Original Research ajog.

org

GYNECOLOGY
Hormone-related side effects in new users of a
levonorgestrel 52-mg intrauterine device
Jennifer L. Kerns, MD, MS, MPH; Lisa M. Keder, MD, MPH; Carrie A. Cwiak, MD, MPH; Carolyn L. Westhoff, MD, MSc;
Mitchell D. Creinin, MD

BACKGROUND: Although the levonorgestrel 52 mg intrauterine de- over the first 180 days did not differ between prior combined hormonal and
vice is locally active and has low systemic hormone exposure, hormonal no hormonal contraception users except for acne (84 [13.0%] vs 73
intrauterine device users sometimes report hormone-related side effects. [8.5%], respectively), P¼.006, odds ratio 1.61 (95% confidence interval
OBJECTIVE: Evaluate hormone-related adverse event rates among all 1.15e2.24). However, this association was weaker after adjustment for
participants and compare these among those who used combined hor- age, race, ethnicity, obesity status, and parity (adjusted odds ratio 1.40,
monal or no hormonal contraception in the month before enrollment. 95% confidence interval 0.99e1.98) At 360 days, prior combined hor-
STUDY DESIGN: A total of 1714 women aged 16 to 45 years old monal contraception users were more likely to report acne (101 [15.7%] vs
received a levonorgestrel 52 mg intrauterine device in a multicenter 91 [10.6%], respectively, P¼.005) and orgasm/libido problems (20
phase 3 trial to evaluate contraceptive efficacy and safety for up to [3.1%] vs 12 [1.4%], respectively, P¼.03). Over the first 180 days, all side
10 years. This analysis evaluated a subset of participants who used effects other than acne were reported in less than 3% of days; acne was
combined hormonal or no hormonal contraception in the month prior to reported an average of 13 days (7.4%) per prior combined hormonal
device placement. We assessed all nonexpulsion, nonbleeding-related contraception user and 9 days (5.0%) per prior nonhormonal contracep-
events with
1% incidence at 180 days with a plan to include tion user (P<.0001). Discontinuation for evaluated side effects occurred in
weight increase regardless of incidence; we excluded events consid- 83 (5.5%) participants with no difference between those who used
ered nonhormonal. We computed 180-day side effect frequency rates combined hormonal (36 [5.6%]) or no hormonal contraception (47 [5.5%],
based on the number of days a side effect was reported during the P¼1.0) before study entry.
study period. We created a multivariable model for side effect inci- CONCLUSION: Using combined hormonal contraception prior to le-
dence at 180 days based on age, race, ethnicity, body mass index at vonorgestrel 52 mg intrauterine device placement is weakly associated
enrollment, parity, and contraception use in the month before enroll- with reporting hormonally related side effects like acne. Only a small
ment. For those side effects with a P value <.2 on univariate com- percentage of levonorgestrel 52 mg intrauterine device users experienced
parison between combined hormonal and no hormonal contraception potentially hormone-related side effects during the initial 6 months of use
users, we secondarily evaluated 360-day event rates. that resulted in discontinuation.
RESULTS: Overall, 644 participants used combined hormonal contra-
ception (primarily oral [n¼499, 77.5%]) and 855 used no hormonal Key words: contraception, intrauterine device, levonorgestrel, Liletta,
method before intrauterine device placement. Individual side effect rates side effects

Introduction as required by regulatory agencies, in- for regulatory approval of Liletta

Contraceptive clinical trials evaluating cludes new events or worsening of (Medicines360, San Francisco, CA and
efficacy require enrollment of hetero- existing conditions that occur after study Abbvie, North Chicago, IL; Liletta is a
sexually active participants; such couples product initiation. registered trademark of Odyssea
may be commonly using a contraceptive Intrauterine devices (IUDs) have the Pharma SPRL [Belgium], an Abbvie
method prior to starting the study. Thus, benefit of being locally active and affiliate).4e6
some side effects reported in the early highly effective. Although hormonal
months of a trial could be related to IUDs have very low systemic hormone Materials and methods
cessation of the prior method. When exposure,1 participants in some IUD This report represents a secondary
these studies are performed for regula- clinical trials reported hormone- analysis of data from the ACCESS IUS
tory approval, adverse event reporting, related side effects.2e5 We hypothe- multicenter trial; the methods of the
sized that some side effects may be main study have been reported previ-
related to recent cessation of estrogen- ously.4,5 A central or local Institutional
Cite this article as: Kerns JL, Keder LM, Cwiak CA, et al. containing combined hormonal Review Board for each center approved
Hormone-related side effects in new users of a levonor- contraception (CHC) and not neces- the study. All participants signed written
gestrel 52-mg intrauterine device. Am J Obstet Gynecol sarily related to initiating a progestin- informed consents before study
2024;231:628.e1-10. only IUD. To evaluate this hypothesis, participation.
0002-9378/$36.00 we reviewed available data from A Briefly, investigators at 29 clinical
ª 2024 Elsevier Inc. All rights are reserved, including those Comprehensive Contraceptive Efficacy sites in the United States invited
for text and data mining, AI training, and similar technologies.
https://doi.org/10.1016/j.ajog.2024.06.049
and Safety Study of an IUS (ACCESS healthy, nonpregnant, sexually active
IUS), the phase 3 study trial performed (at least 4 times monthly), nulliparous

628.e1 American Journal of Obstetrics & Gynecology DECEMBER 2024

ajog.org GYNECOLOGY Original Research

pruritus, cervical dysplasia, contusion,

AJOG at a Glance insomnia, back pain, dizziness, and
Why was this study conducted? presyncope as nonhormonally related.
To evaluate whether side effects reported shortly after levonorgestrel 52-mg in- For those side effects with a P value <.2
trauterine device (IUD) placement are related to other factors, especially the on univariate comparison between CHC
contraceptive method the person was using prior to placement. and NHC users at 180 days, we second-
arily evaluated 360-day event rates. We
Key findings also evaluated side effect frequency rates
Some side effects, specifically acne and orgasm/libido problems, are reported based on the number of days of a side
more frequently in persons who stop combined hormonal contraceptives to start effect reported during the 180-day study
a levonorgestrel 52-mg IUD compared to prior users of nonhormonal contra- period.
ception. Study participants recently using combined hormonal contraception We used Fisher exact and chi-square
more frequently reported acne after starting a levonorgestrel 52-mg IUD as testing for comparisons of proportions.
compared to participants who had been using nonhormonal contraception; We created a multivariable model to
obesity and nulliparity were associated with reporting acne during the first year evaluate correlates of reporting a side
after IUD placement. effect with a significant difference at 180
or 360 days in incidence between prior
What does this add to what is known? CHC and NHC users based on age, race,
No information is available for providers or patients about early side effects re- ethnicity, body mass index at enroll-
ported in new users of a levonorgestrel 52-mg IUD. These findings are novel and ment, parity, and CHC use in the month
will help providers and patients understand which effects might be more related before enrollment. We performed sta-
to stopping a combined hormonal contraceptive. tistical analyses using SAS 9.4 (SAS
Institute, Inc, Cary, NC) with a P value of
.05 considered statistically significant.
and parous women aged 16 to 45 years For this analysis, our primary aim was
(inclusive) with regular menstrual cy- to evaluate nonbleeding-related adverse Results
cles (21e35 days when not using hor- events that could represent a hormone- Of the 1714 with successful IUD inser-
mones and with a variation of typical related side effect from the IUD during tion, 1499 (87.5%) met the criteria for
cycle length of no more than 5 days), the first 180 days of IUD use. We this analysis. Demographic characteris-
and who desired a hormonal IUD for included participants with successful tics differed significantly between prior
contraception to participate. After IUD IUD placement and at least one 28-day CHC and NHC users and appear in
placement (enrollment), participants cycle of follow-up and excluded those Table 1. Among the 644 prior CHC
received a daily paper diary to record who used a progestin-only IUD, pill, or users, the contraceptive method used in
spotting, bleeding, cramping, and implant in the month before enrollment. the month before enrollment was a
additional contraceptive use. During Participants who had recently used pill (n¼499 [77.5%]), ring (n¼136
the first year, participants had in- injectable contraception were not [21.1%]), or patch (n¼9 [1.4%]). The
person follow-up visits at 1, 3, 6, and eligible for enrollment. We included all 855 prior NHC participants reported
12 months and a phone call at events reported, even if the investigator using a nonhormonal/non-IUD method
9 months. During each interaction, assessed the event as not related to the (n¼676 [79.1%]), a copper IUD (n¼30
study staff asked in an open-ended IUD. We assessed the primary outcomes [3.5%]), or no method (n¼149
manner about new health concerns of hormone-related adverse event rates [17.4%]). Overall, 1487 (99.2%), 1440
and considered any new health issue or among all participants in the analysis (96.1%), 1369 (91.3%), and 1237
worsening of an existing condition as and compared these rates among those (82.5%) participants in this analysis
an adverse event. who used CHC or no hormonal continued IUD use at 1, 3, 6 (180 days),
Investigators at each study site contraception (NHC) in the month and 12 (360 days) months, respectively.
assessed the severity of adverse events before enrollment. For the evaluation, Side effect rates over the first 180 days
(safety outcomes) and their relationship we initially included all nonexpulsion, of IUD use are presented in Table 2 for
to IUD use, placement, or removal. We nonbleeding-related events with
1% all participants in the analysis by prior
organized adverse events into standard- incidence at 180 days in prior CHC or CHC and NHC use. Acne was the only
ized terms using the Medical Dictionary NHC users with a plan to include weight side effect that differed between new
for Regulatory Activities (MedDRA) in increase regardless of incidence. We IUD users who had used CHCs
accordance with the International Con- condensed similar MedDRA terms as (84 [13.0%]) vs NHCs (73 [8.5%]),
ference on Harmonization of Technical shown in Appendix 1. We excluded up- P¼.006 before IUD use. Of note, orgasm/
Requirements for Registration of Phar- per respiratory or gastrointestinal ill- libido problems among prior CHC and
maceuticals for Human Use. nesses, vulvovaginal infections or NHC users were infrequent but almost

DECEMBER 2024 American Journal of Obstetrics & Gynecology 628.e2

Original Research GYNECOLOGY ajog.org

TABLE 1
Demographics of participants in a phase 3 study of a levonorgestrel 52-mg IUD based on contraceptive use in the
month prior to enrollmenta
Total CHC NHC
Characteristic N¼1499 n¼644 n¼855 P valueb
Age at enrollment (y) 27.05.6 26.75.1 27.25.9 .06
<25 576 (38.4) 268 (41.6) 308 (36.0) .03
25e45 923 (61.6) 376 (58.4) 547 (64.0)
Race
White 1170 (78.1) 545 (84.6) 625 (73.1) <.001
Black or African American 199 (13.3) 45 (7.0) 154 (18.0)
Asian 62 (4.1) 26 (4.0) 36 (4.2)
Multiracial 40 (2.7) 17 (2.6) 23 (2.7)
c
Other 28 (1.9) 11 (1.7) 17 (2.0)
Ethnicity
Hispanic or Latina 221 (14.7) 56 (8.7) 155 (18.1) <.001
BMI at enrollment (kg/m ) 2d
26.96.7 26.16.3 27.56.9 <.001
Obese (
30.0) 372 (24.8) 130 (20.2) 242 (28.3) <.001
Parity
Nulliparous 907 (60.5) 466 (72.4) 441 (51.6) <.001
Data presented as n (%) or meanstandard deviation.
BMI, body mass index; CHC, combined hormonal contraceptive; IUD, intrauterine device; NHC, no hormonal contraceptive.
a
Excludes 23 women with successful IUD placement and no follow-up information during the study period, 147 with levonorgestrel IUD use in the month prior to enrollment, and 45 with other
progestin-only hormonal contraception use in the month prior to enrollment; b Comparing CHC and NHC users, Fisher exact or chi-square testing; c Includes 19 American Indian or Alaska Native, 5
Native Hawaiian or Other Pacific Islander, and 4 with missing information; d Two persons with missing information.

reached statistical significance at reported in less than 3% of days any of these side effects occurred in only
180 days (13 [2.0%] vs 7 [0.8%], (Appendix 3). Acne, however, was re- 83 (5.5%) participants (Table 4), with no
respectively, P¼.07). We evaluated 6 side ported an average of 13 days (7.4%) per difference between those who used
effects with a P value <.2 at 180 days for prior CHC user and 9 days (5.0%) per CHCs or NHCs prior to study entry
the 360-day analysis (Appendix 2); from prior NHC user (P<.0001). Table 3 in- ((36 [5.6%]) vs (47 [5.5%]), P¼1.0). In
180 to 360 days, the rates for these 6 side cludes the number of days that a the first 180 days, the only side effects
effects increased slightly for acne (10.5% participant who reported the event with a discontinuation rate of 1% or
to 12.8%), alopecia (1.3% to 1.7%), experienced that side effect. Of note, more were mood changes (1.1%) and
breast pain/tenderness (4.5% to 5.3%), days per affected participant were acne (1.0%). Most of the individual
mood changes (4.5% to 5.2%), weight 100 days or more for a few categories, adverse event category numbers are
increase/poor weight loss (2.6% to including acne among 73 prior NHC small and no statistical differences are
3.6%), and orgasm/libido problems users (100 days), alopecia among 12 evident.
(1.3% to 2.1%). At 360 days, more prior prior CHC users (103 days), anxiety/ In multivariable modeling to evaluate
CHC than NHC users reported acne (101 panic issues among 9 prior CHC users factors associated with reporting acne or
[15.7%] vs 91 [10.6%], P¼.005) and (101 days), orgasm/libido issues among orgasm/libido issues during the first
orgasm/libido problems (20 [3.1%] vs 12 12 prior CHC users (106 days) and 7 6 months of IUD use (Table 5), the un-
[1.4%], P¼.03). prior NHC users (111 days), and weight adjusted odds for reporting acne based
When averaged across the study increase/poor weight loss among 27 on method used (CHC or NHC) prior to
population over the first 180 days of prior NHC users (110 days). IUD insertion (odds ratio 1.61, 95%
IUD use, the event frequency rate Despite these high number of adverse confidence interval [CI] 1.15e2.24) are
(number of days of each side effect) of event days reported by a few users over no longer significant (adjusted odds ra-
all side effects other than acne were the first 180 days, discontinuation for tio 1.40, 95% CI 0.99e1.98). Notably,

628.e3 American Journal of Obstetrics & Gynecology DECEMBER 2024

ajog.org GYNECOLOGY Original Research

effect were relatively small for the total

TABLE 2
population of levonorgestrel 52-mg IUD
Nonexpulsion side effectsa over the first 180 days of levonorgestrel 52-mg
users. Over the first 180 days of IUD use,
IUD use
all side effects other than acne were re-
Total CHC usersb NHC usersb ported in less than 3% of days (5 days per
Side effect N¼1499 n¼644 n¼855 P valuec participant). When limiting the calcula-
Abdominal discomfort/pain 43 (2.9) 18 (2.8) 25 (2.9) >.99 tion to those who reported the side ef-
Acne 157 (10.5) 84 (13.0) 73 (8.5) .006
fect, the number of days the event was
experienced was notably high for acne
Alopecia 20 (1.3) 12 (1.9) 8 (0.9) .17 (73 prior NHC users) and alopecia (12
Anxiety/panic issues 26 (1.7) 9 (1.4) 17 (2.0) .43 prior CHC users). We can postulate that
Breast pain/tenderness 67 (4.5) 34 (5.3) 32 (3.7) .16 most of these events were relatively mild
in nature as few discontinued IUD use as
Depression 30 (2.0) 12 (1.9) 18 (2.1) .85
a result.
Dyspareunia 64 (4.3) 31 (4.8) 33 (3.9) .53 Similarly, discontinuation rates also
Fatigue 15 (1.0) 6 (0.9) 9 (1.0) >.99 may provide significant insight into
Headache (nonmigraine) 86 (5.7) 34 (5.3) 52 (6.1) .58 how bothersome the side effect is for a
participant. Overall, few (5.5%) par-
Migraine headache 22 (1.5) 9 (1.4) 13 (1.5) >.99
ticipants discontinued for a potentially
Mood changes 67 (4.5) 22 (3.4) 44 (5.2) .13 hormone-related side effect in the first
Nausead 43 (2.9) 16 (2.5) 27 (3.1) .53 180 days with the most common still
Orgasm/libido problems 19 (1.3) 12 (1.9) 7 (0.8) .10
only having rates of about 1% (mood
changes and acne). The side effects re-
Ovarian cyst 22 (1.5) 11 (1.7) 11 (1.3) .52 ported with more frequency, although
Pelvic pain/discomfort 61 (4.1) 23 (3.6) 38 (4.4) .43 by few participants, did not correlate
Vaginal odor/discharge 55 (3.7) 20 (3.1) 35 (4.1) .33 with the most common side effects
resulting in discontinuation, implying
Weight increase/poor weight loss 39 (2.6) 12 (1.9) 27 (3.2) .14
that the events were relatively mild in
Data presented as n (%).
nature or there were other unmeasur-
CHC, combined hormonal contraceptive; IUD, intrauterine device; NHC, no hormonal contraceptive.
able factors.
a
New events or worsening of preexisting conditions with an incidence
1%; b Method used in the mo before IUD placement;
c
Comparing CHC and NHC users, Fisher exact test; d Vomiting (1.3% overall) not included because we could not separate
participants who had nausea and vomiting from those with just vomiting. Results in the context of what is
known
Patients use CHCs for acne treatment
both obese and nulliparous participants libido issues than participants who had and several placebo-controlled trials
had significantly higher adjusted odds not used any hormones in the month have demonstrated therapeutic benefit
(70% and 60%, respectively) of report- prior to IUD insertion. In the adjusted of combined oral contraceptives (COCs)
ing acne compared to their counterparts. analysis, prior CHC use did not predict for acne.7,8 A retrospective analysis of a
No factors were associated with orgasm/ acne or orgasm/libido issues. The commercial claims database published
libido issues, though the number of cases prior CHC and NHC groups were in 2020 included more than 330,000
was small. demographically very dissimilar, patients, of which 279,144 used a COC,
which demonstrates the inherent dif- 35,597 used a levonorgestrel IUD, 7203
Comment ferences in the characteristics of peo- used a copper IUD, and the remainder
Principal findings ple who use CHCs as compared to used progestin-only pills, implants, or
New levonorgestrel 52-mg IUD users nonhormonal methods. Interestingly, injectables.9 The authors reported
were more likely to report acne in the nonobese and parous participants had similar increases in risk of a clinical
first 180 days of use if they had used a approximately 40% lower odds of encounter for acne among those who
CHC immediately prior to IUD initi- reporting acne with levonorgestrel 52- switched from a COC to a levonorgestrel
ation as compared to those who had mg IUD use. IUD (aHR 1.93, 95% CI 1.69e2.22) and
not used hormones. However, this as- Event rates for any side effect do not copper IUD (aHR 1.70, 95% CI
sociation was weaker after adjustment reflect the significance of that effect since 1.23e2.35). These risks were higher than
for age, race, ethnicity, obesity status, an event that occurs for 1 day counts the the risk of incident acne or treatment
and parity. In extension through the same as an event that occurs for escalation from topical to oral acne
first year of use, prior CHC users were 3 months. Overall, the number of days of treatment during the first year of use
more likely to report acne or orgasm/ experiencing a hormone-related side with either IUD as compared to COC

DECEMBER 2024 American Journal of Obstetrics & Gynecology 628.e4

 Original Research
628.e5 American Journal of Obstetrics & Gynecology DECEMBER 2024

TABLE 3
Frequency (percent of total days) and days of nonexpulsion side effectsa reported over the first 180 days of levonorgestrel 52-mg IUD use
CHC usersb NHC usersb
IUD exposure (d)¼112,373 IUD exposure (d)¼148,085
Side effect Days with side effect [n (%)] # with event n¼644 Days per affected person Days with side effect [n (%)] # with event n¼855 Days per affected person
Abdominal discomfort/pain 1017 (0.91) 18 56.5 1514 (1.02) 25 60.6
Acne 8336 (7.42) 84 99.2 7326 (4.95) 73 100.4
Alopecia 1235 (1.10) 12 102.9 548 (0.37) 8 68.5
Anxiety/panic issues 908 (0.81) 9 100.9 1191 (0.80) 17 70.1

GYNECOLOGY
Breast pain/tenderness 2304 (2.05) 34 67.8 1741 (1.18) 32 54.4
Depression 948 (0.84) 12 79.0 1455 (0.98) 18 80.8
Dyspareunia 1834 (1.63) 31 59.2 2460 (1.66) 33 74.6
Fatigue 276 (0.25) 6 46.0 682 (0.46) 9 75.8
Headache (nonmigraine) 2294 (2.04) 34 67.5 2581 (1.74) 52 49.6
Migraine headache 491 (0.44) 9 54.6 652 (0.44) 13 50.2
Mood changes 1351 (1.20) 22 61.4 3811 (2.57) 44 86.6
c
Nausea 601 (0.54) 16 37.6 873 (0.59) 27 32.3
Orgasm/libido problems 1273 (1.13) 12 106.1 774 (0.52) 7 110.6
Ovarian cyst 717 (0.64) 11 65.2 551 (0.37) 11 50.1
Pelvic pain/discomfort 1011 (0.90) 23 44.0 2173 (1.47) 38 57.2
Vaginal odor/discharge 741 (0.66) 20 37.1 2154 (1.45) 35 61.6
Weight increase/poor weight 1049 (0.93) 12 87.4 2977 (2.01) 27 110.3
loss
CHC, combined hormonal contraceptive; IUD, intrauterine device; NHC, no hormonal contraceptive.
a
New events or worsening of preexisting conditions with an incidence
1%; b Method used in the mo before IUD placement; c Vomiting (1.3% overall) not included because we could not separate participants who had nausea and vomiting from those with just
vomiting.

ajog.org
ajog.org GYNECOLOGY Original Research

control group, we do not have a

TABLE 4
comparator population to know if
Discontinuation rates for nonexpulsion side effectsa over the first 180 days
these hormonal side effect rates exceed
of levonorgestrel 52-mg IUD use
what might occur in a similar popula-
Total CHC usersa NHC usersb tion not using any hormonal
Side effect N¼1499 n¼644 n¼855 P valuec contraception.
Abdominal discomfort/pain 2 (0.1) 0 2 (0.2) .51
Acne 15 (1.0) 8 (1.2) 7 (0.8) .44 Strengths and limitations
A strength of this study is the pro-
Alopecia 2 (0.1) 2 (0.3) 0 .18
spective collection of adverse events in
Anxiety/panic issues 2 (0.1) 0 2 (0.2) .51 a diverse study population. We did
Breast pain/tenderness 0 0 0 not include vulvovaginal infections or
Depression 2 (0.1) 0 2 (0.2) .51 pruritis as hormonal side effects as
levonorgestrel IUDs are not known to
Dyspareunia 3 (0.2) 2 (0.3) 1 (0.1) .58
cause alterations of the vaginal
Fatigue 1 (0.1) 0 1 (0.1) >.99 microbiome.12,13 However, recognizing
Headache (nonmigraine) 5 (0.3) 4 (0.6) 1 (0.1) .17 that some IUD users do feel they have
more vaginal complaints, we did
Migraine headache 0 0 0
include vaginal odor/discharge in the
Mood changes 17 (1.1) 7 (1.1) 10 (1.2) >.99 assessments. We chose to not include
Nausead 2 (0.1) 0 2 (0.2) .51 bleeding events as bleeding changes
Orgasm/libido problems 4 (0.3) 2 (0.3) 2 (0.2) >.99 reflect an outcome, often desired, with
levonorgestrel 52-mg IUD use. Of
Ovarian cyst 3 (0.2) 3 (0.5) 0 .08
note, 3 published reports from this
Pelvic pain/discomfort 13 (0.9) 6 (0.9) 7 (0.8) >.99 study have extensively detailed
Vaginal odor/discharge 2 (0.1) 0 2 (0.2) .51 bleeding changes during the first year
Weight increase/poor weight loss 10 (0.7) 2 (0.3) 8 (0.9) .20 of IUD use.14e16
A weakness of this study is that we did
Data presented as n (%).
not assess whether participants who had
CHC, combined hormonal contraceptive; IUD, intrauterine device; NHC, no hormonal contraceptive.
previously been using CHCs were partly
a
Includes side effects with a reported incidence
1%; b Method used in the month before IUD placement; c Comparing CHC
and NHC users, Fisher exact test; d Vomiting not included because we could not separate participants who had nausea and doing so for secondary reasons, like acne
vomiting from those with just vomiting. treatment. As a secondary analysis, the
prior CHC and NHC groups were
demographically very dissimilar, which
users. The authors concluded that these Clinical implications demonstrates the inherent differences in
observations demonstrate that some of The hormonal side effect data in this the characteristics of people who use
the observed association between non- analysis provide clear information for CHCs as compared to no hormonal
COC use and acne may be related to clinicians to counsel patients on realistic methods. These differences, as demon-
COC discontinuation and its associated expectations after hormonal IUD place- strated in the multivariable analysis,
benefits when starting the new method. ment. Importantly, the discontinuation result in the pre-IUD contraceptive
Our findings from this prospective study rate for these events is relatively low, method no longer being a significant
support those of the large retrospective which implies the severity of these issues predictor of acne or sexual complaints
analysis. is minor for most IUD users. Clinicians with levonorgestrel IUD use.
Though infrequent, some people should advise patients discontinuing
report a decrease in sexual desire with CHCs who have past issues with acne or Conclusions
levonorgestrel IUD use.10,11 The small orgasm/libido issues that discontinuing During the initial 6 months of use, only a
numbers and cross-sectional design of the CHC may result in recurrence of small percentage of levonorgestrel 52-
those evaluations limit their ability to these side effects. mg IUD users will experience poten-
evaluate appropriately for confounders tially hormone-related side effects that
such as prior CHC use. Although we Research implications will result in IUD discontinuation.
found a higher incidence of orgasm/li- This study provides prospective data on Furthermore, using CHCs prior to
bido issues in prior CHC users by 1 year hormonal side effects in the first year of placement is not associated with who
after levonorgestrel 52-mg IUD place- levonorgestrel 52-mg IUD use. Because will report acne. Clinicians can use this
ment, the overall rate remained very low. phase 3 studies like this one have no information to provide relevant

DECEMBER 2024 American Journal of Obstetrics & Gynecology 628.e6

 Original Research
628.e7 American Journal of Obstetrics & Gynecology DECEMBER 2024

TABLE 5
Factors associated with acne and orgasm/libido side effects over 180 d of levonorgestrel 52-mg IUD use (N[1499)
Acne Orgasm/libido issues
Number of
a
Characteristic subjects n (%) Odds ratio Adjusted odds ratio n (%) Odds ratio Adjusted odds ratioa
Age (y)
<25 576 54 (9.4) Referent Referent 8 (1.4) Referent Referent

25 923 103 (11.2) 1.21 (95% CI 0.86e1.72) 1.44 (95% CI 1.00e2.07) 16 (1.7) 1.25 (95% CI 0.53e2.95) 1.35 (95% CI 0.56e3.30)
b
Race

GYNECOLOGY
White 1170 132 (11.3) Referent Referent 18 (1.5) Referent Referent
Non-White 329 25 (7.6) 0.65 (95% CI 0.41e1.01) 0.77 (95% CI 0.49e1.22) 6 (1.8) 1.19 (95% CI 0.47e3.02) 1.32 (95% CI 0.51e3.41)
Ethnicity
Non-Hispanic 1288 138 (10.7) Referent Referent 19 (1.5) Referent Referent
Hispanic or Latina 211 19 (9.0) 0.82 (95% CI 0.50e1.36) 0.98 (95% CI 0.58e1.64) 5 (2.4) 1.62 (95% CI 0.60e4.39) 1.92 (95% CI 0.69e5.33)
2c
BMI at enrollment (kg/m )
<30.0 1125 133 (11.8) 1.94 (95% CI 1.24e3.05) 1.69 (95% CI 1.07e2.69) 17 (1.5) 0.80 (95% CI 0.33e1.94) 0.77 (95% CI 0.31e1.94)

30.0 372 24 (6.5) Referent Referent 7 (1.9) Referent Referent
Parity
Nulliparous 907 112 (12.3) 1.71 (95% CI 1.19e2.46) 1.58 (95% CI 1.07e2.33) 15 (1.7) 1.09 (95% CI 0.47e2.51) 1.13 (95% CI 0.46e2.80)
Parous 592 45 (7.6) Referent Referent 9 (1.5) Referent Referent
d
Contraception use at Enrollment
CHC 644 84 (13.0) 1.61 (95% CI 1.15e2.24) 1.40 (95% CI 0.99e1.98) 14 (2.2) 1.88 (95% CI 0.83e4.26) 2.13 (95% CI 0.90e5.03)
NHC 855 73 (8.5) Referent Referent 10 (1.2) Referent Referent
BMI, body mass index; CHC, combined hormonal contraception; CI, confidence interval; IUD, intrauterine device; NHC, no hormonal contraception.
a
Adjusted odds ratio controlling for all factors in table; b Four persons with missing information; c Three persons with missing information; d Method used in the month before IUD placement.

ajog.org
ajog.org GYNECOLOGY Original Research

counseling to patients considering a le- combined oral contraceptive containing ethi- Francisco, San Francisco, CA (Kerns); Department of
vonorgestrel IUD. n nylestradiol 20 mg plus drospirenone 3mg Obstetrics and Gynecology, Ohio State University, Co-
administered in a 24/4 regimen: a pooled lumbus, OH (Keder); Department of Obstetrics and Gy-
analysis. Eur J Obstet Gynecol Reprod Biol necology, Emory University, Atlanta, GA (Cwiak);
Acknowledgments 2011;155:171–5. Department of Obstetrics and Gynecology, Columbia
The authors thank the participating investigators 9. Barbieri JS, Mitra N, Margolis DJ, Harper CC, University, New York, NY (Westhoff); and Department of
and coordinators at the 29 study centers for Mostaghimi A, Abuabara K. Influence of Obstetrics and Gynecology, University of California,
conduct of the clinical trial and submission of contraception class on incidence and severity of Davis, Sacramento, CA (Creinin).
data (investigators funded by Medicines360 to acne vulgaris. Obstet Gynecol 2020;135: Received Jan. 23, 2024; revised May 29, 2024;
conduct the study). 1306–12. accepted June 30, 2024.
10. Trigo ACM, Maron CC, Pinheiro MSA, da C.A.C. receives royalties from Contraceptive Tech-
References Silva SBL, Brito MB. Female sexual function in nology Communications and Springer. C.L.W. serves on
1. Hofmann BM, Apter D, Bitzer J, et al. women using LARC methods. Gynecol Endo- Data Safety Monitoring Boards for studies sponsored by
Comparative pharmacokinetic analysis of crinol 2022;38:68–72. Bayer and Merck & Co; serves on an Advisory Board for
levonorgestrel-releasing intrauterine systems 11. Malmborg A, Brynhildsen J, Hammar M. Agile Therapeutics; and has been a consultant for HRA
and levonorgestrel-containing contraceptives A survey of young women’s perceptions of the Pharma, Mithra, and Organon. M.D.C. receives speaking
with oral or subdermal administration route. Eur influence of the levonorgestrel-intrauterine sys- honoraria from Gedeon Richter, Mayne, OLIC, and
J Contracept Reprod Health Care 2020;25: tem or copper-intrauterine device on sexual Organon; serves on Advisory Boards for Fuji Pharma,
417–26. desire. Sex Reprod Healthc 2019;21:75–80. Gedeon Richter, GlaxoSmithKline, Mayne, Merck, OLIC,
2. Gemzell-Danielsson K, Schellschmidt I, 12. Brooks JP, Edwards DJ, Blithe DL, et al. and Organon; consults for Estetra SRL, Libbs, Mayne, and
Apter D. A randomized, phase II study Effects of combined oral contraceptives, depot Medicines360. J.L.K. and L.M.K. report no conflict of
describing the efficacy, bleeding profile, and medroxyprogesterone acetate and the interest.
safety of two low-dose levonorgestrel-releasing levonorgestrel-releasing intrauterine system on The Department of Obstetrics, Gynecology and
intrauterine contraceptive systems and Mirena. the vaginal microbiome. Contraception Reproductive Sciences, University of California, San
Fertil Steril 2012;97:616–22.e1-3. 2017;95:405–13. Francisco, receives contraceptive clinical trials
3. Nelson A, Apter D, Hauck B, et al. Two low- 13. Balle C, Happel AU, Heffron R, Jaspan HB. research funding from Bayer, Bioceptive, Medi-
dose levonorgestrel intrauterine contraceptive Contraceptive effects on the cervicovaginal cines360, Sebela, and Teva. The Department of Ob-
systems: a randomized controlled trial. Obstet microbiome: recent evidence including ran- stetrics and Gynecology, Ohio State University, receives
Gynecol 2013;122:1205–13. domized trials. Am J Reprod Immunol 2023;90: research funding for contraceptive clinical trials from
4. Eisenberg DL, Schreiber CA, Turok DK, e13785. Agile Therapeutics, Merck, Medicines360, and Sebela.
Teal SB, Westhoff CL, Creinin MD. Three-year 14. Darney PD, Stuart GS, Thomas MA, The Department of Gynecology and Obstetrics, Emory
efficacy and safety of a new 52-mg Cwiak C, Olariu A, Creinin MD. Amenorrhea University, receives contraceptive clinical trials
levonorgestrel-releasing intrauterine system. rates and predictors during 1 year of levonor- research funding from Medicines360 and Sebela.
Contraception 2015;92:10–6. gestrel 52 mg intrauterine system use. Contra- C.L.W.’s university department receives contraceptive
5. Teal SB, Turok DK, Chen BA, Kimble T, ception 2018;97:210–4. research funding from Sebela and Medicines360. The
Olariu AI, Creinin MD. Five-year contraceptive 15. Schreiber CA, Teal SB, Blumenthal PD, Department of Obstetrics and Gynecology, University of
efficacy and safety of a levonorgestrel 52-mg Keder LM, Olariu AI, Creinin MD. Bleeding pat- California, Davis, receives contraceptive research
intrauterine system. Obstet Gynecol 2019;133: terns for the Liletta levonorgestrel 52 mg in- funding for Dr Creinin from Chemo Research SL, Evo-
63–70. trauterine system. Eur J Contracept Reprod fem, HRA Pharma, Medicines360 (including medical
6. Westhoff CL, Keder LM, Gangestad A, Health Care 2018;23:116–20. and safety oversight of this trial from 2009 to 2021),
Teal SB, Olariu AI, Creinin MD. Six-year contra- 16. Chen BA, Eisenberg DL, Schreiber CA, Merck, and Sebela.
ceptive efficacy and continued safety of a levo- Turok DK, Olariu AI, Creinin MD. Bleeding Clinical trial registration: Clinicaltrials.gov
norgestrel 52 mg intrauterine system. changes after levonorgestrel 52-mg intrauter- NCT00995150 (https://www.clinicaltrials.gov/study/
Contraception 2020;101:159–61. ine system insertion for contraception in NCT00995150?term¼NCT00995150&rank¼1; regis-
7. Redmond GP, Olson WH, Lippman JS, women with self-reported heavy menstrual tration date October 14, 2009).
Kafrissen ME, Jones TM, Jorizzo JL. Norgesti- bleeding. Am J Obstet Gynecol 2020;222: Presented in part as poster abstracts at the Society of
mate and ethinyl estradiol in the treatment of SS888.e1–6. Family Planning 2019 Annual Meeting and the 2020 and
acne vulgaris: a randomized, placebo-controlled 2021 American College of Obstetrics and Gynecology
trial. Obstet Gynecol 1997;89:615–22. Author and article information Annual Clinical Meetings.
8. Koltun W, Maloney JM, Marr J, Kunz M. From the Department of Obstetrics, Gynecology and Corresponding author: Mitchell D. Creinin, MD.
Treatment of moderate acne vulgaris using a Reproductive Sciences, University of California, San mdcreinin@ucdavis.edu

DECEMBER 2024 American Journal of Obstetrics & Gynecology 628.e8

Original Research GYNECOLOGY ajog.org

Appendix 1. Medical Breast tenderness Libido decreased

Dictionary for Nipple pain Loss of libido
Regulatory Activities Depression Orgasm abnormal
terms used for Depression Ovarian cyst
determination of
reported side effects Depressed mood Ovarian cyst
Major depression Ovarian cyst ruptured
Abdominal discomfort or pain
Dyspareunia Pelvic pain or discomfort
Abdominal discomfort Fatigue Pelvic discomfort
Abdominal pain Headache
Pelvic pain
Abdominal pain lower Migraine
Uterine spasm
Migraine
Abdominal pain upper Uterine hypertonus
Acne Migraine with aura
Mood changes Vaginal discharge or odor
Acne
Affect lability Vaginal discharge
Acne cystic
Alopecia Mood altered Vaginal odor
Anxiety or panic issues Mood swings Weight increase or loss poor
Breast pain or tenderness Nausea Abnormal weight gain
Breast discomfort Orgasm or libido problems Weight increased
Breast pain Anorgasmia Weight loss poor

628.e9 American Journal of Obstetrics & Gynecology DECEMBER 2024

ajog.org GYNECOLOGY Original Research

APPENDIX 2
Select nonexpulsion side effectsa over the first 360 days of levonorgestrel
52-mg IUD use
Total CHC usersa NHC usersb
Side effect N¼1499 n¼644 n¼855 P valuec
Acne 192 (12.8) 101 (15.7) 91 (10.6) .005
Alopecia 25 (1.7) 14 (2.2) 11 (1.3) .22
Breast pain/tenderness 79 (5.3) 38 (5.9) 41 (4.8) .35
Mood changes 78 (5.2) 26 (4.0) 52 (6.1) .08
Orgasm/libido problems 33 (2.1) 20 (3.1) 12 (1.4) .03
Weight increase/poor weight loss 54 (3.6) 19 (3.0) 35 (4.1) .26
Data presented as n (%).
CHC, combined hormonal contraceptive; IUD, intrauterine device; NHC, no hormonal contraceptive.
a
New events or worsening of preexisting conditions, only includes events with CHC vs NHC incidence comparison with P value
<.2 at 6 months (180 days); b Method used in the month before IUD placement; c Comparing CHC and NHC users, Fisher
exact test.

APPENDIX 3
Frequency (percent of total days) of nonexpulsion side effectsa for all
participants over the first 180 days of levonorgestrel 52-mg IUD use
CHC usersb NHC usersb
n¼644 n¼855
IUD exposure IUD exposure
Side effect (d)¼112,373 (d)¼148,085 P valuec
Abdominal discomfort/pain 1017 (0.91) 1514 (1.02) .003
Acne 8336 (7.42) 7326 (4.95) <.0001
Alopecia 1235 (1.10) 548 (0.37) <.0001
Anxiety/panic issues 908 (0.81) 1191 (0.80) .92
Breast pain/tenderness 2304 (2.05) 1741 (1.18) <.0001
Depression 948 (0.84) 1455 (0.98) .0002
Dyspareunia 1834 (1.63) 2460 (1.66) .56
Fatigue 276 (0.25) 682 (0.46) <.0001
Headache (nonmigraine) 2294 (2.04) 2581 (1.74) <.0001
Migraine headache 491 (0.44) 652 (0.44) .90
Mood changes 1351 (1.20) 3811 (2.57) <.0001
d
Nausea 601 (0.54) 873 (0.59) .07
Orgasm/libido problems 1273 (1.13) 774 (0.52) <.0001
Ovarian cyst 717 (0.64) 551 (0.37) .52
Pelvic pain/discomfort 1011 (0.90) 2173 (1.47) <.0001
Vaginal odor/discharge 741 (0.66) 2154 (1.45) <.0001
Weight increase/poor weight loss 1049 (0.93) 2977 (2.01) .14
Data presented as n (%).
CHC, combined hormonal contraceptive; IUD, intrauterine device; NHC, no hormonal contraceptive.
a
New events or worsening of preexisting conditions with an incidence
1%; b Method used in the month before IUD
placement; c Comparing CHC and NHC users, chi-square test; d Vomiting (1.3% overall) not included because we could not
separate participants who had nausea and vomiting from those with just vomiting.

DECEMBER 2024 American Journal of Obstetrics & Gynecology 628.e10