight neutral, so should not have altered the weight/BMI findings. Since weight gain is a known side effect of DMPA, controls with a previous use of this contraceptive method were excluded. Despite the fact that only cases had been previously exposed to DMPA, there was still no difference in weight gain between cases and controls. We originally hypothesized that adolescents who had used DMPA prior to ENG insertion would be at increased risk for weight/BMI change. However, we did not find this to be true. This further supports the conclusion that ENG is a weight neutral contraceptive option. Although our study does not suggest that ENG use contributes to weight gain, some of the cases were repeat ENG users; we treated each ENG insertion as a separate case. If longer duration of ENG use is a contributor to weight or BMI change (i.e., patients who underwent replacement/reinsertion of a second ENG), a separate study looking at only repeat ENG users (or duration of continuous ENG use) may be helpful.
Future studies should consider identifying risk factors that may predispose adolescent and young adult women to gain more weight, particularly looking at participants who have gained greater than 2 kg (which is commonly cited as a meaningful weight change in contraceptive research).11 Dickerson et al. found that weight gain with LARC methods was more commonly reported by overweight and obese women, as opposed to their normal weighted counterparts.8 As such, another area of future study would be prospectively studying weight change over time in those that are overweight or obese at baseline.
As the population of adolescent and young adult users of LARC increases, we have the ability and opportunity to look at long-term side effects such as weight change. Weight gain is often a major concern and deterrent for adolescent patients when choosing a hormonal contraceptive method. Although the safety, efficacy and acceptability of LARCs are well established, there has not been significant data reported on weight effects in this dynamic population, until now. This study supports the current recommendation of the etonogestrel implant as a weight neutral, first-line hormonal contraceptive option for adolescents. We believe that this study adds to the contraceptive literature by providing evidence-based counseling recommendations that the etonogestrel implant does not carry a risk of weight gain in adolescent and young adult users. Given the highly effective contraceptive efficacy of ENGs, adolescent and young adult women should be discouraged from changing to a less effective contraceptive option due to perceived weight gain. Adolescents and young adult women should continue to be counseled and reassured that ENGs are a safe and weight neutral contraceptive method.
Acknowledgments
The authors thank Li Wang, MS for her help with statistical and data analysis. This project was supported by CTSA award No. UL1TR000445 from the National Center for Advancing Translational Sciences. Its contents are solely the responsibility of the authors and do not necessarily represent official views of the National Center for Advancing Translational Sciences or the National Institutes of Health.
The authors would like to thank Li Wang, MS for her help with statistical and data analysis.
All individuals who contributed significantly to this work are listed above.
Abbreviations:
- ENG
etonogestrel subdermal implant
- BMI
body mass index
- LARC
long-acting reversible contraception
- IUD
intra-uterine device
- DMPA
depo medroxyprogesterone acetate
- SD
standard deviation
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Ott MA, Sucato GS. AAP Policy Statement Contraception for Adolescents. Pediatrics 2014;134:e1244–56. [DOI] [PubMed] [Google Scholar]
- 2.Mestad R, Secura G, Allsworth JE, Madden T, Zhao Q, Peipert JF. Acceptance of long-acting reversible contraceptive methods by adolescent participants in the Contraceptive CHOICE Project. Contraception 2011;84:493–498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lopez LM, Edelman A, Chen M, Otterness C, et al. Progestin-only contraceptives: Effects on weight. Cochrane Database Syst Rev 2013;7(7): CD008815. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Beksinska ME, Smit JA, Kleinschmidt I, et al. Prospective study of weight change in new adolescent users of DMPA, NET-EN, COCs, nonusers and discontinuers of hormonal contraception. Contraception 2010;81:30–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Depo-provera – Medroxyprogesterone acetate injection Package insert. From Pharmacia and Upjohn Company. Accessed August 18, 2016 from: http://labeling.pfizer.com/ShowLabeling.aspx?id=522#section-5.10
- 6.Dal’Ava N, Bahamondes L, Bahamondes V, et al. Body weight and composition in users of levonorgestrel-releasing uterine system. Contraception 2012;86:350–3. [DOI] [PubMed] [Google Scholar]
- 7.Lakha F, Glasier AF. Continuation rates of Implanon in the UK: data from an observational study in a clinical setting. Contraception 2006;74:287–9. [DOI] [PubMed] [Google Scholar]
- 8.Dickerson LM, Diaz VA, Jordan J, et al. Satisfaction, early removal, and side effects associated with long-acting reversible contraception. Family Medicine 2013;45(10):701–707. [PubMed] [Google Scholar]
- 9.Vickery Z, Madden T, Zhao Q, et al. Weight change at 12 months in users of three progestin-only contraceptive methods. Contraception 2013;88:503–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bahamondes L, Brache V, Moazzam Ali, et al. A multicenter randomized clinical trial of etonogestrel and levonorgestrel contraceptive implants with nonrandomized copper intrauterine device controls: effect on weight variations up to 3 years after placement. Contraception 2018;98(3):181–187. [DOI] [PubMed] [Google Scholar]
- 11.Gallo MF, Legardy-Williams J, Hylton-Kong T, et al. Association of progestin contraceptive implant and weight gain. Obstetrics & Gynecology 2016;127(3):573–576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.R Core Team (2014). R: A language and environment for statistical computing R Foundation for Statistical Computing, Vienna, Austria. [Google Scholar]
- 13.Sznajder KZ, Tomaszewski KS, Burke AE, et al. Incidence of discontinuation of long-active reversible contraception among adolescent and young adult women served by an urban primary care clinic. Journal of Pediatric and Adolescent Gynecology 2017;30(1):53–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Silva Dos Santos PN, Madden T, Omvig K, et al. Changes in body composition in women using long-acting reversible contraception. Contraception 2017;95(4):382–389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Modesto W, Dal Ava N, Monteiro I, et al. Body composition and bone mineral density in users of the etonogestrel-releasing contraceptive implant. Arch Gynecol Obstet 2015;292:1387–91. [DOI] [PubMed] [Google Scholar]
- 16.Obijuru L, Bumpus S, Auinger P, et al. Etonogestrel implants in adolescents: Experience, satisfaction, and continuation. Journal of Adolescent Health 2016;58:284–289. [DOI] [PubMed] [Google Scholar]
- 17.Berlan E, Mizraji K, Bonny AE. Twelve-month discontinuation of etonogestrel implant in an outpatient pediatric setting. Contraception 2016;94:81–86. [DOI] [PubMed] [Google Scholar]