Original Research ajog.

org

GYNECOLOGY
Placement of an intrauterine device within 48 hours
after early medical abortion—a randomized controlled
trial
Sara Hogmark, MD; Karin Lichtenstein Liljeblad, MD; Niklas Envall, RNM, PhD; Kristina Gemzell-Danielsson, MD, PhD;
Helena Kopp Kallner, MD, PhD

BACKGROUND: Intrauterine devices are safe, well-tolerated, and successful intrauterine device placement were similar between the
known to reduce the risk of unwanted pregnancies. At medical abortion, groups. Patients in the intervention group had lower pain scores at
intrauterine devices are placed at a follow-up visit. Patients who miss this placement of the intrauterine device (mean pain score [visual analogue
visit risk being left without contraception. scale], 32.3; standard deviation, 29) compared with the control group
OBJECTIVE: This study aimed to investigate if placement of an intra- (mean pain score [visual analogue scale], 43.4; standard deviation, 27.9;
uterine device within 48 hours of completed medical abortion at up to 63 P¼.002). Patients preferred their allocated time of placement significantly
days’ gestation leads to higher user rates at 6 months after the abortion more often in the intervention group (83/111, 74.8%) than in the control
compared with placement at 2 to 4 weeks after abortion. Furthermore, we group (70/114, 61.4%; P¼.03). Use of ultrasound at intrauterine device
aimed to compare continued use of intrauterine devices, safety, and pa- placement (because of doubts concerning complete abortion) was more
tient satisfaction between groups. common in the intervention group (43/108, 39.8%) than in the control
STUDY DESIGN: We performed an open-label, randomized, group (15/101, 14.9%; P<.001), and in one patient in the control group a
controlled, multicenter, superiority trial (phase 3). A total of 240 patients retained gestational sac was found. Three patients in the intervention
requesting medical abortion at up to 63 days’ gestation and opting for group and 2 in the control group had a vacuum aspiration. No difference
an intrauterine device were allocated to placement within 48 hours of was found in intrauterine device expulsion rates between the groups.
complete medical abortion (intervention group) or at 2 to 4 weeks after Expulsion during the first 6 months after abortion was experienced by 9 of
abortion (control group). We defined the abortion as complete after 97 (9.3%) patients in the intervention group and 4 of 89 (4.5%; P¼.25) in
bleeding with clots and cessation of heavy bleeding following the use of the control group. There were no perforations or infections requiring
misoprostol. Patients answered questionnaires at 3, 6, and 12 months. antibiotic treatment.
The primary outcome was use of intrauterine device at 6 months CONCLUSION: Placement of an intrauterine device within 48 hours
postabortion. Secondary outcomes included expulsion rate, pain at after medical abortion at 63 days’ gestation does not lead to higher user
placement, adverse events and complications from the abortion, rates at 6 months after the abortion compared with intrauterine device
acceptability, and pregnancies and their outcomes. Differences in placement at 2 to 4 weeks after abortion. When compared with placement
nonparametric continuous variables were analyzed with the at a follow-up visit after 2 to 4 weeks, intrauterine device placement within
ManneWhitney U test and differences in dichotomous variables with the 48 hours after early medical abortion seems safe, is preferred by patients,
chi square or Fisher exact tests. A P value of <.05 was considered and is associated with lower pain scores.
statistically significant.
RESULTS: In the intervention group, 91 of 111 (82%) participants used Key words: contraception, copper intrauterine devices, family planning
an intrauterine device at 6 months after the abortion vs 87 of 112 (77.7%) services, hormone-releasing intrauterine device, induced abortion, long-
in the control group, with a difference in proportion of 4.3% (95% confi- acting reversible contraception, medicated intrauterine devices, post-
dence interval,
0.062 to 0.148; P¼.51). Attendance rate and rate of abortion intrauterine device insertion

Introduction have been shown to benefit from long- IUDs are safe, well-tolerated, and
Individuals having an abortion are often acting reversible contraception, such as known to reduce the risk of unwanted
at risk for a new unplanned pregnancy intrauterine devices (IUDs).1 Because pregnancies and subsequent need for
with a subsequent need for abortion, and fertility may return already 8 to 10 days abortions.1,5e7 To increase access to
after abortion, and resumption of sexual immediate use of contraception after
activity within few weeks of abortion is pregnancy, placement of an IUD at the
Cite this article as: Hogmark S, Lichtenstein Liljeblad K,
Envall N, et al. Placement of an intrauterine device within common, immediate initiation of time of cesarean delivery or after vaginal
48 hours after early medical abortion—a randomized contraception is important for patients birth is routinely performed in many
controlled trial. Am J Obstet Gynecol 2023;228:53.e1-9. who wish to avoid a subsequent preg- settings today. Infection and expulsion
0002-9378
nancy.2,3 Sweden has the highest abor- rates are low after planned cesarean de-
ª 2022 The Authors. Published by Elsevier Inc. This is an tion rate in Western Europe (18/1000 livery, whereas expulsion rates are
open access article under the CC BY-NC-ND license (http:// women of fertile age). Half of the abor- considerably higher after vaginal
creativecommons.org/licenses/by-nc-nd/4.0/).
https://doi.org/10.1016/j.ajog.2022.07.063
tion cases are among individuals with at birth.8e10 IUD placement at the time
least 1 previous abortion.4 of first-trimester surgical abortion is

JANUARY 2023 American Journal of Obstetrics & Gynecology 53.e1

Original Research GYNECOLOGY ajog.org

expected that early placement would

AJOG at a Glance increase the long-term use of IUDs
Why was this study conducted? with maintained safety and patient
We aimed to compare immediate placement of an intrauterine device (IUD) satisfaction.
within 48 hours after medical abortion with placement at 2 to 4 weeks after
abortion. Materials and Methods
The design was an open-label, random-
Key findings ized, controlled, multicenter, superiority
Immediate IUD placement after medical abortion does not lead to higher user trial (phase 3). We recruited patients
rates at 6 months after abortion compared with placement after 2 to 4 weeks. aged 18 years requesting medical
Immediate IUD placement is safe, without increased rates of expulsion, perfo- abortion with gestation of 63 days and
ration, or infection. opting for postabortion IUD at the gy-
Immediate IUD placement results in lower pain scores and is preferred by necology clinics of Danderyd, Stock-
patients. holm South General, Falun/Mora,
Uppsala University, and Helsingborg
What does this add to what is known? hospitals in Sweden. Exclusion criteria
This study provides evidence for clinical guidelines to include IUD placement were contraindications for medical
within 48 hours after medical abortion and thereby improve patient satisfaction abortion or IUD use, inability to give
with maintained safety and improved access to IUDs. informed consent, and abortion-related
complications (septic abortion,
bleeding >1000 mL, uterine atony, and
placental retention). We provided writ-
well-studied and regarded as convenient, patients attend and that adverse events ten and oral information about the study
safe, and effective.11 However, gradually are few.14,15 for patients with an appointment for
since its introduction, medical abortion In addition to these studies, Korjamo medical abortion at the involved clinics
has become the primary method of et al16 performed a randomized who had chosen to have an IUD for
choice in many settings (95% of abor- controlled trial (RCT) with placement of postabortion contraception. All patients
tions in Sweden), and many patients opt a hormonal IUD within 3 days of miso- had the opportunity to ask questions and
for IUD as postabortion contraception.4 prostol administration or at 2 to 4 weeks all study procedure followed the Decla-
Moreover, the vast majority (85%) of after medical abortion at up to 63 days’ ration of Helsinki recommendations for
abortions are early abortions at <63 gestation. In this trial, rates of total IUD physicians in biomedical research
days’ gestation. In contrast to surgical expulsion were comparable, but rates of involving human subjects. After signing
abortion, IUDs are traditionally placed partial expulsion were higher in the early informed consent, we randomized pa-
at a follow-up visit 2 to 4 weeks after a group. The low number of participants tients who took mifepristone to place-
medical abortion. In Sweden, the current (n¼108) and a loss to follow-up rate of ment of an IUD within 48 hours after
recommendation is to place IUDs after nearly 30% in the intervention group complete abortion (intervention group)
medical abortion within 1 week of limit conclusions that could be drawn or to IUD placement at a scheduled
misoprostol administration. However, from this study. follow-up visit 2 to 4 weeks after abor-
adherence to this guideline is low.12 Pa- Although guidelines include immedi- tion according to routine care (control
tients who miss this follow-up visit risk ate placement of IUDs after medical group). The abortion was defined as
being left without contraception.13 abortion,17 there have been, to the best complete after patients reported
Previous trials have examined the of our knowledge, no previous studies bleeding with clots and cessation of
effectiveness and safety of IUD place- on the placement of IUDs as early as heavy bleeding following the use of
ment after medical abortion within 1 within 48 hours after complete abortion misoprostol and the providers found no
week of mifepristone administration or without routine use of ultrasound after reason to suspect an incomplete abor-
at a later time point.14,15 However, these early medical abortion. The aim of this tion on the basis of patient history. To
trials have been performed with an ul- RCT was to compare the rate of IUD use confirm complete abortion, all patients
trasound examination before IUD at 6 months following medical abortion took a low-sensitivity urine pregnancy
placement as part of the protocol. These at up to 63 days’ gestation between pa- test at 2 to 4 weeks after the abortion,
studies have shown similar expulsion tients who had an IUD placed within 48 either self-administered at home, or by
rates with earlier and later times of hours after complete abortion and those clinic staff at a follow-up visit. The
placement and that postabortion endo- who had it placed 2 to 4 weeks after numbers of included patients were not
metrial thickness does not correlate to abortion. We hypothesized that place- expected to be equal at each site, which
risk of IUD expulsion. Moreover, with ment of an IUD early after medical was considered in the statistical analysis.
earlier placement visits, these studies abortion would lead to placement rates The randomization ratio between the
have shown that a greater proportion of of close to 100%. Furthermore, we intervention and control group was 1:1

53.e2 American Journal of Obstetrics & Gynecology JANUARY 2023

ajog.org GYNECOLOGY Original Research

in permuted blocks of 4 to 8. The study copper IUD is provided for free in most which is commonly observed in abortion
was approved by the Regional Ethics settings. All IUDs were placed by staff at studies. With a power of 90% and an
Committee of Stockholm (permit the same clinic. If a patient missed the alpha of 0.05, we needed to randomize
number 2016/1685-31/1) with an placement visit, 3 attempts were made to 240 patients. An interim analysis was
amendment for new study centers reach the patient by telephone to performed when 50% of patients had
(permit number 2021-02625). The study reschedule. been recruited, with the predefined de-
was also approved by the European Included patients were asked to cision to stop inclusion in case of
Medicines Agency (EudraCT number participate in the study for 12 months. expulsion rates exceeding 20% or
2018-000287-29). Follow-up was at 3, 6, and 12 months, acceptability rates <50% at the 3-month
The medical termination of preg- either by a phone call and/or an e-mail follow-up in any group.
nancy was carried out according to the with a link to a structured questionnaire We performed statistical analyses us-
World Health Organization guidelines.18 with multiple questions related to the ing IBM SPSS Statistics for Windows,
We performed a chlamydia polymerase primary and secondary outcomes of the version 26 (IBM Corp, Armonk, NY).
chain reaction test, unless the patient study. There was no mandatory follow- The main analysis for the primary
actively abstained, and screening for up visit after the IUD placement visit. outcome was a modified intention-to-
bacterial vaginosis using Amsel’s criteria. The primary outcome of the study was treat (mITT) analysis including all ran-
Patients with bacterial vaginosis received IUD use at 6 months postabortion, domized patients with medical abortion
treatment started before or at the same evaluated as the proportion of patients and follow-up recorded at 6 months.
time as the abortion. No routine antibi- using IUD vs not using IUD. The sec- Hence, also patients with no IUD
otics were given. Ultrasound verification ondary outcomes were rates of IUD placement and patients experiencing
of complete abortion was not mandatory placement at allocated time, reasons for expulsion were included in the mITT
according to protocol, except in the case nonplacement of IUD, expulsion rate, population. The analyses included the
of doubt concerning complete abortion. pain at placement, adverse events and full dataset, and all results were based on
Included patients who had home complications from the abortion, observed outcomes without imputation
administration of misoprostol and were acceptability, and pregnancies and of missing data. Nonparametric contin-
allocated to the intervention group were abortions evaluated at the 3-, 6-, and uous variables are presented as medians
scheduled for IUD placement within 48 12-month follow-up. with minimum and maximum values;
hours of misoprostol administration. We measured pain scores at IUD differences between groups were
Patients treated in the clinic could have placement using a visual analogue scale analyzed by the Fisher exact test.
their IUD placed immediately after (VAS) ranging from 0 to 100, where 0 is Dichotomous variables are presented as
assumed complete abortion or return equal to no pain and 100 to the worst proportions with differences between
within 48 hours for placement. Patients imaginable pain. We asked patients to groups analyzed by the chi square or
in the control group were scheduled for indicate the pain before IUD placement, Fisher exact test, as appropriate. All dif-
an appointment for placement after 2 to at placement of tenaculum, at sounding, ferences between groups were consid-
4 weeks. The study drugs approved for at placement of the IUD, and before they ered as statistically significant if they had
this study were: Mirena (levonorgestrel- left the clinic. a P value <.05.
releasing intrauterine system [LNG- The rate of expulsion reported by
IUS], 52 mg), Kyleena (LNG-IUS, 19.5 patients was limited to complete expul- Results
mg), Jaydess (LNG-IUS, 13.5 mg, mar- sion because no clinical examination was From January 2019 to February 2021, a
keted as Skyla in some countries), and undertaken to detect partial expulsion as total of 240 patients having early medical
NovaT 380 (Cu-IUD), manufactured by part of scheduled follow-up. Partial abortion at up to 63 days’ gestation and
Bayer GmBH, Leverkusen, Germany. We expulsion could hence only be diagnosed opting for IUD postabortion were
placed all products according to the in- at a clinically motivated visit or if pa- included in the trial. The flow of patients
structions by the manufacturer and tients felt the IUD. is described in Figure 1. A total of 120
provided all IUDs at no cost according to We hypothesized that the use of IUD 6 patients were randomized to IUD
regulations by the Medical Products months after abortion would be at least placement within 48 hours after
Agency. This is a deviation from clinical 20% higher in the intervention group. assumed expulsion of the pregnancy,
practice. In Sweden, women aged <26 The sample size was calculated on the and 103 (85.8%) of these received the
years normally pay a sum corresponding basis of the hypothesis of 80% IUD use allocated intervention. We randomized
to approximately $10/V10 for the de- in the intervention group and 60% use in 120 patients to the control group with
vice. Women aged >25 years pay the control group at 6 months after IUD placement after 2 to 4 weeks. Of
approximately $95 for the hormonal abortion. Three percent to 5% were these, 92 (76.7%) received the allocated
IUDs used in this study. The placement estimated to need a vacuum aspiration intervention. Three patients were
visit is always provided for free within because of incomplete abortion and/or excluded in the intervention group
the healthcare system in Sweden if the prolonged bleeding, and approximately because of withdrawal of consent (n¼2)
IUD is placed by a nurse-midwife. The 15% loss to follow-up was expected, or not having an abortion (n¼1). In the

JANUARY 2023 American Journal of Obstetrics & Gynecology 53.e3

Original Research GYNECOLOGY ajog.org

FIGURE 1
CONSORT 2010 flow diagram

CONSORT, Consolidated Standards of Reporting Trials; ITT, intention-to-treat; IUD, intrauterine device; mITT, modified intention-to-treat.
Hogmark. Immediate vs delayed placement of intrauterine devices after early medical abortion. Am J Obstet Gynecol 2023.

53.e4 American Journal of Obstetrics & Gynecology JANUARY 2023

ajog.org GYNECOLOGY Original Research

control group, 3 patients withdrew

TABLE 1
consent. The groups were comparable
Baseline characteristics
regarding baseline characteristics and
the IUDs chosen (Table 1). Characteristics Intervention (n¼120) Control (n¼120)
This report includes analysis of pri- Demographic characteristics
mary outcome and secondary outcomes
at 6 months in an mITT analysis. For the Age
primary outcome and for IUD expulsion Median 31 30
rates, we also present per-protocol ana- IQR 26e35 26e35.75
lyses. Patients in both groups who did Min-max 18e48 18e48
not come for IUD placement, who had
the IUD placed outside the allocated Missing 1
time window or during surgery, or for Number of school years, n (%)
whom IUD placement failed were 9 3 (2.5) 6 (5)
included in the mITT analysis but 10e12 60 (50.4) 58 (48.3)
removed from the per-protocol popula-
tion. One patient in the intervention >12 56 (47.1) 56 (46.7)
group had 68 days’ gestation and was Missing 1
included in the mITT analysis. In the Other baseline characteristics
intervention group, 91 of 111 (82%)
Gestational age at mifepristone intake
(Table 2) used an IUD at 6 months after
the abortion vs 87 of 112 (77.7%) in the Median 43 42
control group, with a difference in pro- IQR 40e51.25 38e49.75
portion of 4.3% (95% confidence Min-max 28e68 28e63
interval,
0.062 to 0.148; P¼.51). A
Missing 2
sensitivity analysis with imputation of
the results with these proportions did Parous women, n (%) 88 (73.3) 84 (70)
not change results significantly. Previous abortion, n (%) 72 (60.5) 72 (60)
Attendance rate and rate of successful Missing 1
IUD placement were similar between
Misoprostol taken at home, n (%) 104 (87.4) 101 (84.2)
groups. In the intervention group, 108 of
117 (92.3%) patients attended the IUD Missing 1
placement visit. IUDs in the intervention Type of IUD placed, n (%)
group were placed at a median of 42 Mirena 59 (54.6) 45 (44.1)
hours (interquartile range, 21e46 hours,
Kyleena 39 (36.1) 49 (48)
minimum 0 and maximum 144 hours)
after completed abortion, and all place- Jaydess 3 (2.8) 1 (1.0)
ments were successful except one. The Copper IUD Nova T 7 (6.5) 7 (6.9)
failure was because of severe pain and IUD not placed 12 18
inability to pass the IUD through the
Baseline characteristics of patients having medical abortion at up to 63 days’ gestation and opting for IUD as postabortion
internal cervical os. The patient declined contraception (N¼240).
another try with anesthetics, and at 6 IQR, interquartile range; IUD, intrauterine device.
months postabortion she used condoms Hogmark. Immediate vs delayed placement of intrauterine devices after early medical abortion. Am J Obstet
Gynecol 2023.
for contraception. In the control group,
103 of 118 (87.3%) patients attended the
IUD placement visit. There were 3 failed
placements. Two were owing to severe placed later on, but became pregnant consisted of 97 patients in the interven-
pain, and the third one was because of with an ectopic pregnancy, as described tion group and 89 in the control group.
signs of infection. Of the 2 patients with below. The use of IUD at 6 months was 84 of 97
severe pain at placement, one had an The per-protocol analysis included all (86.6%) (Table 2) in the intervention
IUD placed under general anesthesia patients with medical abortion at 63 group and 79 of 89 (88.8%) in the con-
later on, but then withdrew consent to days’ gestation and without surgical trol group (P¼.82).
continue participation. The other pa- intervention who had an IUD placed Use of ultrasound at IUD placement
tient withdrew consent when contacted within the allocated time window and was more common in the intervention
at the 3-month follow-up. The patient who filled out the 6-month follow-up group than in the control group
with signs of infection had an IUD questionnaire. The per-protocol population (P<.001). In the intervention group,

JANUARY 2023 American Journal of Obstetrics & Gynecology 53.e5

Original Research GYNECOLOGY ajog.org

There were no perforations or infections

TABLE 2
requiring antibiotic treatment.
Proportion of patients using an intrauterine device at 6 months
No difference was found in IUD
Intention-to-treat Per-protocol expulsion rate between the groups ac-
cording to the mITT analysis (Table 3).
Intervention Control Intervention Control
IUD users n¼111 n¼112 P value n¼97 n¼89 P value IUD expulsion during the first 6 months
after abortion was experienced by 10 of
At 6 mo, 91 (82%) 87 (77.7%) .51 84 (86.6%) 79 (88.8%) .82
n (%) 111 (9.0%) patients in the intervention
group (7 parous and 3 nulliparous) and 4
Intervention¼placement of an intrauterine device within 48 hours after early medical abortion; control¼placement of an in-
trauterine device at 2 to 4 weeks after early medical abortion. Proportions of patients using an intrauterine device at 6 months of 112 (3.6%; P¼.11) patients in the
following early medical abortion (N¼240). P value calculated with Fisher exact test. control group (2 parous and 2 nullipa-
IUD, intrauterine device. rous). At 6 months after the abortion, 4
Hogmark. Immediate vs delayed placement of intrauterine devices after early medical abortion. Am J Obstet
Gynecol 2023. of the 14 patients who had experienced
IUD expulsion were using an IUD. All
expelled IUDs were hormonal IUDs (10
Mirena and 4 Kyleena). In the per-
ultrasound was used in 43 of 108 Healthcare providers rated the ease of protocol analysis of expulsion rates, the
(39.8%) patients. There were no pa- IUD placement as very easy/easy/neither difference was also not significant
tients with retained products of easy nor difficult in 106 of 108 (98.1%) (P¼.25) (Table 3).
conception. In the control group, ul- of the patients in the intervention group Within 3 months after the abortion, 1
trasound was used in 15 of 101 (14.9%) and as difficult/very difficult in 2 of 108 patient in each group became pregnant.
patients, and in one patient a retained patients (1.9%). Corresponding The patient in the intervention group
gestational sac was found. This patient numbers in the control group were 93 of had not returned for IUD placement and
had the IUD placed at the time of 101 (92.1%) and 8 of 101 (7.9%), had a subsequent abortion. The patient
vacuum aspiration. respectively (P¼.05). Patients preferred in the control group had an ectopic
Pain scores were normally distributed. their allocated time of IUD placement pregnancy with a Kyleena in situ, which
Patients in the intervention group had significantly more often in the inter- was removed at the time of surgery.
significantly lower pain scores at place- vention group (83/111, 74.8%) than in During the time period of 3 to 6 months
ment of the IUD (mean pain score VAS, the control group (70/114, 61.4%; after the abortion, 4 patients in the
32.3; standard deviation [SD], 29) P¼.03). intervention group and 3 patients in the
compared with the control group (mean A total of 3 patients in the intervention control group became pregnant. Among
pain score VAS, 43.4; SD, 27.9; P¼.002). group had a vacuum aspiration after the the patients in the intervention group, 1
A description of pain scores at different medical abortion because of retained had not come for IUD placement and
points of measurement is shown in products of conception. In the control decided to keep the pregnancy. One had
Figure 2. group, 2 patients had vacuum aspiration. an expulsion of a Kyleena and had an
abortion. Two patients had had their
Mirena extracted because of side effects.
At the 6-month follow-up, none of these
FIGURE 2
2 patients had decided how to proceed
Mean pain score during placement of intrauterine devices
with the pregnancy. In the control group,
the patient with an ectopic pregnancy at
3 months became pregnant again and
decided to keep that pregnancy. One
patient did not come for placement and
had a miscarriage. One patient had a
copper IUD extracted because of side
effects, became pregnant again, and had
a miscarriage.

Comment
Principal findings
Placement of an IUD within 48 hours
Error bars represent2 standard errors. Asterisk indicates significant difference. after complete abortion with 63 days’
IUD, intrauterine device. gestation does not lead to higher user
Hogmark. Immediate vs delayed placement of intrauterine devices after early medical abortion. Am J Obstet Gynecol 2023. rates 6 months after the abortion
compared with IUD placement at 2 to 4

53.e6 American Journal of Obstetrics & Gynecology JANUARY 2023

ajog.org GYNECOLOGY Original Research

TABLE 3
Expulsions of intrauterine devices by time postabortion
Time postabortion Intervention Control
P value,
Overall overall
Expulsion Complete n (%) Partial n (%) n (%) Complete n (%) Partial n (%) Overall n (%) expulsions
Within 3 mo
mITT 4/112 (3.6) 4/112 (3.6) 8/112 (7.1) 2/114 (1.8) 1/114 (0.9) 3/114 (2.6) .13
Per-protocol 4/97 (4.1) 3/97 (3.1) 7/97 (7.2) 2/89 (2.2) 1/89 (1.1) 3/89 (3.4) .33
Between 3e6 mo
mITT 2/111 (1.8) 0 2/111 (1.8) 1/112 (0.9) 0 1/112 (0.9) .62
Per-protocol 2/97 (2.1) 0 2/97 (2.1) 1/89 (1.1) 0 1/89 (1.1) 1.00
Within 6 mo
mITT 6/111 (5.4) 4/111 (3.6) 10/111 (9.0) 3/112 (2.7) 1/112 (0.9) 4/112 (3.6) .11
Per-protocol 6/97 (6.2) 3/97 (3.1) 9/97 (9.3) 3/89 (3.4) 1/89 (1.1) 4/89 (4.5) .25
Expulsions of intrauterine devices within 6 months following medical abortion. P values calculated with Fisher exact test.
mITT, modified intention-to-treat.
Hogmark. Immediate vs delayed placement of intrauterine devices after early medical abortion. Am J Obstet Gynecol 2023.

weeks after abortion. When compared that 57% of patients attended follow-up those of previous studies with placement
with placement at a follow-up visit after at a primary healthcare center. These of IUDs within 2 to 4 weeks after abor-
2 to 4 weeks, IUD placement within 48 results support the positive impact of tion.14,15,20 In contrast, in the interven-
hours after early medical abortion seems scheduled follow-ups, preferably at the tion group of our study, the 9.3%
safe, is preferred by patients, and is abortion clinic. expulsion rate at 6 months postabortion
associated with lower pain scores. Sääv et al15 reported continued use of was lower than the 12.5% expulsion rate
IUDs in 68% of participants in the at 3 months found by Korjamo et al.16
Results in the context of what is intervention group and 72% in the However, most expulsions in the Kor-
known delayed group. In 2011, Shimoni et al14 jamo study were asymptomatic partial
The relatively high attendance for studied IUD use in patients in New expulsions, mainly diagnosed by vaginal
placement in our study may be explained York, the United States, randomized to ultrasound at a scheduled follow-up
by booked appointments for follow-up placement of a copper IUD within 1 visit. We found similar expulsion rates
at the abortion clinic, and additionally, week of mifepristone administration in the early placement group in Sääv
provision of IUDs for free. Korjamo et al compared with 4 to 6 weeks after medi- et al.15 However, in that study the time of
and Sääv et al15 both performed their cal abortion. They reported a nonstatis- expulsion in relation to placement was
studies in high-resource settings similar tical difference, with 69% use after 6 not reported, nor did they differentiate
to ours, with the Sääv study performed months in the intervention group and between partial and total expulsion.
in Sweden and the Korjamo study per- 60% in the control group. Korjamo Studies have not been able to show that
formed in Finland. In the Korjamo et al16,19 did not report IUD use at 6 ultrasound before IUD placement can
study, women were randomized to “fast- months but at 1 year, at which continued predict risk of expulsion.21 We actively
track insertion” within 3 days of miso- use in patients with medical abortion at chose not to include an ultrasound ex-
prostol administration or placement at a <64 days’ gestation was only reported as amination as part of our follow-up
visit 2 to 4 weeks after the abortion. In “best” or “worst case” scenario. In the protocol to increase generalizability.18
the Sääv study, women were randomized “best case” scenario, 80% in the imme- In our study, ultrasound examinations
to IUD placement 5 to 9 days or 3 to 4 diate placement group and 71.7% in the did not show any retained products of
weeks after mifepristone administration. delayed group used an IUD at 1 year,19 as conception. Thus, our results are reas-
Both of these prospective studies had opposed to 82% in the intervention suring concerning the safety of early
attendance rates >85% with follow-up group and 77.7% in the control group at placement of IUDs after medical abor-
at the abortion clinic.15,16 In contrast, 6 months postabortion in our study. tion without previous ultrasound
the Pohjoranta et al study,13 also per- Expulsion rate of IUDs in the control examination. All patients had a low-
formed in Finland, retrospectively found group in our study was comparable to sensitivity pregnancy test for

JANUARY 2023 American Journal of Obstetrics & Gynecology 53.e7

Original Research GYNECOLOGY ajog.org

confirmation of abortion completion Strengths and limitations References

according to guidelines. The strength of this study was the 1. Heikinheimo O, Gissler M, Suhonen S. Age,
One hypothesis that has not been robust RCT design. It was not deemed parity, history of abortion and contraceptive
explored previously was that immediate choices affect the risk of repeat abortion.
feasible to blind participants. The study
Contraception 2008;78:149–54.
IUD placement could be less painful size was larger compared with previous 2. Schreiber CA, Sober S, Ratcliffe S,
than delayed placement because of dila- studies and powered to detect the stip- Creinin MD. Ovulation resumption after medical
tion of the cervix after the abortion. We ulated difference, but we did not reach abortion with mifepristone and misoprostol.
measured pain scores immediately on a statistical significance in the primary Contraception 2011;84:230–3.
VAS, and results clearly showed that 3. Boesen HC, Rørbye C, Nørgaard M, Nilas L.
outcome. Less common safety out-
Sexual behavior during the first eight weeks after
patients in the intervention group had comes such as perforations and serious legal termination of pregnancy. Acta Obstet
significantly lower pain scores during infections could not be assessed in a Gynecol Scand 2004;83:1189–92.
IUD placement compared with patients study of our size. We had a very low loss 4. The National Board of Health and Welfare.
in the control group. to follow-up rate for a study on abor- Statistik om aborter. Socialstyrelsens. 2021.
Available at: https://www.socialstyrelsen.se/
tion. It may be that patients who choose
statistik-och-data/statistik/statistikamnen/
Clinical implications IUD for postabortion contraception aborter/. Accessed March 20, 2022.
We have shown that placement of any constitute a subgroup among patients 5. Rose SB, Lawton BA. Impact of long-acting
IUD within 48 hours can be used in who have abortion. In addition, reversible contraception on return for repeat
clinical settings with maintained effec- participation in a study differs from abortion. Am J Obstet Gynecol 2012;206:37.
tiveness and safety. Expulsion rates were e1–6.
clinical practice and may attract a
6. Winner B, Peipert JF, Zhao Q, et al. Effec-
low and could have been because of certain subset of patients. Two factors tiveness of long-acting reversible contraception.
minimal cervical dilation in early in our study design differ from clinical N Engl J Med 2012;366:1998–2007.
medical abortion. Early placement is practice. One is that patients were 7. Kilander H, Alehagen S, Svedlund L,
preferred by patients and results in provided with the device for free. This Westlund K, Thor J, Brynhildsen J. Likelihood of
lower pain scores than placement after repeat abortion in a Swedish cohort according
was a requirement of the Medical
to the choice of post-abortion contraception: a
2 to 4 weeks postabortion. We hy- Products Agency. In addition, patients longitudinal study. Acta obstet gynecol Scand
pothesize that the lower pain scores are were given an appointment and did not 2016;95:565–71.
related to the naturally dilated cervix have to contact a provider to arrange 8. Lopez LM, Bernholc A, Hubacher D, Stuart G,
postabortion. IUD placement. This also may have Van Vliet HA. Immediate postpartum insertion of
We observed high attendance rates for intrauterine device for contraception. Cochrane
increased the proportion of patients
Database Syst Rev 2015:CD003036.
placement when IUDs are provided for who came for placement. Following 9. Makins A, Taghinejadi N, Sethi M, et al. FIGO
free, and patients are given an appoint- guidelines for placement and not hav- postpartum intrauterine device initiative:
ment for placement at the abortion ing ultrasound as a part of the protocol complication rates across six countries. Int J
clinic. The fact that patients did not have increases generalizability of results to Gynaecol Obstet 2018;143:20–7.
to physically collect an IUD at the 10. Kapp N, Curtis KM. Intrauterine device
settings where ultrasound is not
insertion during the postpartum period: a sys-
pharmacy, pay for the IUD, and book an accessible. tematic review. Contraception 2009;80:
appointment may have increased atten- 327–36.
dance for placement. Our results did not Conclusions 11. Okusanya BO, Oduwole O, Effa EE. Imme-
raise any concerns about the imple- Placement of an IUD within 48 hours diate postabortal insertion of intrauterine de-
vices. Cochrane Database Syst Rev 2014:
mentation of these practices into guide- after medical abortion with 63 days’
CD001777.
lines and clinical practice. gestation is safe and can be performed 12. Egardt A, Algovik M, Kopp Kallner H.
without ultrasound examination. It Adherence to Swedish national recommenda-
Research implications does not lead to higher user rates 6 tions for long acting reversible contraceptive
months after the abortion, but is provision at the time of medical abortion.
This study reports the primary outcome Contraception 2021;103:261–4.
and secondary outcomes of our trial at preferred by patients and associated
13. Pohjoranta E, Mentula M, Suhonen SP,
6-month follow-up. The study is part of with lower pain scores when compared Heikinheimo O. Predicting poor compliance with
a larger study in which placement of with IUD placement at 2 to 4 weeks follow-up and intrauterine contraception ser-
IUD within 48 hours after medical after abortion. n vices after medical termination of pregnancy.
BMJ Sex Reprod Health 2018 [Epub ahead of
abortion is studied up to 22 weeks of print].
pregnancy. The impact of gestational Acknowledgments 14. Shimoni N, Davis A, Ramos ME, Rosario L,
length and type of IUD on effectiveness, We wish to acknowledge the research co- Westhoff C. Timing of copper intrauterine device
pain at placement, and risk of expulsion ordinators and nurse-midwives who have been insertion after medical abortion: a randomized
involved in this study by identifying patients controlled trial. Obstet Gynecol 2011;118:
remains unknown. In addition, long-
suitable for recruitment and by helping with 623–8.
term continued use of IUDs after follow-up. We also wish to acknowledge the 15. Sääv I, Stephansson O, Gemzell-
medical abortion remains to be individuals who have kindly participated in the Danielsson K. Early versus delayed insertion of
explored. study. intrauterine contraception after medical

53.e8 American Journal of Obstetrics & Gynecology JANUARY 2023

ajog.org GYNECOLOGY Original Research

abortion - a randomized controlled trial. PLoS holm, Sweden (Drs Lichtenstein Liljeblad and Kopp submitted work. The remaining authors report no conflict
One 2012;7:e48948. Kallner); School of Health and Welfare, Dalarna University, of interest.
16. Korjamo R, Mentula M, Heikinheimo O. Fast- Falun, Sweden (Dr Envall); Department of Women’s and This research was funded by the Swedish Society of
track vs. delayed insertion of the levonorgestrel- Children’s Health, Karolinska Institutet, Stockholm, Medicine (SLS-692651) and Stockholm Region/Kar-
releasing intrauterine system after early medical Sweden (Drs Envall and Gemzell-Danielsson); and World olinska Institutet (ALF, LS-2016-1376, RS 2019-1054. K
abortion - a randomized trial. Contraception Health Organization Collaborating Centre for Research 0138-2015, LS 2018-1257). Grants paid for hormonal
2017;96:344–51. and Research Training in Human Reproduction, Depart- intrauterine devices used in the study and salaries for
17. Abortion care guideline. Geneva: World ment of Gynecology and Reproductive medicine, Kar- participating researchers and staff.
Health Organization; 2022. Licence: CC BY-NC- olinska University Hospital, Stockholm, Sweden (Dr The Center for Clinical Research Dalarna, Uppsala
SA 3.0 IGO. Gemzell-Danielsson). University (CKFUU-789891, CKFUU-895571, CKFUU-
18. World Health Organization. Medical man- Received April 8, 2022; revised July 29, 2022; 934198, CKFUU-936612, CKFUU-940743, CKFUU-
agement of abortion, vol. 2021; 2018. accepted July 31, 2022. 963459) provided financial support for S.H. when
19. Korjamo R, Mentula M, Heikinheimo O. Im- S.H. reports personal fees from Gedeon Richter and conducting the study and preparing the article during her
mediate versus delayed initiation of the Bayer, outside the submitted work; N.E. reports personal doctoral studies and for salaries for participating staff.
levonorgestrel-releasing intrauterine system fees from Bayer Sweden AB for educational activities The funding sources were not involved in study design,
following medical termination of pregnancy - 1 outside the submitted work; K.G.D. reports honoraria for the collection, analysis, or interpretation of data, the
year continuation rates: a randomised controlled educational activities and consultations for MSD/ writing of the report, or the decision to submit the article
trial. BJOG 2017;124:1957–64. Organon, Bayer AG, Gedeon Richter, Exeltis, Azanta, for publication.
20. Betstadt SJ, Turok DK, Kapp N, Feng KT, HRA-Pharma, Mithra, Natural Cycles, CampusPharma, The clinical trial identification number and the URL of
Borgatta L. Intrauterine device insertion after Cirqle Biomedical, Medicine360, MedinCell, Myovant the registration site
medical abortion. Contraception 2011;83:517–21. Sciences, and Exelgyn, and her clinic has participated in i) Date of registration July 27, 2018
21. Shimoni N, Davis A, Westhoff C. Can ultra- clinical trials conducted by Exeltis, Mithra, Bayer, MSD, ii) Date of initial participant enrollment January 16,
sound predict IUD expulsion after medical RemovAid, and Myovant Sciences. H.K.K. reports per- 2019
abortion? Contraception 2014;89:434–9. sonal fees for educational activities for Actavis, Bayer, iii) Clinical trial identification number NCT03603145
CampusPharma, Gedeon Richter, Exeltis, Nordic Pharma, iv) URL of the registration site ClinicalTrials.gov
Natural Cycles, Mithra, Consilient Health, Ferring Phar- v) Data will be made available upon request from the
Author and article information maceuticals, MSD/Organon, Preglife, provision of expert corresponding author
From the Department of Clinical Sciences, Danderyd opinions for Bayer, Evolan Pharma, Gedeon Richter, A summary of results was presented at the 16th
Hospital, Karolinska Institutet, Stockholm, Sweden (Drs Exeltis, Merck & Co., Teva Pharmaceuticals, TV4 and Congress of the European Society of Contraception and
Hogmark, Lichtenstein Liljeblad, Envall, and Kopp Kall- Natural Cycles, Pharmiva, Dynamic Code, Ellen, Ester- Reproductive Health, Ghent, Belgium, May
ner); Center for Clinical Research Dalarna, Uppsala Uni- care, and Leia, and acting as an investigator in trials 26e28, 2022.
versity, Falun, Sweden (Dr Hogmark); Department of sponsored by Bayer, MSD, Mithra, Ethicon, Azanta, Corresponding author: Sara Hogmark, MD. sara.
Obstetrics and Gynecology, Danderyd Hospital, Stock- Gedeon Richter, Pharmiva, and Gedea outside the hogmark@regiondalarna.se

JANUARY 2023 American Journal of Obstetrics & Gynecology 53.e9