Received: 5 April 2018 | Revised: 28 August 2018 | Accepted: 30 January 2019 | First published online: 20 February 2019

DOI: 10.1002/ijgo.12778

CLINICAL ARTICLE
Gynecology

Effects of a levonorgestrel intrauterine system versus a copper

intrauterine device on menstrual changes and uterine
artery Doppler

Mohamed Rezk1 | Elsayed Elshamy1,2,* | Abd-Elhamid Shaheen1 | Mohamed Shawky3 |

Hala Marawan4

1
Department of Obstetrics and
Gynecology, Menoufia University,
Menoufia, Egypt
2
Department of Obstetrics and
Gynecology, King Abdul-Aziz Airbase
Hospital, Dhuhran, Saudi Arabia
3
Department of Radiology, Menoufia
University, Menoufia, Egypt
4
Department of Public Health and
Community Medicine, Faculty of
Medicine, Menoufia University,
Menoufia, Egypt
*Correspondence
Elsayed Elshamy, 041/9 Prince Sattam
street, Al-Khobar, Saudi Arabia.
Email: s_shamy77@ yahoo.com
Abstract
Objective: To compare the effects of a levonorgestrel-­releasing intrauterine system
(LNG-­IUS) and a copper intrauterine device (Cu-­IUD) on menstrual changes and uterine
artery Doppler indices.
Methods: A randomized clinical trial was conducted at Menoufia University Hospital,
Egypt, between December 2016 and August 2017. 306 multiparous women desiring
intrauterine contraception were randomly assigned to LNG-­IUS (n=152) or Cu-­IUD
(n=154). Uterine artery pulsatility index (PI) and resistant index (RI) were measured
before use, and 3 and 6 months after insertion, and associations with abnormal bleeding
were evaluated.
Results: Irregular bleeding was initially reported by 31 (74%) of 42 women in the LNG-­
IUS group, and heavy menstrual bleeding by 53 (67%) of 79 women in the Cu-­IUD
group. Incidence of abnormal bleeding decreased over the 6-­month study period.
Uterine artery PI was significantly correlated with abnormal bleeding at a cutoff of 1.35
with area under the curve (AUC) 0.93, sensitivity 88%, and specificity 100%, whereas
uterine artery RI was significantly correlated with abnormal bleeding at a cutoff of 0.62
with AUC 0.1, sensitivity 96%, and specificity 100%.
Conclusion: LNG-­IUS-­related abnormal bleeding was associated with changes in uterine
artery blood flow that were not evident among Cu-­IUD users.
Pan African Clinical Trials Registry: PACTR201701001900640.

KEYWORDS
Levonorgestrel-releasing intrauterine system; Long-acting reversible contraception

1 | INTRODUCTION insertion that declined by the end of the first year.2 By contrast, the
Cu-­IUD has been associated with a 30%–50% increase in menstrual
Long-­acting reversible contraception methods including the blood loss, and approximately 10%–20% of users request its removal
levonorgestrel-­releasing intrauterine system (LNG-­IUS) and copper intra- during the first year of use.3
uterine device (Cu-­IUD) are highly effective and safe methods for pre- Although some studies have reported that the LNG-­IUS increases
1
venting pregnancy, with a low failure rate similar to that of sterilization. uterine artery Doppler indices owing to increased impedance to
In a previous study, the LNG-­IUS was associated with prolonged blood flow in the uterine arteries 3 months after use,4–6 others found
or irregular bleeding by, respectively, 22% and 67% of users early after no changes in the uterine artery pulsatility index (PI) or resistance

18 | © 2019 International Federation of wileyonlinelibrary.com/journal/ijgo Int J Gynecol Obstet 2019; 145: 18–22
Gynecology and Obstetrics
Rezk ET AL. | 19

index (RI) among users of the LNG-­IUS 3 months after insertion.7–10
Similarly, studies of uterine artery Doppler indices to identify the
association between Cu-­IUD use and heavy menstrual bleeding have
reported conflicting results.11–13
Understanding changes in the uterine vasculature among IUD
users is important in terms of both physiologic and pathophysiologic
changes in order to help counsel women about their contraceptive
choices.6 The aim of the present study was therefore to assess the
impact of the LNG-­IUS versus the Cu-­IUD on menstrual changes and
uterine artery Doppler indices at 3 and 6 months after use.
2 | MATERIALS AND METHODS
The present randomized clinical trial among new IUD users was con-
ducted at the Department of Obstetrics and Gynecology in collabora-
tion with the Radiology and Public Health and Community Medicine
Departments, Faculty of Medicine, Menoufia University, Shibin El-­kom
City, Menoufia governorate, Egypt, between December 15, 2016, and
August 31, 2017. The study protocol was approved by the research
ethical committee of Menoufia Faculty of Medicine (no. 456M/2016)
in accordance with the Declaration of Helsinki. All participants signed
informed consent prior to commencement of the study.
The study recruited multiparous women with a normal menstrual
cycle attending the family planning clinic who requested long-­acting
reversible contraception and fulfilled the criteria for intrauterine con-
traception after providing a complete history and undergoing a clinical
examination and transvaginal ultrasonography. Exclusion criteria were
nulliparity, pregnancy, undiagnosed uterine bleeding, genital infection
in the preceding month, medical disorders, and any contraindication to
progestin administration or IUD insertion.
The sample size was calculated on the basis of a reported 20% dif-
ference between the two methods in abnormal bleeding.2 Thus, 150
women were required in each group for the study to have 90% power
at the 5% significance level. To allow for study dropout, 165 partici-
pants were recruited in each group.
Participants were randomized by use of computer-­generated simple
random tables to one of two groups. In group 1, women received the LNG-­
IUS (Mirena, Bayer HealthCare, Berlin, Germany); in group 2, the women
received the copper T-­380 IUD (Copper IUD, DPK, Egypt). Blinding was
not done due to the clear difference between both types of IUD.
Uterine artery Doppler was performed at the Radiology
Department before the contraceptive device was inserted, and 3
and 6 months after insertion, using a 7-­MHz transvaginal probe and
an EUB-­7000 ultrasound instrument (Hitachi, Tokyo, Japan). To avoid
interobserver variability, all examinations were performed by a single
ultrasonographer. Pulsed-­wave Doppler was applied at a sampling
gate of 2 mm with an insonation angle of less than 30°. The pulsa-
tility index (PI) and resistance index (RI) were determined automati-
cally from the mean of three similar waveforms. Confounding factors
that might affect measurements were avoided by asking the women
to abstain from taking non-­steroidal anti-­inflammatory drugs 24 hours
before the exam and by performing all scans between 8:00 and 11:00
am to eliminate diurnal variation.
The primary outcome measures were menstrual changes and
changes in uterine artery Doppler indices 3 and 6 months after inser-
tion. All women affected by abnormal bleeding at the 3-­ or 6-­month
follow-­up were included in a regression analysis to assess the correla-
tion between uterine artery Doppler indices and subsequent irregular
or heavy menstrual bleeding. The secondary outcomes were adverse
effects of the method and patient acceptability.
Women were given menstrual calendars to record the days on
which they experienced bleeding as recommended by the WHO. The

Women attending family planning clinic (n=896)

menstrual segment included both a bleeding episode and bleeding-­
free interval. Vaginal spotting was defined as very slight bleeding
requiring no sanitary protection, whereas heavy bleeding was defined
Assessed for eligibility (n=350)
Excluded (n=20)
as bleeding requiring sanitary protection. Women's acceptability in
Enrollment
Not meeting inclusion terms of compliance with follow-­up, overall satisfaction, and rec-
criteria (n=14).
Declined to participate ommendation of the method to other women was assessed at the
(n=6).
6-­month follow-­up via a pre-­designed questionnaire. Inquiry about
Randomized (n=330) adverse effects was done at each scheduled visit with appropriate
Allocation management of reported symptoms. Accurate evaluation of the symp-
toms was done to verify their association with the used method.
LNG-IUS inserted Cu-IUD inserted The data were tabulated, and analyzed by SPSS version 22 (IBM,
(n=165) (n=165)
Armonk, NY, USA). Categoric variables were summarized as number
(percentage), and continuous variable as median (interquartile range).
The X2 test was used to compare two qualitative variables; the Mann-­
Follow up
Whitney U test was used to compare two independent quantitative
Lost to follow-up (n=13) Lost to follow-up (n=11) variables. The association between PI or RI and bleeding was assessed
by Spearman correlation. Multiple logistic regression was used to
Analysis determine predictors of abnormal bleeding. Receiver operator charac-
Analyzed (n=152) Analyzed (n=154) teristic (ROC) curve analysis was used to determine the optimal cutoff
value, and sensitivity, and specificity of PI and RI to predict bleeding. A
FIGURE 1 CONSORT flow diagram. P value of less than 0.05 was considered to be statistically significant.
20 | Rezk ET AL.

TABLE 1 Uterine artery pulsatility index and resistance index by study group.

LNG-­IUS group (n=152) Cu-­IUD group (n=154)

Index Index (95% CI) P valuea Index (95% CI) P valuea Mann-­Whitney U P value

PI
Initial 1.9 (1.9–2.1) 1.75 (1.2–1.9) 9.61 <0.001
3 mo 1.9 (1.22–2.1) <0.001 1.9 (1.17–1.95) <0.001 1.19 0.232
6 mo 1.9 (1.4–2.1) <0.001 1.9 (1.2–2.1) <0.001 2.16 0.031
RI
Initial 0.75 (0.71–0.78) 0.72 (0.62–0.78) 5.86 <0.001
3 mo 0.77 (0.62–0.8) <0.001 0.77 (0.61–0.8) <0.001 0.947 0.344
6 mo 0.75 (0.62–0.82) <0.001 0.75 (0.6–0.82) <0.001 0.672 0.502

Abbreviations: CI, confidence interval; Cu-­IUD, copper intrauterine device; LNG-­IUS, levonorgestrel-­releasing intrauterine system; PI, pulsatility index; RI,
resistance index.
a
Versus initial value.

Uterine artery PI and RI are summarized in Table 1. There was a sig-

3 | RESULTS
nificant difference between the two groups in initial PI (P<0.001) and
RI (P<0.001), which were both higher among women in the LNG-­IUS
Overall, 330 women met the study criteria and were recruited to
group. In both groups, PI and RI were significantly lower at both 3 and
the trial. Of these, 306 women attended the 6-­month follow-­up
6 months as compared with the initial Doppler indices (all P<0.001).
and completed the study (LNG-­IUS group, n=152; Cu-­IUD group,
Multiple logistic regression analysis of participant characteristics
n=154) (Fig. 1). There was no significant difference between the
and abnormal bleeding revealed a significant correlation with uterine
two groups in age, parity, or body mass index (calculated as weight
artery PI (P<0.001) and RI (P<0.001) at baseline (Table 2). In ROC curve
in kilograms divided by the square of height in meters), as summa-
analysis, uterine artery PI was correlated with abnormal bleeding at a
rized in Table S1.
cutoff value of 1.35 with an area under the curve (AUC) of 0.93, sensi-
The adverse effects and acceptability of the contraceptive meth-
tivity of 88%, and specificity of 100%. Uterine artery RI was correlated
ods reported by the women are summarized in Table S2. A higher
with abnormal bleeding at a cutoff value of 0.62 with an AUC of 0.1,
percentage of women in the Cu-­IUD group reported lower abdom-
sensitivity of 96%, and specificity of 100% (Fig. 2 and Table 3).
inal pain, back pain, and abnormal bleeding at the end of the study
(P<0.001). Fewer women reported abnormal bleeding in the LNG-­IUS
(42/152, 27.6%) than in the Cu-­IUD (79/154, 51.3%) group at the start
of the study (P<0.001). Among those reporting abnormal bleeding, the 4 | DISCUSSION
bleeding was mainly in the form of irregular bleeding (31/42, 74%)
in the LNG-­IUS group and heavy menstrual bleeding (53/79, 67%) in In the present study, abnormal uterine bleeding among new LNG-­IUS

the Cu-­IUD group. In both groups, the number of women reporting users, but not Cu-­IUD users, was found to be positively correlated

abnormal bleeding had decreased at 6 months. Method acceptability with hemodynamic changes in the uterine artery with a significant

in the form of higher compliance with follow-­up, recommendation decrease in the PI and RI.

to other women, and overall satisfaction was higher in the LNG-­IUS A positive effect of LNG-­IUS on the uterine artery PI was first

group (P<0.001). reported by Järvelä et al.,4 who assessed uterine artery PI among
13 symptomatic postmenopausal women who received transder-

T A B L E 2 Multiple logistic regression analysis of participant mal estradiol for 1 month before combining LNG-­IUS with estrogen
characteristics and abnormal bleeding. replacement therapy. In a study of 27 fertile, regularly menstruating
women, the same group subsequently reported that LNG-­IUS inser-
Characteristic X2 P value Odds ratio (95% CI)
tion seemed to increase impedance to uterine arterial blood flow in
Age 0.705 0.401 1.03 (0.96–1.12) the mid-­luteal phase in association with serum LNG and a concomitant
BMI 0.505 0.477 0.97 (0.89–1.05) decrease in serum progesterone.5
Uterine PI 13.21 <0.001 5.86 (2.25–15.2) Similarly, Jiménez et al.6 found that, after controlling for age and
Uterine RI 21.47 <0.001 8.25 (2.79–17.63) parity, the LNG-­IUS was independently associated with increased uter-
Parity 3.23 0.072 0.76 (0.56–1.02) ine artery PI among 27 women, whereas there was no change among
25 women who received a TCU 380A IUD, in a study that assessed PI
Abbreviations: BMI, body mass index (calculated as weight in kilograms
divided by the square of height in meters); CI, confidence interval; PI, pul- just before insertion in the mid-­luteal phase and 3 months after inser-
satility index; RI, resistance index. tion. More recently, Bastianelli et al.7 conducted a longitudinal study
Rezk ET AL.
F I G U R E 2 ROC curve analysis of the association between
abnormal bleeding and uterine artery pulsatility index (PI) and
resistance index (RI).
of 32 women with measurement of uterine artery PI and RI before and
6 months after LNG-­IUS insertion; they concluded that the LNG-­IUS
not only altered endometrial thickness but, among women with pro-
longed bleeding, also significantly changed uterine artery blood flow.
By contrast, Zalel et al.8 reported no changes in blood flow in the uter-
ine artery after LNG-­IUS insertion among 47 women and Cu-­IUD inser-
tion among 35 women when assessed before and after use.
Dane et al.9 found decreased subendometrial vascularization and
higher uterine artery PI in the spiral arteries among 25 new users of
Depo-­medroxyprogesterone acetate injection, but no changes among
25 new LNG-­IUS users. More recently, Cihangir et al.10 reported that
there was no significant change in ovarian volume or uterine artery RI
and PI at 6 or 12 months relative to pre-­insertion values among 49
LNG-­IUS users.
With regard to the Cu-­IUD, most recent studies have reported
an absence of significant changes in uterine artery PI and RI with or
without adverse effects including dysmenorrhea, heavy menstruation,
and dyspareunia at 3 or 6 months after insertion.6,13–15 Heavy men-
strual bleeding associated with the Cu-­IUD might be due to induced
production of vasoactive agents other than prostaglandins in the
surrounding tissue.16
T A B L E 3 Sensitivity and specificity of uterine artery Doppler
indices to predict abnormal bleeding.
Parameter Uterine artery PI Uterine artery RI
AUC 0.93 0.10
Cutoff value 1.35 0.62
Sensitivity, % 88 96
Specificity, % 100 100
Abbreviations: AUC, area under the curve; PI, pulsatility index; RI,
­resistance index.
| 21
The discrepancies among the above studies might be attributed
to the small sample sizes, or to measurement of Doppler indices at
different times of the menstrual cycle or times of the day, or by dif-
ferent ultrasonographers. These confounding factors were avoided in
the current study. Indeed, the study design and the larger number of
participants constitute the main strengths of the study.
The study has some limitations. It was not possible to extend the fol-
low-­up period or to compare other long acting contraceptive methods.
In conclusion, LNG-­IUS-­related abnormal bleeding was associated
with changes in the uterine artery blood flow that were not evident
among Cu-­IUD users. Uterine artery Doppler is a simple, inexpen-
sive, and readily available tool to assess women with abnormal uter-
ine bleeding after insertion of the LNG-­IUS for contraception. Future
research should focus on the predictive value of uterine artery Doppler
indices and subsequent heavy menstruation among new Cu-­IUD users.
AU T HO R CO NT R I B U T I O NS
MR, EE, and A-­ES contributed to project development, data collec-
tion, and manuscript writing. MS contributed to project development,
Doppler studies, and manuscript writing. HM contributed to data col-
lection, statistical analysis, and manuscript writing.
CO NFL I C TS O F I NT ER ES T
The authors have no conflicts of interest.
REFERENCES
1. Winner B, Peipert JF, Zhao Q, et al. Effectiveness of long-­acting
reversible contraception. N Engl J Med. 2012;366:1998–2007.
2. Suvisaari J, Lähteenmäki P. Detailed analysis of menstrual bleeding
patterns after postmenstrual and postabortal insertion of a copper
IUD or a levonorgestrel-­releasing intrauterine system. Contraception.
1996;54:201–208.
3. Momtaz M, Zayed M, Rashid K, Idriss O. Doppler study of the uter-
ine artery in patients using an intrauterine contraceptive device.
Ultrasound Obstet Gynecol. 1994;4:231–234.
4. Järvelä I, Raudaskoski T, Tekay A, Jouppila P. Effect of the
levonorgestrel-­releasing intrauterine system on the uterine artery
pulsatility index in postmenopausal hormone replacement therapy.
Ultrasound Obstet Gynecol. 1997;10:350–355.
5. Järvelä I, Tekay A, Jouppila P. The effect of a levonorgestrel-­releasing
intrauterine system on uterine artery blood flow, hormone concen-
trations and ovarian cyst formation in fertile women. Hum Reprod.
1998;13:3379–3383.
6. Jiménez MF, Vetori D, Fagundes PA, de Freitas FM, Cunha-Filho JS.
Subendometrial microvascularization and uterine artery blood flow
in IUD-­induced side effects (levonorgestrel intrauterine system and
copper intrauterine device). Contraception. 2008;78:324–327.
7. Bastianelli C, Farris M, Rapiti S, Vecchio RB, Benagiano G. Different
bleeding patterns with the use of levonorgestrel intrauterine sys-
tem: Are they associated with changes in uterine artery blood flow?
Biomed Res Int. 2014;2014:815127.
8. Zalel Y, Shulman A, Lidor A, Achiron R, Mashiach S, Gamzu R. The
local progestational effect of the levonorgestrel-­releasing intrauter-
ine system: A sonographic and Doppler flow study. Hum Reprod.
2002;17:2878–2880.
22 | Rezk ET AL.
9. Dane B, Akca A, Dane C, Evcimen S, Cetin A. Comparison of the
effects of the levonorgestrel-­releasing intrauterine system (Mirena)
and depot-­medroxyprogesterone acetate (Depo-­Provera) on suben-
dometrial microvascularisation and uterine artery blood flow. Eur J
Contracept Reprod Health Care. 2009;14:240–244.
10. Cihangir U, Ebru A, Murat E, Levent Y. Mechanism of action of the
levonorgestrel-­releasing intrauterine system in the treatment of
heavy menstrual bleeding. Int J Gynecol Obstet. 2013;123:146–149.
11. Hurskainen R, Teperi J, Paavonen J, Cacciatore B. Menorrhagia and
uterine artery blood flow. Hum Reprod. 1999;14:186–189.
12. Frajndlich R, von Eye Corleta H, Frantz N. Color Doppler sonographic
study of the uterine artery in patients using intrauterine contracep-
tive devices. J Ultrasound Med. 2000;19:577–579.
13. De Souza MA, Geber S. Doppler color flow analysis of the uterine
arteries before and after intrauterine device insertion: A prospective
study. J Ultrasound Med. 2006;25:153.
14. Shen O, Ben Chetrit A, Margalioth EJ, Lavie O, Rabinowitz RR. Effects of
a copper-­medicated intrauterine device on ovarian artery, uterine artery,
and intrauterine blood flow. Gynecol Obstet Invest. 2008;66:253–256.
15. Mutlu I, Demir A, Mutlu MF. Can uterine artery Doppler parame-
ters predict copper intrauterine device-­induced side effects? Eur J
Contracept Reprod Health Care. 2014;19:51–56.
16. Järvelä I, Tekay A, Jouppila P. The effect of diclofenac on uterine
artery blood flow resistance during menstruation in patients with
and without a copper intrauterine device. Hum Reprod. 1998;
13:2480–2483.
S U P P O RT I NG I NFO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section at the end of the article.
Table S1. Characteristics of the study participants.
Table S2. Adverse effects and method acceptability among the
study participants.