Unintended pregnancy remains an important public health problem. Approximately 3 million unplanned or mistimed pregnancies occur each year in the United States.1 Despite a decrease worldwide,2 unintended pregnancy rates remain unchanged in the United States1,3 Increasing access to effective contraception may prevent the negative health and economic consequences associated with unplanned pregnancy.4
Long-acting reversible contraceptive (LARC) methods, including the intrauterine device (IUD) and subdermal implant, offer women the most effective protection against pregnancy. Adolescents and women who use oral contraceptive pills, the contraceptive patch, or vaginal ring have a 20-fold increased risk of failure rate compared with women who use LARC methods.5 Results from the Contraceptive CHOICE Project have shown that LARC methods have higher rates of continuation and satisfaction rates at 12 and 24 months compared with short-acting reversible contraception and that continuation is equally high among adolescent females.6–8 Despite these high rates of effectiveness and continuation, fewer than 10% of women in the United States currently use an IUD or implant compared with many European countries with higher rates of uptake.9,10
The American College of Obstetricians and Gynecologists has recommended LARC methods as first-line contraceptive options for both adults and adolescents.11,12 However, concerns among health care providers about rapid discontinuation of LARC methods, especially among adolescents, may discourage use.13 Other studies have demonstrated that contraceptive dissatisfaction contributes to early discontinuation.14,15 Rates of LARC discontinuation at 6 months vary between 6% and 17% with cramping and changes in bleeding patterns commonly reported as reasons for removal of the IUD and implant.16–18 Our objective in this analysis was to measure 6-month discontinuation rates among users of the levonorgestrel intrauterine system, copper IUD, and etonogestrel implant and to investigate factors associated with early discontinuation of LARC methods.
MATERIALS AND METHODS
The methods of CHOICE have been previously described in detail19; therefore, we present a brief review and detailed description of the procedures relevant to this analysis. CHOICE is a prospective cohort study of 9,256 adolescents and women who were provided with no-cost reversible contraception. The primary goal of the project was to promote the use of LARC methods and decrease the rate of unintended pregnancy. CHOICE was a convenience sample of adolescents and women living in the St. Louis area. Eligible participants were between 14 and 45 years of age, residents of St. Louis City or County, at risk for pregnancy and did not desire pregnancy in the next year, and willing to start a new contraceptive method. Participants completed follow-up surveys by telephone at 3 and 6 months and then every 6 months for 3 years (first 5,090 participants) or 2 years (remainder of the cohort). We enrolled participants at our university clinical research site and 13 partner community clinics. Participants who enrolled at the university research site were provided with standardized, comprehensive contraceptive counseling.20 Participants who enrolled at partner clinics underwent the usual counseling for that site. All participants were offered screening for Neisseria gonorrhoeae and Chlamydia trachomatis infection at the time of enrollment with self-collected vaginal swabs. The study was approved by the Washington University Human Research Protection Office and written informed consent was obtained from participants before enrollment.
For inclusion in this analysis, participants must have received a LARC method at or within 3 months of enrollment. We obtained information about LARC method discontinuation from survey data and from our ancillary removal log (a separately maintained record of any participant who came to the university clinical research site for a removal, called to report a removal, or reported a removal during a follow-up telephone survey). Those who experienced an expulsion of an IUD were offered a replacement device. Participants who had a second IUD inserted within 28 days of the expulsion were defined as continuers in this analysis. To determine reasons for removal, two research assistants extracted and coded data regarding the reason for removal from the removal log and participant charts (D.S.G. and T.C.). Up to three reasons were coded for each participant. Codes were reviewed by the principal investigator (T.M.) and any discrepancies were resolved through discussion, chart review, and consensus. We then calculated frequencies for the primary reasons for requesting removal for each discontinuer.
Our analysis compared the baseline characteristics between 6-month continuers and discontinuers. Percentages and frequencies were calculated to compare early discontinuation rates among LARC users. χ2s were used to compare categorical variables. Continuation at 6 months was evaluated between continuers and discontinuers for each LARC method. We investigated factors associated with discontinuation before 6 months using crude and adjusted logistic regression. We calculated odds ratios (ORs) and 95% confidence intervals (CIs). Any covariates that were significant in the crude analysis were included in the adjusted model. We planned a priori to include age in the final model because we hypothesized that adolescents and young women (aged 14–19 years) may have higher rates of early discontinuation of LARC methods. Interaction terms were used to test for significant relationships between race and age, low socioeconomic status, and history of a sexually transmitted infection (STI). None of the interactions were found to be statistically significant and, therefore, were not included in the model. Statistical analyses were performed using SAS 9.2 software, and the significance level (alpha) was set at .05.
Among the 9,256 adolescents and women in the CHOICE cohort, there were 7,074 participants who chose an IUD or an implant at baseline. Of these, 760 (8%) participants did not receive their baseline chosen method within the first 3 months and 147 (2%) started their method before enrollment leaving 6,167 participants eligible for this analysis. The participant flowchart is shown in Figure 1. There were 239 (3%) participants without any data about method use at 6 months and were considered lost to follow-up. These participants were more likely to have chosen the implant at baseline (34% compared with 23%, P<.01), be of Hispanic ethnicity (8% compared with 5%, P=.04), report low socioeconomic status (70% compared with 59%, P<.01), have higher parity (21% compared with 12%, P<.01), and have reported a prior unintended pregnancy (75% compared with 67%, P=.01) compared with those who followed up. They were also less likely to have been enrolled at the university clinical research site (46% compared with 68%, P<.01), have a college education (9% compared with 23%, P<.01), and to be single (49% compared with 58%, P=.03). Participants lost to follow-up were equally likely to be aged 14–19 years as those who followed up (14% for both groups). Therefore, a total of 5,928 participants met our inclusion criteria and had follow-up data available. Of these, 3,610 (61%) received the levonorgestrel intrauterine system, 952 (16%) chose the copper IUD, and 1,366 (23%) received the implant at or within 3 months of baseline. There were 5,495 (93%) participants who were still using their baseline LARC method at 6 months and were considered continuers. The remaining 433 (7%) had discontinued their baseline method within the first 6 months of use. When stratified by age, there was no significant differences in early discontinuation by LARC method between adolescents and young women (age 14–19 years) and older women (8.5% compared with 7.1%, respectively, P=.72).
Table 1 compares the baseline characteristics between the continuers and discontinuers. Our overall sample was racially and ethnically diverse; 49% were black, and approximately 5% were of Hispanic ethnicity. There were 835 (14%) adolescents and young women aged 14–19 years in our sample. A history of a prior unintended pregnancy was common, reported by 67% of the sample. Characteristics among continuers and discontinuers were similar except that discontinuers were less likely to be married (29% compared with 36%, P=.01) and more likely to be black (55% compared with 48%, P=.04), have low socioeconomic status (63% compared with 58%, P=.04), and to have reported a past STI (35% compared with 30%, P=.02). Discontinuation rates at 6 months were similar across the three LARC methods; 7.3% of levonorgestrel intrauterine system users, 8.0% of copper IUD users, and 6.9% of implant users (P=.60).
Table 2 presents the results of the crude and adjusted logistic regression investigating participant characteristics associated with early discontinuation of any LARC method. In the crude model, we found that black participants were 29% more likely to discontinue their method early compared with white participants (OR 1.29, CI 1.05–1.59). Marital status, socioeconomic status, and history of STI were also associated with a slight increase in discontinuation. Therefore, we present both crude and adjusted ORs from our analyses. After adjusting for age, race, marital status, low socioeconomic status, and past STI, we found that single (adjusted OR 1.26, 95% CI 1.01–1.59) and divorced, separated, and widowed (adjusted OR 1.62, 95% CI 1.11–2.37) participants were more likely to discontinue their LARC method by 6 months. Young age (14–19 years) was not associated with early discontinuation in either the crude or adjusted model (adjusted OR 1.16, 95% CI 0.87–1.54).
Table 3 presents the most commonly reported reason for requesting removal of each of the LARC methods. Participants could report up to three reasons for discontinuation; however, we included only the primary or first reason given by each person for discontinuation during the follow-up surveys, telephone contact, or clinic visits because we assumed women would list their most bothersome complaint first. The most common reason for requesting removal among IUD users was pain or cramping: 41 (28%) levonorgestrel intrauterine system discontinuers and 18 (35%) copper IUD discontinuers (P=.38). Complaints of heavy bleeding were most common among copper IUD discontinuers (17%). Removal for irregular or frequent bleeding was cited across all three LARC methods: 9% of levonorgestrel intrauterine system discontinuers, 19% of copper IUD discontinuers, and 53% of implant discontinuers (P for IUD comparison=.08; P for IUD and implant comparison <.001). Only 3% of participants requested removal of their LARC method at or before 6 months as a result of desired pregnancy.
Our study found low rates of discontinuation by 6 months across the three LARC methods, which do not differ among adolescents and young women compared with older women. The only characteristic associated with early discontinuation in our sample was being unmarried, although the effect sizes were relatively small. Unmarried participants may be less likely to be in stable relationships or sexually active and thus more likely to discontinue their contraceptive method as compared with married or cohabitating participants. There were no other significant associations between early discontinuation and most demographic or reproductive characteristics.
Previously published analyses from the CHOICE cohort show high rates of LARC continuation at 12 and 24 months. At 12 months, continuation rates were 88% for the levonorgestrel intrauterine system, 84% for the copper IUD, and 83% for the implant.6 At 24 months, continuation rates were 79% for the levonorgestrel intrauterine system, 77% for the copper IUD, and 69% for the implant.7 The results of this analysis together with the previously described 12- and 24-month continuation rates suggest that discontinuation happens steadily over time rather than a large proportion discontinuing early in use. Although there are limited data about continuation of LARC at 6 months, other studies have found similar or slightly lower rates of continuation. A study of 283 copper IUD users found that 11% had discontinued by 6 months, and younger age was not associated with discontinuation.16 A study of 439 IUD and implants users found 89% and 83%, respectively, were still using their method at 6 months.17 A retrospective study of 976 implant users found that 94% were still using the implant at 6 months, and younger age was not associated with continuation.18
Among the IUD discontinuers in our study, cramping was the most common reason for requesting removal by 6 months. Although irregular bleeding has been a common reason reported for levonorgestrel intrauterine system removal in other studies,17,21 only 9% of levonorgestrel intrauterine system discontinuers reported bleeding disturbances as the reason for discontinuation. Our study did, however, identify irregular vaginal bleeding as the main concern among implant discontinuers. All CHOICE participants enrolled at the university clinical research site underwent a standardized contraceptive counseling session focused on evidence-based information.20 Participants were counseled about the possibility of irregular bleeding with the levonorgestrel intrauterine system and the implant. Participants were encouraged to call study staff or make an appointment to address method complaints, and staff encouraged continuation by providing reassurance about the common side effects. Clinicians offered medical management for bleeding irregularities such as nonsteroidal anti-inflammatory drugs or oral contraceptive pills to users with persistent complaints.22 , 23 The majority of bleeding irregularities with progestin-only contraception typically occurs within the first 6 months of method initiation.22 Therefore, 6-month discontinuers may reflect the population with intolerable bleeding side effects.
Our study had a number of strengths, which include a relatively large, racially and socioeconomically diverse cohort of reproductive-aged adolescents and women using LARC methods, including more than 800 participants aged 14–19 years. We systematically collected data regarding discontinuation at 3 and 6 months and used multiple data sources to code discontinuation and reasons for discontinuation. There are also several limitations to our study. Because this was a secondary analysis, we did not plan to enroll an adequate sample of adolescents to specifically stratify continuation by age. However, a post hoc power calculation shows that we had greater than 90% power to detect a statistically significant difference between the observed discontinuation rate of 7% among older women and a twofold increased risk of discontinuation in participants younger than 20 years of age. There were differences in baseline characteristics between those who completed follow-up surveys and those lost to follow-up. However, our overall loss to follow-up was very low at 3% so this likely has a minimal effect on our findings. In addition, data about method continuation was self-reported and could be subject to misclassification bias. Because CHOICE was restricted to a single geographical area, it is possible that our findings are not generalizable to other populations or the entire United States. However, our large and diverse cohort increases the likelihood that these findings may be generalizable to other populations of adolescents and women seeking contraception.
Long-acting reversible contraceptive methods have a number of potential advantages for adolescent females.12 Many adolescents desire to delay pregnancy for a number of years; therefore, the long-lasting action of IUDs and implant is beneficial. In addition, LARC methods do not require the user to be adherent, which may be especially advantageous for adolescents. Lastly, adolescents have higher failure rates with shorter-acting methods such as oral contraceptive pills, the contraceptive patch, or the vaginal ring. However, LARC methods are equally effective in adolescents and older women.5 Our study findings show that rates of early discontinuation are low among both adolescents and older women. As a result of the numerous benefits of LARC , IUDs and the implant should be considered as first-line contraceptive options among all adolescents and women to reduce unintended pregnancy.
1. Finer LB, Zolna MR. Unintended pregnancy in the United States: incidence and disparities, 2006. Contraception 2011;84:478–85.
2. Singh S, Sedgh G, Hussain R. Unintended pregnancy: worldwide levels, trends, and outcomes. Stud Fam Plann 2010;41:241–50.
3. Finer LB, Henshaw SK. Disparities in rates of unintended pregnancy in the United States, 1994 and 2001. Perspect Sex Reprod Health 2006;38:90–6.
4. Gipson JD, Koenig MA, Hindin MJ. The effects of unintended pregnancy on infant, child, and parental health: a review of the literature. Stud Fam Plann 2008;39:18–38.
5. Winner B, Peipert JF, Zhao Q, Buckel C, Madden T, Allsworth JE, et al.. Effectiveness of long-acting reversible contraception. N Engl J Med 2012;366:1998–2007.
6. Peipert JF, Zhao Q, Allsworth JE, Petrosky E, Madden T, Eisenberg D, et al.. Continuation and satisfaction of reversible contraception. Obstet Gynecol 2011;117:1105–13.
7. Rosenstock JR, Peipert JF, Madden T, Zhao Q, Secura GM. Continuation of reversible contraception in teenagers and young women. Obstet Gynecol 2012;120:1298–305.
8. O'Neil-Callahan M, Peipert JF, Zhao Q, Madden T, Secura GM. Twenty-four–month continuation of reversible contraception. Obstet Gynecol 2013;122:1083–91.
9. Sonfield A. Popularity disparity: attitudes about the IUD in Europe and the United States. Guttmacher Policy Rev 2007;10:19–4.
10. Finer LB, Jerman J, Kavanaugh ML. Changes in use of long-acting contraceptive methods in the United States, 2007–2009. Fertil Steril 2012;98:893–7.
11. Increasing use of contraceptive implants and intrauterine devices to reduce unintended pregnancy. ACOG Committee Opinion No. 450. American College of Obstetricians and Gynecologists. Obstet Gynecol 2009;114:1434–8.
12. Adolescents and long-acting reversible contraception: implants and intrauterine devices. Committee Opinion No. 539. American College of Obstetricians and Gynecologists. Obstet Gynecol 2012;120:983–8.
13. Kavanaugh ML, Frohwirth L, Jerman J, Popkin R, Ethier K. Long-acting reversible contraception for adolescents and young adults: patient and provider perspectives. J Pediatr Adolesc Gynecol 2013;26:86–95.
14. Moreau C, Cleland K, Trussell J. Contraceptive discontinuation attributed to method dissatisfaction in the United States. Contraception 2007;76:267–72.
15. Behringer T, Reeves MF, Rossiter B, Chen BA, Schwarz EB. Duration of use of a levonorgestrel IUS amongst nulliparous and adolescent women. Contraception 2011;84:e5–e10.
16. Garbers S, Haines-Stephan J, Lipton Y, Meserve A, Spieler L, Chiasson MA. Continuation of copper-containing intrauterine devices at 6 months. Contraception 2013;87:101–6.
17. Wong RC, Bell RJ, Thunuguntla K, McNamee K, Vollenhoven B. Implanon users are less likely to be satisfied with their contraception after 6 months than IUD users. Contraception 2009;80:452–6.
18. Harvey C, Seib C, Lucke J. Continuation rates and reasons for removal among Implanon users accessing two family planning clinics in Queensland, Australia. Contraception 2009;80:527–32.
19. Secura GM, Allsworth JE, Madden T, Mullersman JL, Peipert JF. The Contraceptive CHOICE Project: reducing barriers to long-acting reversible contraception. Am J Obstet Gynecol 2010;203:e115.e1–7.
20. Madden T, Mullersman JL, Omvig KJ, Secura GM, Peipert JF. Structured contraceptive counseling provided by the Contraceptive CHOICE Project. Contraception 2013;88:243–9.
21. Jensen JT, Nelson AL, Costales AC. Subject and clinician experience with the levonorgestrel-releasing intrauterine system. Contraception 2008;77:22–9.
22. Abdel-Aleem H, d'Arcangues C, Vogelsong K, Gülmezoglu AM. Treatment of vaginal bleeding irregularities induced by progestin only contraceptives. The Cochrane Database of Systematic Reviews 2007, Issue 2. Art. No.: CD003449. DOI: 10.1002/14651858.CD003449.pub2.
23. Madden T, Proehl S, Allsworth JE, Secura GM, Peipert JF. Naproxen or estradiol for bleeding and spotting with the levonorgestrel intrauterine system: a randomized controlled trial. Am J Obstet Gynecol 2012;206:129.e1–8.