Background
Intrauterine devices, including levonorgestrel-releasing and copper devices, are highly effective long-acting reversible contraceptives. The potential risks associated with intrauterine devices are low and include uterine perforation and device expulsion.
Objective
This study aimed to evaluate the risk of perforation and expulsion associated with levonorgestrel-releasing devices vs copper devices in clinical practice in the United States.
Study Design
The Association of Perforation and Expulsion of Intrauterine Device study was a retrospective cohort study of women aged ≤50 years with an intrauterine device insertion during 2001 to 2018 and information on intrauterine device type and patient and medical characteristics. Of note, 4 research sites with access to electronic health records contributed data for the study: 3 Kaiser Permanente–integrated healthcare systems (Northern California, Southern California, and Washington) and 1 healthcare system using data from a healthcare information exchange in Indiana (Regenstrief Institute). Perforation was classified as any extension of the device into or through the myometrium. Expulsion was classified as complete (not visible in the uterus or abdomen or patient reported) or partial (any portion in the cervix or malpositioned). We estimated the crude incidence rates and crude cumulative incidence by intrauterine device type. The risks of perforation and expulsion associated with levonorgestrel-releasing intrauterine devices vs copper intrauterine devices were estimated using Cox proportional-hazards regression with propensity score overlap weighting to adjust for confounders.
Results
Among 322,898 women included in this analysis, the incidence rates of perforation per 1000 person-years were 1.64 (95% confidence interval, 1.53–1.76) for levonorgestrel-releasing intrauterine devices and 1.27 (95% confidence interval, 1.08–1.48) for copper intrauterine devices; 1-year and 5-year crude cumulative incidence was 0.22% (95% confidence interval, 0.20–0.24) and 0.63% (95% confidence interval, 0.57–0.68) for levonorgestrel-releasing intrauterine devices and 0.16% (95% confidence interval, 0.13–0.20) and 0.55% (95% confidence interval, 0.44–0.68) for copper intrauterine devices, respectively. The incidence rates of expulsion per 1000 person-years were 13.95 (95% confidence interval, 13.63–14.28) for levonorgestrel-releasing intrauterine devices and 14.08 (95% confidence interval, 13.44–14.75) for copper intrauterine devices; 1-year and 5-year crude cumulative incidence was 2.30% (95% confidence interval, 2.24–2.36) and 4.52% (95% confidence interval, 4.40–4.65) for levonorgestrel-releasing intrauterine devices and 2.30% (95% confidence interval, 2.18–2.44) and 4.82 (95% confidence interval, 4.56–5.10) for copper intrauterine devices, respectively. Comparing levonorgestrel-releasing intrauterine devices with copper intrauterine devices, the adjusted hazard ratios were 1.49 (95% confidence intervals, 1.25–1.78) for perforation and 0.69 (95% confidence intervals, 0.65–0.73) for expulsion.
Conclusion
After adjusting for potential confounders, levonorgestrel-releasing intrauterine devices were associated with an increased risk of uterine perforation and a decreased risk of expulsion relative to copper intrauterine devices. Given that the absolute numbers of these events are low in both groups, these differences may not be clinically meaningful.
Introduction
An estimated 14% of women worldwide use intrauterine devices (IUDs). IUDs provide highly effective contraception, with rates of unintended pregnancy of <1%. Of note, 2 types of IUDs are widely used: levonorgestrel (LNG)-releasing IUDs, which are approved by the US Food and Drug Administration (FDA) for use up to 3 to 6 years, and copper IUDs, which are approved for use up to 10 years before removal or replacement.
Why was this study conducted?
It is important to understand how intrauterine device (IUD) type (levonorgestrel [LNG] releasing or copper) is associated with the risk of uterine perforation and device expulsion in US clinical practice.
Key findings
The rates of uterine perforations and device expulsions were low in both users of LNG-IUDs and copper IUDs, but perforation risk was approximately 50% higher for women using LNG-IUDs than those using copper IUDs. The risk of device expulsion was 30% lower for women using LNG-IUDs than copper IUDs.
What does this add to what is known?
In light of low absolute rates of uterine perforation and device expulsion, differences in risk by device type may not be clinically meaningful.
The risks associated with IUD use are rare and include uterine perforation and IUD expulsion. The prospective observational European Active Surveillance Study for Intrauterine Devices (EURAS-IUD) evaluated the incidence and risk of uterine perforation by IUD type, finding a borderline higher risk of perforation among LNG-IUD users than copper IUD users. , Earlier analyses of expulsion rate by IUD type have yielded varied findings. In the Contraceptive CHOICE study, the 36-month cumulative expulsion rate did not differ by IUD type, whereas a meta-analysis evaluating the rates of IUD expulsion by IUD type found a greater rate of expulsion with LNG-IUDs than with copper IUDs after vaginal delivery.
We conducted a multisite US cohort study, Association of Perforation and Expulsion of Intrauterine Devices (APEX-IUD), to evaluate risk factors associated with IUD-related uterine perforations and IUD expulsion as observed in clinical practice. Associations between postpartum timing of IUD insertion and breastfeeding with uterine perforation and IUD expulsion have been reported previously. Here, we compared the incidence and risk of uterine perforation and expulsion associated with LNG-IUDs vs copper IUDs. At the time of the study, 4 brands of LNG-IUDs (Mirena [Bayer HealthCare Pharmaceuticals Inc, Whippany, NJ], Liletta [Allergan USA Inc, Irvine, CA; Medicines360, San Francisco, CA], Skyla [Bayer HealthCare Pharmaceuticals Inc, Whippany, NJ], and Kyleena [Bayer HealthCare Pharmaceuticals Inc, Whippany, NJ]) and 1 brand of copper IUD (ParaGard [CooperSurgical, Inc, Trumbull, CT]) were approved for use in the United States.
Materials and Methods
The 4 research sites contributed data for the study: 3 integrated healthcare systems—Kaiser Permanente Northern California (KPNC), Kaiser Permanente Southern California (KPSC), and Kaiser Permanente Washington (KPWA)—and 1 healthcare system using data from a healthcare information exchange in Indiana—Regenstrief Institute (RI). The study methods and validation of algorithms to identify uterine perforation and IUD expulsion and confirm the availability of breastfeeding status at the time of IUD insertion have been described in detail previously. ,
Study cohort
The APEX-IUD population included 326,658 women aged ≤50 years at the time of IUD insertion between 2001 and 2018. Women were required to have at least 12 months of enrollment in the healthcare plans or to have had a clinical encounter within the health information exchange more than 12 months before IUD insertion. Of this population, 322,898 women had information on IUD type recorded in their electronic health record (EHR) and were included in the current analysis ( Figure 1 ). The first date for inclusion in the study was the beginning of 2001 for RI, 2007 for KPWA, 2009 for KPSC, and 2010 for KPNC; the last date for inclusion at all research study sites was April 30, 2018. Only data for a woman’s first IUD insertion in the study period were included in this analysis. All participating research study sites received approval or exemption for the conduct of this study by their respective institutional review boards. In addition, KPSC received approval from the California Health and Human Services Agency and the California Department of Public Health Center for Health Statistics and Informatics for the use of vital statistics.
Exposure and covariates
The variables were ascertained retrospectively from the research sites’ data systems using a mixture of structured data (National Drug Code; International Classification of Diseases, Ninth Revision and Tenth Revision, Clinical Modification; Healthcare Common Procedure Coding System; and Current Procedural Terminology codes) and unstructured data (clinical notes via natural language processing [NLP]). The primary exposure of interest, IUD type, was identified via drug, procedural, or diagnostic codes and NLP.
The date of IUD insertion was used as the index date. For the outcome evaluations, person-time at risk was calculated from the IUD insertion date to the first occurrence of a study outcome (uterine perforation or IUD expulsion) or censoring event. Censoring events included IUD expiration (based on FDA-approved duration of use for the IUD brand, plus 3 months to allow for the normal variability in clinical visits for removal or replacement of IUDs), removal, or reinsertion; pregnancy, hysterectomy, bilateral oophorectomy, or other sterilization procedure; disenrollment from the healthcare system (for the Kaiser Permanente sites) or date of last clinical encounter in the healthcare system (for RI); death; or end of the study period (June 30, 2018). For RI, the last clinical encounter before December 1, 2018, was considered the most recent.
Covariates at index date included research site, demographics (age and race and ethnicity), postpartum status (nonpostpartum [ie, nulliparous or >52 weeks after delivery] or postpartum timing of insertion [for women ≤52 weeks after delivery]), breastfeeding status, clinical characteristics, procedure- and provider-related characteristics, and other potential risk factors at the time of IUD insertion based on all available information during the look-back period, which extended to the earliest enrollment date (Kaiser Permanente sites) or clinical encounter (RI). The risk factors included smoking status during the past 12 months, body mass index (BMI, kg/m 2 ), reproductive history, gynecologic diagnoses (eg, dysmenorrhea and uterine fibroids), and information about the IUD insertion procedure (year and indicators of difficult insertion). Women in the nonpostpartum group were assumed not to be breastfeeding at IUD insertion. Breastfeeding status was classified as yes (last breastfeeding date ≤30 days before IUD insertion or any time after IUD insertion, up to 12 months after delivery), no (documentation of not breastfeeding at or before insertion or breastfeeding data were missing at insertion and the most recent documentation of breastfeeding [yes] was >30 days before insertion), or undetermined.
Outcomes
The outcomes of interest for this analysis were any IUD-related uterine perforation and any IUD expulsion. Uterine perforation could be complete (ie, clinical evidence of IUD in the pelvis or abdominal cavity) or partial (ie, IUD removed after being visualized as partially embedded in the myometrium on imaging or hysteroscopy or partial perforation noted by the clinician at the time of removal). IUD expulsion could be complete (ie, IUD located in the vagina, not present in the uterus or abdomen on imaging, or patient reported that the IUD was expelled or “fell out”) or partial (ie, any portion of IUD in the cervix on imaging, IUD extending through the cervix on examination, or IUD considered malpositioned on imaging and removed by the clinician). These outcomes were previously validated at each study site. Medical record abstraction and/or clinical note review was performed to validate EHR-based algorithms or confirm perforations.
Statistical analysis
Descriptive analyses for all covariates were stratified by IUD type. For categorical variables, frequencies and percentages were calculated for each level. For continuous variables, mean, standard deviation, minimum, maximum, median, and quartiles were calculated. Missing data were treated as missing, and no imputation was performed. The variables BMI and parity included a “missing” category for analyses.
Crude incidence rates were calculated as the number of IUD-related uterine perforations and IUD expulsions occurring during the person-time at risk divided by the total person-time at risk (in person-years) and were reported as point estimates (number of cases per 1000 person-years) with 95% confidence intervals (CIs). Crude cumulative incidence, the number of IUD-related uterine perforations and IUD expulsions occurring up to a time point of the total number of IUD insertions, was estimated using the Kaplan-Meier method.
Cox proportional-hazards regression models were used to estimate crude hazard ratios (HRs) and are reported as point estimates with 95% CIs. The first 5 years of person-time at risk were considered in the HR analyses. Adjusted HRs were estimated using a Cox model with propensity score overlap weighting. Propensity score models were developed separately for uterine perforation and IUD expulsion. Covariates were assessed for inclusion in propensity score models based on association with the study outcome if the crude HR was >1.11 or <0.90; additional confounders were included if they yielded a ≥10% change in effect estimate of IUD type. Details for the propensity score models and the overlap weights have been described previously and are presented in Supplemental Appendix A . The following variables were included in the propensity score models for adjustment for both perforation and expulsion outcomes: postpartum status (4 categories), breastfeeding status, menorrhagia diagnosis in the last 12 months, age (tertiles), race and ethnicity, calendar year of index date (categorical), BMI (categorical), dysmenorrhea, uterine fibroids, parity (0, >0, or missing), cesarean delivery at any time before the index date, live birth for the most recent delivery, concomitant gynecologic procedure, indicator of difficult IUD insertion, provider experience (quartiles of IUD insertions performed in the past year), and research site. For the propensity score model for perforation, additional variables included recent smoking, duration of look-back period (quartiles), and cesarean delivery for the most recent delivery. To further adjust for postpartum categories after a lack of balance in initial propensity score models, an interaction of postpartum timing (4 categories) by site was added for the expulsion outcome.
To assess the balance between IUD-type groups, absolute standardized differences were calculated before and after weighting. Groups were considered balanced if the standardized difference was <0.20 (generally considered small).
All analyses were performed using the SAS software (version 9.3 or higher; SAS Institute, Cary, NC).
Role of the funding source
Authors affiliated with the study sponsor (4 of 27) participated in designing the study, interpreting the data, writing the report, and deciding to submit the article for publication.
Results
Cohort characteristics
The cohort for this analysis included 322,898 women: 259,234 (80.3%) with LNG-IUDs and 63,664 (19.7%) with copper IUDs. The median duration of continuous enrollment after IUD insertion was 2.4 years (range, 0–16.7).
The average age of copper IUD users was slightly lower than that of LNG-IUD users (31.2 vs 32.2 years), with a lower proportion of users aged ≥37 years (24.6% vs 31.9%) ( Table 1 ). Proportionally fewer LNG-IUD users than copper IUD users were breastfeeding at IUD insertion (18.7% vs 24.1%), and proportionally fewer LNG-IUD users than copper IUD users were postpartum (28.7% vs 35.4%). Proportionally more LNG-IUD users than copper IUD users had a BMI indicating obesity (34.4% vs 27.2%) and had a diagnosis of menorrhagia in the 12 months before IUD insertion (12.1% vs 1.3%) ( Table 1 ). Proportionally more LNG-IUD users than copper IUD users had a diagnosis of dysmenorrhea in the year before IUD insertion (3.8% vs 1.6%) and had a history of uterine fibroids before IUD insertion (6.0% vs 2.5%). After propensity score weighting, all variables included in the analyses of uterine perforation and IUD expulsion had satisfactory balance (absolute standardized difference of <0.2) (data not shown).
| Characteristics | IUD type | Unweighted absolute standardized differences a | |
|---|---|---|---|
| LNG-IUD (n=259,234) | Copper IUD (n=63,664) | ||
| Person-years at risk | 507,151.2 | 127,587.0 | |
| Age (y), mean (SD) | 32.2 (8.5) | 31.2 (7.4) | 0.121 |
| Age categories (y), n (%) | |||
| ≤28 | 94,007 (36.3) | 23,929 (37.6) | 0.027 |
| 29–36 | 82,637 (31.9) | 24,083 (37.8) | 0.125 |
| 37–50 | 82,590 (31.9) | 15,652 (24.6) | 0.162 |
| Race and ethnicity, n (%) | |||
| Asian or Pacific Islander | 29,574 (11.4) | 8995 (14.1) | 0.082 |
| Hispanic Black | 572 (0.2) | 117 (0.2) | 0.008 |
| Hispanic Other | 44,261 (17.1) | 11,705 (18.4) | 0.034 |
| Hispanic White | 32,736 (12.6) | 9543 (15.0) | 0.068 |
| Non-Hispanic Black | 23,750 (9.2) | 4211 (6.6) | 0.095 |
| Non-Hispanic White | 110,479 (42.6) | 24,423 (38.4) | 0.087 |
| Other or multiple | 12,862 (5.0) | 3367 (5.3) | 0.015 |
| Unknown | 5000 (1.9) | 1303 (2.0) | 0.008 |
| Body mass index (kg/m 2 ), n (%) | |||
| Underweight | 2807 (1.1) | 845 (1.3) | 0.022 |
| Normal weight | 87,499 (33.8) | 25,045 (39.3) | 0.116 |
| Overweight | 75,421 (29.1) | 19,817 (31.1) | 0.044 |
| Obese | 89,069 (34.4) | 17,344 (27.2) | 0.155 |
| Missing | 4438 (1.7) | 613 (1.0) | 0.065 |
| Breastfeeding status, n (%) | |||
| Yes | 48,447 (18.7) | 15,330 (24.1) | 0.132 |
| No | 23,754 (9.2) | 6674 (10.5) | 0.044 |
| Undetermined or no delivery in the past year | 187,033 (72.1) | 41,660 (65.4) | 0.145 |
| Recent smoker, n (%) | 26,502 (10.2) | 5559 (8.7) | 0.051 |
| Postpartum status (wk), n (%) | |||
| ≤6 | 15,631 (6.0) | 4228 (6.6) | 0.025 |
| >6 to ≤14 | 42,760 (16.5) | 12,934 (20.3) | 0.099 |
| >14 to ≤52 | 16,110 (6.2) | 5379 (8.4) | 0.086 |
| Nonpostpartum (>52 wk or no delivery) | 184,733 (71.3) | 41,123 (64.6) | 0.143 |
| Parity, n (%) | |||
| 0 | 48,399 (18.7) | 12,960 (20.4) | 0.043 |
| >0 | 180,134 (69.5) | 44,247 (69.5) | 0.000 |
| Missing | 30,701 (11.8) | 6457 (10.1) | 0.054 |
| Menorrhagia in 12 mo before IUD insertion, n (%) | 31,291 (12.1) | 834 (1.3) | 0.441 |
| Dysmenorrhea diagnosis in 12 mo before IUD insertion, n (%) | 9724 (3.8) | 995 (1.6) | 0.136 |
| History of uterine fibroids before IUD insertion, n (%) | 15,553 (6.0) | 1621 (2.5) | 0.171 |
| Any difficult insertion, n (%) | 24,666 (9.5) | 4648 (7.3) | 0.080 |
| Number of IUD insertions performed by provider, mean (SD) | 44.7 (36.1) | 43.5 (31.8) | 0.035 |
| Concomitant gynecologic procedure, n (%) | 23,104 (8.9) | 2662 (4.2) | 0.192 |
| Calendar year of IUD insertion, n (%) | |||
| 2001–2009 | 12,233 (4.7) | 3516 (5.5) | 0.036 |
| 2010 | 24,346 (9.4) | 6891 (10.8) | 0.048 |
| 2011 | 25,918 (10.0) | 6450 (10.1) | 0.004 |
| 2012 | 28,878 (11.1) | 7378 (11.6) | 0.014 |
| 2013 | 27,243 (10.5) | 6855 (10.8) | 0.008 |
| 2014 | 27,342 (10.5) | 6404 (10.1) | 0.016 |
| 2015 | 29,966 (11.6) | 7256 (11.4) | 0.005 |
| 2016 | 33,202 (12.8) | 7678 (12.1) | 0.023 |
| 2017 | 37,593 (14.5) | 8406 (13.2) | 0.038 |
| 2018 | 12,513 (4.8) | 2830 (4.4) | 0.018 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
