Abstract
Background:
Intrauterine devices (IUDs) are highly effective, yet acceptability and continuation differ by user and device characteristics. We evaluated the efficacy of the NTCu380 Mini, a copper 380 mm2 IUD available in Europe and Canada, and compared safety and continuation to the TCu380A IUD, the copper 380 mm2 IUD available in the United States.
Methods:
Participants aged 16–40 were randomly assigned 4:1 to the NTCu380 Mini or TCu380A and assessed over 37 months. The primary outcome was NTCu380 Mini efficacy measured by Pearl Index, (defined as the number of pregnancies per 100 woman-years) in participants aged ≤35 years. Additional outcomes included NTCu380 Mini time-to-pregnancy by survival analysis, and comparative adverse events (AEs) and continuation rates of NTCu380 Mini to TCu380A.
Results:
We randomly assigned 887 participants to NTCu380 Mini (744 [83.9%] nulliparous) and 218 to TCu380A (183 [83.9%] nulliparous) of which 875 (98.6%) and 213 (97.7%), respectively, had successful placement. The NTCu380 Mini IUD 3-year cumulative Pearl Index was 1.86 (95% CI 1.20–2.74) pregnancies/100 women-years; cumulative probability of pregnancy through year 3 was 4.8% (95% CI 2.8–6.9%). Serious AEs occurred in 31 (3.5%) NTCu380 Mini and 4 (1.9%) TCu380A users (p=0.28). Discontinuation occurred in 449 (51.3%) NTCu380 Mini and 122 (57.3%) TCu380A users (p=0.07). Adverse events led to discontinuation less frequently among NTCu380 Mini (152 [20.8%]) vs. TCu380A users (57 [33.2%]), p=0.001, particularly bleeding or pelvic pain AEs (103 [14.5%] vs. 46 [27.3%], respectively, p<0.001).
Conclusions:
In this predominantly nulliparous population, NTCu380 Mini IUD had an approximate 5% pregnancy rate over 3 years with fewer side effects leading to discontinuation compared with TCu380A users.
Introduction
Intrauterine devices (IUDs) are known to provide highly effective, safe, convenient, reversible, and cost-effective contraception. In the United States (U.S.), several levonorgestrel-releasing IUDs with Nova-T frames are available, but ParaGard (TCu380A, Cooper Surgical, Trumbull, CT), an intrauterine device with a Tatum-T frame that entered the market in 1988, is the only U.S. Food and Drug Administration (FDA) approved non-hormonal IUD. In contrast, many non-hormonal copper IUDs are approved and in clinical use outside of the U.S.
People seeking long-acting, reversible contraception may choose a copper IUD because it is hormone-free, allows for cyclic menstrual bleeding, and is long-acting with a quick return to fertility upon removal. However, increased menstrual bleeding and cramping experienced by some users can result in diminished satisfaction.1 Up to 20% of users discontinue use within the first year and nulliparity is associated with a higher incidence of TCu380A-associated side effects and removals.2,3 On average, nulliparous individuals have smaller uterine cavities (mean cavity volume of 57 cm3) compared to parous individuals (mean volumes of 82 cm3 to 105 cm2)4. The TCu380A used in the U.S. is the largest T-shaped copper IUD available globally, measuring 32 mm wide and 36 mm long. A smaller copper IUD might be better tolerated by nulliparous individuals and improve available contraceptive options.
Given the limited non-hormonal IUD options in the U.S. and the value of access to a range of contraceptive methods for the benefit of individuals and the public health, we sought to assess the efficacy of a smaller copper IUD with a Nova-T frame which is currently approved for nulliparous and parous users in some European countries and Canada (Mona Lisa® NTCu380 Mini, Mona Lisa, N.V., Belgium) in a predominately nulliparous U.S. population, and to compare the adverse effects and discontinuation rates of the NTCu380 Mini with the TCu380A in a 3-year randomized trial.
Methods
Trial Design
This multicenter, participant-masked, phase 3 randomized clinical trial was conducted at 16 U.S. sites; the trial design has been described previously in a planned interim report that presented 1-year continuation outcomes in nulliparous participants.5 We evaluated the contraceptive efficacy of the NTCu380 Mini6 and compared safety and secondary endpoints of discontinuation and product satisfaction between the NTCu380 Mini and the TCu380A.7 A central IRB and center-specific IRBs, as required, approved the trial; all participants provided written informed consent before any trial procedures occurred. The trial is registered on Clinicaltrials.gov (NCT03124160).
Participants
We recruited healthy individuals 16–40 years of age at risk of pregnancy, who were seeking contraception. While participants ≤35 years of age were enrolled to assess the primary efficacy outcome of pregnancy, we also offered enrollment to individuals aged 36–40 years and included those participants in the safety analyses. In accordance with standard approaches in contraceptive clinical trials8, to be at risk of pregnancy, eligibility criteria included metrics of fertility including age, regular menstrual cycles between 21 and 35 days, engagement in regular heterosexual vaginal intercourse (at least once per month), an adequate washout of other contraceptives, and no use other contraceptives or pregnancy prevention during the trial. We excluded those with abnormal cervical cytology requiring treatment, known HIV infection, known infertility or previous permanent contraception, contraindications to copper IUD use, recent injectable contraceptive use, or intention to become pregnant within 37 months after enrollment. Uterine sounding measurement of uterine length was not a trial entry criterion but was done as part of standard-of-care clinical procedures prior to IUD insertion; those participants with cavity length of < 55 mm were excluded and insertion was not attempted (Figure 1). There was no upper limit on uterine sounding measurement. We planned to recruit approximately 80% nulliparous participants, with parous defined as a prior delivery at or after 24 weeks of pregnancy.
Figure 1. Analysis population flow diagram for participants.
Treated population refers to participants who had IUD insertion attempted. User population refers to participants who had IUD successfully inserted. 1No IUD insertion attempt was made due to inability to sound the uterus prior to attempting the IUD insertion, the uterus did not sound to a minimum of 55 mm, medical conditions, or other reasons. 2Two participants were randomized to the NTCu380 Mini group but received TCu380A, and 1 participant was randomized to the TCu380A group but received NTCu380 Mini; this resulted in a net loss of one for NTCu380 Mini, leaving 876 in treated population, and a net gain of one for TCu380A, leaving 214 in the treated population. IUD=intrauterine device.
Interventions/Exposures
The NTCu380 Mini measures 24 mm (horizontal arms) × 30 mm (vertical stem) with a copper wire on the stem, providing 380 mm2 surface area of copper exposure (Figure S1).6 The TCu380A measures 32 mm (horizontal arms) × 36 mm (vertical stem) with copper on the arms and stem, totaling 380 mm2 surface area of copper exposure.7 The insertion tube external diameters for the NTCu380 Mini and TCu380A are 3.7 mm and 4.5 mm, respectively.6,7 Clinician-investigators were highly experienced in placement of the TCu380A and were trained on NTCu380 Mini placement prior to the trial. After successful screening, each participant underwent informed consent, trial enrollment, and IUD placement, with follow-up clinic visits at 6 weeks, then at 3, 6, 12, 24, and 37 months. Participants completed pain assessments using visual analog scales (VAS ranging from 0 for “no pain” to 10 for “worst pain imaginable”) at baseline after speculum placement and then immediately and ten minutes after IUD placement. Site investigators verified IUD presence by visual (speculum) or digital examination of the IUD string and pelvic ultrasound as needed; participants completed urine pregnancy testing at each visit. Telephone calls to collect additional data were scheduled at 9, 18, and 30 months. At the 37-month exit visit, participants with a NTCu380 Mini underwent removal while those randomly assigned to TCu380A could have removal or elect to continue use under the care of their clinical providers. Participants used home pregnancy tests 17 days after the exit visit or whenever they suspected pregnancy and reported the results to their respective sites via telephone.
Endpoints
The trial was designed primarily to measure the contraceptive efficacy of the NTCu380 Mini. Per guidance from the US Food and Drug Administration (FDA), the primary measure of contraceptive efficacy of a new method is the Pearl Index (defined as the number of pregnancies per 100 woman-years) in participants ≤35 years of age at time of enrollment, because the likelihood of pregnancy decreases with advancing age.
For the primary endpoint of pregnancy, investigators estimated date of conception based on menstrual history, ultrasonographic measurements of the pregnancy, and other clinical data. We calculated the Pearl Index overall and by year of use, including only cycles in which vaginal intercourse occurred and no other method was used, or when an on-treatment pregnancy occurred. We did not prespecify calculating the Pearl Index for the TCu380A, as it is already marketed in the U.S. and this trial was not designed to compare efficacy of the two IUDs. We additionally computed the cumulative probability of pregnancy using Kaplan-Meier survival analysis. We censored any participant without an on-treatment pregnancy from the pregnancy survival analysis at the time of the last known date of using the trial device.
We conducted two post-hoc sensitivity analyses. First, during the course of the trial, investigators observed a possible association between menstrual cup use and IUD expulsion.9 In response, 1 year after beginning enrollment, we advised new and already enrolled participants to stop using menstrual cups and to break the vaginal seal prior to cup removal if a cup was used. We also modified the protocol to collect information prospectively on menstrual hygiene product use and conducted a sensitivity analysis to evaluate the effect of menstrual cup use. Second, to more closely evaluate the pregnancy rate among those for whom the NTCu380 Mini is generally intended (those with uterine cavity length of ≤8 cm, according to the European label 6), we calculated the Pearl Index among those with uterine cavity length of ≤8 cm and compared it with those with cavities >8 cm.
Safety and acceptability endpoints included comparison of the NTCu380 Mini and TCu380A, adverse events (AEs) including self-reported pelvic pain and bleeding, IUD continuation, IUD expulsion (complete and partial), and accidental IUD removal. We categorized an expulsion as complete if the IUD was completely outside of the uterus and cervix, as evidenced by visualization of the IUD outside the uterus or complete absence of the IUD on ultrasound examination with no evidence of an intra-abdominal IUD by X-ray. We defined partial expulsion as visualization or palpation of any portion of an IUD (other than strings) at the external cervical os, or ultrasound identification of any part of the IUD in the cervical canal, regardless of symptoms. We defined accidental self-removal as any expulsion (complete or partial) that occurred as a result accidentally pulling on the IUD string.
We also evaluated participant reports of changes in pain from baseline with IUD insertion as measured by Visual Analogue Scale (VAS), and bleeding outcomes from participant daily diaries during the first year including the proportion of cycles with heavy bleeding, prolonged bleeding episodes (longer than 7 days),10 any bleeding/spotting, or bleeding in more than 1/3 of days in a cycle.
We included all participants (aged 16–40) for analysis of adverse events. Routine laboratory monitoring for safety was not performed as the trial involved non-hormonal products. AEs and serious adverse events (SAEs) were collected at each trial visit.
Statistical Analysis
For our sample size calculation, we estimated that 800 NTCu380 Mini users ≤35 years of age at enrollment would provide about 17,700 cycles over 3 years and enable us to verify that the upper bound of the exact 95% confidence interval (CI) of the Pearl Index would be ≤1.0 if the Pearl Index did not exceed 0.4 per 100 women-years (or 6 pregnancies over the trial period). We planned to randomly assign eligible participants in a 4:1 ratio to the NTCu380 Mini and TCu380A resulting in approximately 800 NTCu380Mini and 200 TCu380A participants who were less than or equal to 35 years of age. We planned to recruit an additional 100 participants (10% of the primary efficacy sample size) age 36–40 years for the non-efficacy outcome assessments only. We used a SAS-based computer-generated randomization scheme stratified by parity and stopped enrollment for parous participants at approximately 20% (n=220) of the entire trial population. We defined the user population as those with successful IUD placement.
Baseline characteristics were presented by treatment group. Categorical variables, and continuous variables that have been categorized at discrete levels, were summarized by frequencies and percentages. Continuous variables were summarized by means, standard deviations, medians, interquartile ranges, minima and maxima. We compared the secondary and safety outcomes between IUD groups using two-sided Fisher’s exact or Cochran-Mantel-Haenszel tests. Changes in pain with and shortly after insertion from baseline were described by summary statistics and were compared using Mann-Whitney tests or analysis of variance. We estimated discontinuation probability and 95% CI for each group using the Kaplan-Meier method with the long-rank test for group comparison. We note that all secondary analyses in the trial were considered exploratory and hypothesis generating and were not a priori powered to detect clinically meaningful effect sizes. P-values are reported for safety and adverse event outcomes only (including discontinuation); values less than 0.05 were considered statistically significant and were not adjusted for multiple comparisons. We performed analyses using SAS® V9·4 (SAS Institute, Inc., Cary, NC, USA).
Results
We screened 1329 people and enrolled 1105 eligible participants (Table 1). Among those enrolled, 15 participants did not have an insertion attempt (10 in the NTCu380 Mini arm and 5 in the TCu380A arm) and one insertion failure occurred in each group (Figure 1). Therefore, the user population consisted of 875 NTCu380 Mini users and 213 TCu380A users.
Table 1.
Baseline Characteristics of the Treatment Population by Trial Arm
| Characteristic | NT 380 Mini (n=875) | TCu380A (n=213) |
|---|---|---|
|
| ||
| Median age at enrollment (IQR) – years | 26.0 (22 to 30) | 26.0 (23 to 30) |
| 16 – 35 | 827 (94.5) | 201 (94.4) |
| 36 – 40 | 48 (5.5) | 12 (5.6) |
| Ethnicity – no. (%) | ||
| Hispanic or Latino | 114 (13.0) | 31 (14.6) |
| Not Hispanic or Latino | 756 (86.4) | 182 (85.4) |
| Not reported | 5 (0.6) | 0 (0.0) |
| Race – no. (%) | ||
| American Indian or Alaska Native | 9 (1.0) | 3 (1.4) |
| Asian | 81 (9.3) | 16 (7.5) |
| Black or African American | 78 (9.0) | 19 (8.9) |
| Native Hawaiian or Other Pacific Islander | 4 (0.5) | 0 (0.0) |
| White | 620 (71.5) | 156 (73.2) |
| Multiple | 56 (6.5) | 15 (7.0) |
| Other | 19 (2.2) | 4 (1.9) |
| Not reported | 8 (0.9) | 0 (0.0) |
| Parity – no. (%) | ||
| 0 | 734 (83.9) | 178 (83.6) |
| 1 | 65 (7.4) | 20 (9.4) |
| 2 | 47 (5.4) | 9 (4.2) |
| 3+ | 29 (3.3) | 6 (2.8) |
| Marital status – no. (%) | ||
| Never married | 687 (78.5) | 168 (78.9) |
| Married | 132 (15.1) | 36 (16.9) |
| Separated | 13 (1.5) | 1 (0.5) |
| Divorced | 43 (4.9) | 8 (3.7) |
| Highest education – no. (%) | ||
| Less than high school | 7 (0.8) | 1 (0.5) |
| High school grad or equivalent | 86 (9.8) | 17 (8.0) |
| Some college, no degree | 270 (30.9) | 52 (24.4) |
| Associate’s/bachelor’s degree | 385(44.0) | 118 (55.4) |
| Post-graduate degree | 127 (14.5) | 25 (11.7) |
| Mean Body Mass Index (SD) – kg/m2 * | 26.4 ± 6.6 | 26.7 ± 6.8 |
| ≥30 | 197 (22.5) | 49 (23.0) |
Body Mass Index range 15 to 55 kg/m2 for NTCu380 Mini group and 17 to 53 kg/m2 for CuT380A group.
Over the trial duration, investigators confirmed 25 pregnancies in the NTCu380 Mini group among participants ≤35 years of age; one pregnancy (year 2 of use) was ectopic. The 3-year Pearl Index, i.e., pregnancies/100 women-years, for NTCu380 Mini IUD users ≤35 years of age was 1.86 (95% CI 1.20–2.74) pregnancies per 100 women-years, or 0.98 (95% CI 0.36–2.13) in year 1, 1.65 (95% CI 0.66–3.39) in year 2, and 3.91 (95% CI 2.02–6.84) in year 3 (Table 2). The cumulative probability of pregnancy among participants ≤35 years of age at enrollment through year 3 was 4.80% (95% CI 2.75–6.86%) among NTCu380 Mini users (Figure S2).
Table 2.
Efficacy Outcomes for the NT380 Mini IUD (participants ≤35 years of age.
| No. Users | No. Cycles* | No. Pregnancies | Pearl Index (95% CI) | |
|---|---|---|---|---|
|
| ||||
| All participants | ||||
| Year 1 | 801 | 7967 | 6 | 0.98 (0.36 – 2.13) |
| Year 2 | 541 | 5530 | 7 | 1.65 (0.66 – 3.39) |
| Year 3 | 413 | 3986 | 12 | 3.91 (2.02 – 6.84) |
| Cumulative through year 3 | 804 | 17483 | 25 | 1.86 (1.20 – 2.74) |
| Nulliparous participants | ||||
| Year 1 | 687 | 6822 | 5 | 0.95 (0.31 – 2.22) |
| Year 2 | 474 | 4787 | 7 | 1.90 (0.76 – 3.92) |
| Year 3 | 361 | 3463 | 9 | 3.38 (1.54 – 6.41) |
| Cumulative through year 3 | 690 | 15072 | 21 | 1.81 (1.12 – 2.77) |
| Parous participants | ||||
| Year 1 | 114 | 1144 | 1 | 1.14 (0.03 – 6.33) |
| Year 2 | 67 | 744 | 0 | 0.00 (0.00 – 6.45) |
| Year 3 | 52 | 522 | 3 | 7.47 (1.54 – 21.83) |
| Cumulative through year 3 | 114 | 2410 | 4 | 2.16 (0.59 – 5.52) |
| No menstrual cup use † | ||||
| Year 1 | 598 | 6053 | 4 | 0.86 (0.23 – 2.20) |
| Year 2 | 409 | 4157 | 5 | 1.56 (0.51 – 3.65) |
| Year 3 | 314 | 3043 | 8 | 3.42 (1.48 – 6.73) |
| Cumulative through year 3 | 599 | 13253 | 17 | 1.67 (0.97 – 2.67) |
| Any menstrual cup use | ||||
| Year 1 | 183 | 1859.86 | 1 | 0.70 (0.02 – 3.89) |
| Year 2 | 132 | 1372.88 | 2 | 1.89 (0.23 – 6.84) |
| Year 3 | 99 | 942.85 | 4 | 5.52 (1.50 – 14.12) |
| Cumulative through year 3 | 185 | 4175.59 | 7 | 2.18 (0.88 – 4.49) |
| Length of uterine cavity ≤8 cm | ||||
| Year 1 | 724 | 7208 | 5 | 0.90 (0.29 – 2.10) |
| Year 2 | 492 | 5022 | 4 | 1.04 (0.28 – 2.65) |
| Year 3 | 374 | 3618 | 10 | 3.59 (1.72 – 6.61) |
| Cumulative through year 3 | 726 | 15849 | 19 | 1.56 (0.94 – 2.43) |
| Length of uterine cavity >8 cm | ||||
| Year 1 | 73 | 726 | 1 | 1.79 (0.05 – 9.98) |
| Year 2 | 47 | 491 | 3 | 7.94 (1.64 – 23.22) |
| Year 3 | 37 | 349 | 2 | 7.44 (0.90 – 26.89) |
| Cumulative through year 3 | 74 | 1566 | 6 | 4.98 (1.83 – 10.84) |
Efficacy was determined in each cycle and was determined as pregnancy in a participant while on treatment or in a participant who reported sexual intercourse and no use of other contraceptives in the cycle.
Pearl Index was defined as the number of on-treatment pregnancies (i.e., contraceptive failures) per 100 women-years of exposure.
We did not include a provision for correcting for multiplicity when conducting tests of clinical efficacy, thus results are reported as point estimates and 95% confidence intervals. The widths of the confidence intervals have not been adjusted for multiplicity, so the intervals should not be used in clinical decisions
Rounded to the nearest integer.
42 participants without valid menstrual hygiene product use information were excluded.
In post-hoc sensitivity analyses, the 3-year Pearl Indices for NTCu380 Mini users were1.67 (0.97–2.67) and 2.18 (95% CI 0.88–4.49) for participants not using and using menstrual cups, respectively, and the 3-year Pearl Indices for those with uterine cavity length ≤8 cm and >8 cm were 1.56 (95% CI 0.94–2.43) and 4.98 (95% CI 1.83–10.84), respectively.
Although not powered for this comparison, two pregnancies occurred among TCu380A users among participants ≤35 years of age over the 3-year trial duration with a cumulative probability of pregnancy through year 3 of 2.52% (95% CI 0.00–6.06%).
The proportion of treatment-emergent AEs experienced by at least 10% of users was similar between the NTCu380 Mini and TCu380A groups (Table 3). SAEs occurred in 31 (3.5%) NTCu380 Mini users and 4 (1.9%) TCu380A (p=0.28) users. Given the world-wide experience with mini-copper IUD’s, investigators assigned all but 3 serious AEs as unrelated to trial product use. The 3 related serious AEs, all in the NTCu380 Mini arm, included a uterine perforation, a tubo-ovarian abscess, and an ectopic pregnancy. No deaths occurred. The incidence of all AEs and SAEs are presented in Table S1.
Table 3.
Most Common (>10% of participants) Treatment Emergent Adverse Events* in Copper 380 mm2 IUD Users
| Treatment Group | ||||
|---|---|---|---|---|
|
|
||||
| NTCu380 Mini (n=875) | TCu380A (n=213) | |||
|
|
||||
| System Organ Class / Preferred Term | No. Events | No. (%) Participants (95% CI) | No. Events | No. (%) Participants (95% CI) |
|
| ||||
| Any | 2678 | 775 (88.6) (86.3–90.6) | 597 | 189 (88.7) (83.7–92.7) |
|
| ||||
| Dysmenorrhea | 503 | 396 (45.3) (41.9–48.6) | 124 | 100 (46.9) (40.1–53.9) |
|
| ||||
| Menorrhagia | 479 | 317 (36.2) (33.0–39.5) | 112 | 82 (38.5) (31.9–45.4) |
| Metrorrhagia | 346 | 258 (29.5) (26.5–32.6) | 62 | 55 (25.8) (20.1–32.3) |
| Pelvic pain | 273 | 205 (23.4) (20.7–26.4) | 56 | 53 (24.9) (19.2–31.3) |
| Dyspareunia | 102 | 94 (10.7) (8.8–13.0) | 19 | 18 (8.5) (5.1–13.0) |
| Injury, poisoning and procedural complications | 586 | 515 (58.9) (55.5–62.1) | 153 | 137 (64.3) (57.5–70.8) |
| Procedural pain | 586 | 515 (58.9) (55.5–62.1) | 153 | 137 (64.3) (57.5–70.8) |
| Infections | 389 | 257 (29.4) (26.4–32.5) | 71 | 51 (23.9) (18.4–30.3) |
| Bacterial vaginosis | 136 | 104 (11.9) (9.8–14.2) | 24 | 21 (9.9) (6.2–14.7) |
|
| ||||
| Urinary tract infection | 130 | 95 (10.9) (8.9–13.1) | 24 | 19 (8.9) (5.5–13.6) |
|
| ||||
| Nasopharyngitis | 123 | 99 (11.3) (9.3–13.6) | 23 | 20 (9.4) (5.8–14.1) |
IUD: intrauterine device.
Treatment emergent adverse events (TEAEs) were defined as events that emerge during treatment, having been absent pretreatment, or worsen relative to the pretreatment state.
Treatment group comparison refers to number of participants; Confidence intervals based on exact binomial distribution.
Participants in both arms reported a median VAS pain score of 4.9 after IUD insertion (range 0–10, with higher numbers indicating more severe pain, Table S2). From bleeding diaries collected during the first year of use (Table S3), NTCu380 Mini and TCu380A users experienced similar frequencies of prolonged bleeding episodes, (60.1% vs. 59.7%, p=0.70). The number of bleeding or spotting days decreased similarly in both treatment groups through Cycle 3 before leveling off to approximately 7 to 8 days of bleeding or spotting in both treatment groups for Cycles 4 through 13 (data not shown).
Product continuation rates for year 1 were 73.6% (95% CI 68.7–78.5%) for NTCu380 Mini users and 69.0% (95% CI 59.7–78.3%) for TCu380A users (p=0.18). The cumulative 24-month continuation rates were 59.2% (95% CI 55.3–63.1%) and 51.2% (95% CI 44.3–58.0%), respectively (p=0.03). The cumulative 36-month rates were 48.7% (95% CI 45.5–51.9%) and 42.7% (95% CI 37.0–48.5%), respectively (p=0.07; Figure 2).
Figure 2. Three-year probabilities of overall product continuation.
We pre-specified that if ≥20% of participants discontinued early due to increased menstrual blood loss or due to intermenstrual bleeding/spotting, and the proportions of these two reasons for discontinuation were notably different between treatment groups (P-value < 0.1), then the analyses of observed bleeding diary data are likely to be biased due to informative censoring and attrition. Because the pre-specified criteria were not met, no further censoring related analyses were completed.
The most common reason for early discontinuation was adverse events, accounting for 20.8% (95% CI 17.4–24.2%) and 33.2% (95% CI 25.3–41.1%) of discontinuations among NTCu380 Mini and TCu380A, respectively (p=0.001, Table 4). Bleeding and pelvic pain were the main AEs leading to early discontinuation; NTCu380 Mini users had significantly lower rates of discontinuation due to bleeding or pelvic pain through 3 years compared to TCu380A users (14.5% versus 27.3%, respectively; p<0.001, Table 4).
Table 4.
3-Year Discontinuation, by Adverse Event and Product
| Primary Reasons for Discontinuation | NTCu380 Mini (N=875) | TCu380A (N=213) | ||
|---|---|---|---|---|
|
| ||||
| Percentage (%) | No. (95% CI) | Percentage (%) | No. (95% CI) | |
|
| ||||
| Overall discontinuation | 51.3 | 449 (48.0–54.6) | 57.3 | 122 (50.6–63.9) |
| Any adverse event | 20.8 | 152 (17.4 – 24.2) | 33.2 | 57 (25.3 – 41.1) |
| Bleeding and/or pain adverse event† | 14.5 | 103 (11.4 – 17.6) | 27.3 | 46 (19.5 – 35.1) |
| Increased menstrual blood loss | 6.0 | 40 (3.8 – 8.2) | 10.5 | 16 (4.6 – 16.5) |
| Intermenstrual bleeding | 1.7 | 12 (0.5 – 2.9) | 5.0 | 7 (0.6 – 9.4) |
| Intermenstrual pain | 2.6 | 17 (1.1 – 4.1) | 10.7 | 16 (4.7 – 16.6) |
| Dysmenorrhea | 5.1 | 35 (3.1 – 7.2) | 4.2 | 7 (0.2 – 8.3) |
| Any IUD expulsion‡ | 9.6 | 66 (6.9 – 12.3) | 12.7 | 22 (6.3 – 19.1) |
| Spontaneous complete expulsion | 4.2 | 29 (2.3 – 6.1) | 2.4 | 4 (0.0 – 5.5) |
| Spontaneous partial expulsion | 5.6 | 37 (3.5 – 7.8) | 10.5 | 18 (4.6 – 16.5) |
| Accidental self-removal of IUD | 3.9 | 27 (2.1 – 5.7) | 2.3 | 3 (0.0 – 5.4) |
| Malpositioned IUD removed by clinician | 1.5 | 9 (0.4 – 2.7) | 1.8 | 3 (0.0 – 4.6) |
| Other adverse events§ | 5.3 | 36 (3.2 – 7.4) | 4.1 | 6 (0.1 – 8.1) |
| Other reasons¶ | 14.2 | 87 (11.2 – 17.3) | 16.9 | 25 (9.8 – 23.9) |
Includes IUD removals due to increased menstrual blood loss, intermenstrual bleeding, intermenstrual pain, and dysmenorrhea.
Includes complete and partial (IUD in cervix or at os) expulsions.
Includes cervicovaginal infections, anemia, pelvic inflammatory disease, IUD perforation, and psychological reasons.
Includes desire for pregnancy, withdrawal of consent, protocol violation, investigator/sponsor decision, use of prohibited medications, moved out of area, other health concerns, other personal reasons.
NTCu380 Mini and TCu380A users had similar rates of trial discontinuation due to IUD expulsion over the trial period (66 [9.6%] versus 22 [12.7%], respectively; p=0.13), but partial expulsions were more common than complete expulsions and significantly higher for the TCu380A (10.5%) compared to the NTCu380 Mini (5.6%), p=0.008.
Discussion
In this trial of mostly nulliparous participants, we observed the NTCu380 Mini IUD to have a 1-year Pearl Index of about 1 per 100 women-years (0.98 [95% CI 0.36–2.13 per 100 women-years]) and a 3-year cumulative probability of pregnancy of about 5% (4.80% [95% CI 2.75–6.86%]). We designed the trial with the expectation that among participants ≤35 years at enrollment, no more than 6 pregnancies would occur and that 410/800 (51.3%) would complete 3 years of NTCu380 use. Our observed pregnancy rate was about 4 times higher than we hypothesized (n=25) and the number of participants completing 3 years was approximately the rate expected (48.7%). In our sensitivity analysis to account for factors that might have contributed to our higher-than-expected pregnancy estimates (uterine cavity length and menstrual cup use), we found that the 1- and 3-year Pearl Index point estimates were reduced when accounting for these factors, but the upper 95% CIs still exceeded 1.0 in this analysis. We previously reported on the relationship between menstrual cup use and IUD expulsion in this trial.8 While we did not consider uterine length or menstrual cup use in our a priori statistical assumptions, these factors may be worth considering in future trials of smaller IUDs. Since, uterine measurement by sounding is also frequently done prior to IUD insertion in clinical practice. Thus, discussion of uterine cavity length and use of menstrual cups may be relevant counseling points at the time of contraceptive decision-making for marketed products, especially in countries where the NTCu380 Mini is available.9
Though we did not design this trial with the statistical power to compare pregnancy rates between the NTCu380 Mini and the TCu380A IUD, our calculated cumulative probability of pregnancy through year 3 was 2.52% (95% CI 0.00–6.30%) among TCu380A users and our trial provides some contemporary efficacy data for the TCu380A in a mostly nulliparous U.S. population. Our efficacy results align with a study conducted from 2003 to 2008 that evaluated TCu380A pregnancy rates and estimated a 1- and 3-year cumulative pregnancy rate of 1.1 (95% CI 0.6–2.2) and 2.8 (95% CI 1.3–6.0), per 100 women-years, respectively.11
The overall safety of the NTCu380 Mini IUD and TCu380A were similar in our trial. We also found that, compared to the copper IUD currently marketed in the U.S. (TCu380A), participants randomized to the NTCu380 Mini IUD experienced fewer adverse events leading to early discontinuation, in particular events related to menstrual blood loss, intermenstrual bleeding, and intermenstrual pain. By design, our trial population was disproportionately nulliparous (84%). The associated product differences we observed among secondary endpoints appeared to be driven by the nulliparous population, as we found no clearly evident product performance differences between the two IUD types in the parous population.
Despite the smaller insertion tube diameter and size of the NTCu380 Mini IUD, we did not observe differences by IUD type with insertion pain. Each site adhered to their clinical standards for management of insertional pain. Attention to insertional pain control is now recommended by the Centers for Disease Control and Prevention and is a key component of patient-centered care.12
We identified potential advantages related to side effects to using the NTCu380 Mini IUD in nulliparous individuals. NTCu380 Mini use resulted in fewer removals due to bleeding or pelvic pain compared to the TCu380A. While overall and complete expulsion rates did not differ between products, partial expulsions occurred more frequently in the TCu380A arm; this difference could be due to dissimilar IUD designs (shape, size, and flexion of arms) or could be due to timing of frequency of the examination with relation to the expulsion process (i.e., partial expulsions may become complete expulsions over time).
Strengths of our trial include its large sample size, randomized design with long-term follow up and frequent in person and telephone visits, and multiple trial sites across the U.S. supporting external validity. Additionally, only a small number of participants (n=6) were lost to follow-up. By design, this randomized clinical trial had strict inclusion/exclusion in order to provide estimates of device efficacy; while this is a design strength, as in all clinical trials, generalizability may be limited. For example, the 1-year discontinuation rates in this trial were 24.6% and 31% among those randomized to the NTCu380 Mini IUD and TCu380A, respectively as compared to the 18% discontinuation reported in the nulliparous population of an observational study that included individuals who chose to use theTCu380A3. A limitation of our trial, related to budgetary constraints and sample size requirements for a reliable Pearl Index and regulatory review of a new IUD, was our 4:1 randomization ratio. By design, we were unable to make efficacy comparisons between the two randomization groups, and the smaller sample size of the TCu380A arm resulted in a wider 95% CI and lower power for detection of differences; this may have may have prevented optimal comparisons for other outcomes, as well. Conversely, this design enabled better understanding of product performance in a population whose contraceptive needs are less well met by the TCu380A IUD.
People of reproductive age gather information about contraception from a range of sources, including social media.13 Recent trends in information-seeking indicate the desire to avoid hormonal contraceptive methods, particularly among women who are concerned about perceived mood, sexual, and health-related side effects, even in the setting of reassuring data. Environmental and sexual-equity concerns have also been noted.13,14 Currently, options for people who wish to avoid hormonal contraception, for any reason, are few, and include only barrier methods, natural family planning, and copper IUDs. In the U.S., only one copper IUD is currently available. The results of our trial suggest that the NTCu380 Mini IUD may offer design and use characteristics preferable to some potential users. Because all contraceptive methods have trade-offs, increasing the options available can promote reproductive justice15 and may help overcome access barriers.
In conclusion, our findings provide data on the efficacy, safety and continuation of the NTCu380 Mini in a U.S. population. While we did not generate comparative efficacy data, the 3-year efficacy data point estimates suggest higher pregnancy rates than other long-acting reversible (LARC) methods; the efficacy through the second year of use may be closer to what users expect with a LARC method.16 The NTCu380 Mini IUD may offer some advantages over the TCu380A IUD, particularly among nulliparous individuals. Copper IUDs are the only long-acting, non-hormonal form of reversible contraception currently available; the major challenge to their use has been an increased duration and amount of menstrual flow.17 While not available in the U.S., for many potential IUD users, the level of pregnancy risk with the NTCu380 Mini may be acceptable given the lower incidence of bleeding and pelvic pain compared to TCu380A. Our trial demonstrates that a combination of user and device characteristics does modify outcomes among IUD users, specifically continuation rates related to side effects. These factors are key considerations when bringing new effective and acceptable methods to market and inform a shared decision-making approach to contraceptive counseling for maximal public health benefit.
Supplementary Material
Acknowledgements
We thank Shacondra Johnson (FHI 360) and Gabe Aguilar (Premier Research) for their efforts on this trial. We thank the participants and study staff for their time, efforts, and contributions.
(Funded by Bill & Melinda Gates Foundation, Eunice Kennedy Shriver National Institute of Child Health and Human Development, and Mona Lisa, N.V. (Belgium); ClinicalTrials.gov number NCT03124160.)
The findings and conclusions in this article are those of the authors and do not necessarily reflect the views of Planned Parenthood Federation of America, Inc.
Financial Disclosure
Supported by a grant from Bill & Melinda Gates Foundation (no. OPP 1055878 to FHI 360), Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), and Mona Lisa, N.V. (Belgium). The views expressed in this article are those of the authors and do not necessarily reflect the views of the authors’ institutions or funders.
Data Sharing Statement
Metadata will be available for public access after final regulatory review by the U.S. FDA or two years after publication of this manuscript, whichever comes first. Interested parties may submit a request via email to the corresponding author (or subsequently named alternative) with a brief sketch of what will be done with the data. De-identified individual participant data and a data dictionary defining each raw data variable used in this analysis will be made available. No additional study documents will be provided.
References
- 1.Diedrich JT, Desai S, Zhao Q, Secura G, Madden T, Peipert JF. Association of short-term bleeding and cramping patterns with long-acting reversible contraceptive method satisfaction. Am J Obstet Gynecol 2015;212(1):50.e1–8. (In eng). DOI: 10.1016/j.ajog.2014.07.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hubacher D Copper intrauterine device use by nulliparous women: review of side effects. Contraception 2007;75(6 Suppl):S8–11. (In eng). DOI: 10.1016/j.contraception.2006.12.005. [DOI] [PubMed] [Google Scholar]
- 3.Abraham M, Zhao Q, Peipert JF. Young Age, Nulliparity, and Continuation of Long-Acting Reversible Contraceptive Methods. Obstet Gynecol 2015;126(4):823–829. (In eng). DOI: 10.1097/aog.0000000000001036. [DOI] [PubMed] [Google Scholar]
- 4.Benacerraf BR, Shipp TD, Lyons JG, Bromley B. Width of the normal uterine cavity in premenopausal women and effect of parity. Obstet Gynecol 2010;116(2 Pt 1):305–310. (In eng). DOI: 10.1097/AOG.0b013e3181e6cc10. [DOI] [PubMed] [Google Scholar]
- 5.Hubacher D, Schreiber CA, Turok DK, et al. Continuation rates of two different-sized copper intrauterine devices among nulliparous women: Interim 12-month results of a single-blind, randomised, multicentre trial. EClinicalMedicine 2022;51:101554. (In eng). DOI: 10.1016/j.eclinm.2022.101554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mona Lisa 5 Mini (NTCu380 Mini). [Package Insert]. Mona Lisa N.V, Belgium; August 2016. [Google Scholar]
- 7.Paragard (intrauterine copper contraceptive). [Package Insert]. Cooper Surgical, Trumbull, CT; February 2020. [Google Scholar]
- 8.U.S. Food and Drug Administration. Establishing Effectiveness and Safety for Hormonal Drug Products Intended to Prevent Pregnancy Guidance for Industry. https://www.regulations.gov/docket/FDA-2019-D-21532019.
- 9.Brown JE, Creinin MD, Wu H, et al. Menstrual cup use and intrauterine device expulsion in a copper intrauterine device randomized trial. Contraception 2024;134:110415. (In eng). DOI: 10.1016/j.contraception.2024.110415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Creinin MD, Vieira CS, Westhoff CL, Mansour DJA. Recommendations for standardization of bleeding data analyses in contraceptive studies. Contraception 2022;112:14–22. (In eng). DOI: 10.1016/j.contraception.2022.05.011. [DOI] [PubMed] [Google Scholar]
- 11.Bahamondes L, Brache V, Meirik O, Ali M, Habib N, Landoulsi S. A 3-year multicentre randomized controlled trial of etonogestrel- and levonorgestrel-releasing contraceptive implants, with non-randomized matched copper-intrauterine device controls. Hum Reprod 2015;30(11):2527–38. (In eng). DOI: 10.1093/humrep/dev221. [DOI] [PubMed] [Google Scholar]
- 12.Nguyen AT, Curtis KM, Tepper NK, et al. U.S. Medical Eligibility Criteria for Contraceptive Use, 2024. MMWR Recomm Rep 2024;73(4):1–126. (In eng). DOI: 10.15585/mmwr.rr7304a1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Le Guen M, Schantz C, Régnier-Loilier A, de La Rochebrochard E. Reasons for rejecting hormonal contraception in Western countries: A systematic review. Soc Sci Med 2021;284:114247. (In eng). DOI: 10.1016/j.socscimed.2021.114247. [DOI] [PubMed] [Google Scholar]
- 14.Pfender EJ, Fowler LR. Social Media Is Influencing Contraceptive Choice. J Womens Health (Larchmt) 2024;33(5):563–564. (In eng). DOI: 10.1089/jwh.2023.1152. [DOI] [PubMed] [Google Scholar]
- 15.Cadena DS, Chaudhri A, Scott C. Contraceptive Care Using Reproductive Justice Principles: Beyond Access. Am J Public Health 2022;112(S5):S494–s499. (In eng). DOI: 10.2105/ajph.2022.306915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Stoddard A, McNicholas C, Peipert JF. Efficacy and safety of long-acting reversible contraception. Drugs 2011;71(8):969–80. (In eng). DOI: 10.2165/11591290-000000000-00000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Teal S, Edelman A. Contraception Selection, Effectiveness, and Adverse Effects: A Review. JAMA 2021;326(24):2507–2518. DOI: 10.1001/jama.2021.21392. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Metadata will be available for public access after final regulatory review by the U.S. FDA or two years after publication of this manuscript, whichever comes first. Interested parties may submit a request via email to the corresponding author (or subsequently named alternative) with a brief sketch of what will be done with the data. De-identified individual participant data and a data dictionary defining each raw data variable used in this analysis will be made available. No additional study documents will be provided.