Risk factors for intrauterine device embedment in postmenopausal women: an analysis of 731 participants undergoing hysteroscopy

Prior intrauterine device (IUD) removal failure, V-shaped IUDs, and a smaller uterine volume were risk factors for IUD embedment in menopausal women.

Abstract

Objective

This study aimed to explore the risk factors for intrauterine device (IUD) embedment in postmenopausal women.

Methods

A total of 731 women who underwent hysteroscopy for IUD removal from January 2019 to December 2021 were included in the trial, of whom 301 were diagnosed with IUD embedment (embedded group) and 430 had a normal IUD location (control group). The following data were collected from electronic medical records: demographic traits, ultrasound findings, IUD shapes, uterine features, and hysteroscopy findings. Bivariate contingency analysis and multiple logistic regression were performed to identify the significant independent variables that affect IUD embedment.

Results

Prior IUD removal failure (14.3% vs 5.1%; adjusted odds ratio [aOR], 3.29; 95% CI, 1.90-5.71) and V-shaped IUD (18.9% vs 10.7%; aOR, 2.05; 95% CI, 1.33-3.16) were risk factors that were independently linked with IUD embedment. Uterine volume (22.7 [20.9-24.5] cm³ vs 27.9 [24.8-30.9] cm³; aOR, 0.99; 95% CI, 0.982-0.999) was negatively linked with IUD embedment. There were no differences in age, parity, menopause length, years with an IUD in situ, uterine position, or abortion times between the two groups. There were no differences in existing leiomyoma, prior cervical LEEP (loop electrosurgical excision procedure), or myomectomy between the two groups.

Conclusions

Prior IUD removal failure, V-shaped IUD, and a smaller uterine volume were risk factors for IUD embedment in menopausal women.

Intrauterine devices (IUDs) are the most widely used reversible method of contraception.¹ The use of IUD by 41% of women (approximately 141 million) in China drives the high IUD prevalence worldwide.² Between 1982 and 1990, an average of approximately 10 million women nationwide had an IUD placed each year because of China's one-child fertility policy.³ Recent studies have suggested that IUD use has continued to increase, with more than half of women of reproductive age using an IUD over the past 30 years.⁴ However, approximately one fourth of women older than 45 years have not yet had their IUD removed, largely because they are unaware of when an IUD should be removed and have insufficient health education to improve the on-time removal of IUD at menopause.⁵

Ultrasonography is crucial for determining the IUD position and assessing its complications.⁶ The correct position of the IUD is at the uterine fundus, with the arms completely extended toward the uterine cornua and the vertical stem in the uterine cavity. Intrauterine device embedment is defined as the myometrial penetration of the IUD without serosal extension.

Copper IUDs, which are nonhormone devices used by young women, have high continuation rates, provide reliable protection for 10 to 12 years, and allow the rapid restoration of fertility after removal.⁷ Furthermore, after extensive myomectomy or any other procedures that could cause intrauterine adhesions, the use of an IUD is an effective way to prevent adhesions and restore the anatomy of the uterine cavity.^8,9 Levonorgestrel-releasing IUDs are gaining popularity as a method of contraception; they can treat heavy menstrual bleeding, endometriosis-related discomfort, endometrial hyperplasia, and malignancy.^10-12 Thus, IUDs are an appealing contraceptive option for women.

Diagnostic procedures, including bimanual pelvic examinations, ultrasonography, dilation and curettage, and/or laparoscopy, are commonly performed in women with IUD-related disorders.¹³ Hysteroscopy with or without anesthesia is reportedly a feasible method of retrieving or repositioning an IUD, especially in cases in which the strings are not visible or fragmented pieces are located within the uterine cavity or cervical canal.^14-16 The visual dimensions offered by hysteroscopy improve diagnostic accuracy and minimize the limitations associated with other diagnostic procedures, such as blind curettage and hysterosalpingography (HSG).

Although there are a great number of women older than 45 years needing IUD removal in China, little is known as yet about the variables that predispose postmenopausal women to IUD embedment. Uterine retroflexion, congenital uterine malformations, fibroids, symptoms such as bleeding and discomfort, suspected adenomyosis, past Cesarean section, uterine diameter, and endometrial thickness have been linked to an increased risk of IUD malpositioning.^17-19 In this study, we aimed to identify the potential risk factors for IUD embedment and explore the outcomes of postmenopausal women who underwent hysteroscopic IUD removal. We hypothesized that the IUD type, uterine volume, history of IUD removal failure, and existing uterine fibroids would be associated with a higher incidence of IUD embedment.

METHODS

This retrospective study was conducted at a family planning department in a tertiary hospital in Shanghai between January 2019 and December 2021. Approval was obtained from the institutional review board for this study (GKLW-202255; December 5, 2022), and the need for informed consent was waived. Postmenopausal women who underwent hysteroscopic removal of retained or fragmented IUD were included in the study. The exclusion criteria were as follows: (1) nonmenopausal women; (2) pharmacotherapeutically induced menopause in women with breast cancer; (3) women with IUD perforation, ectopic IUD, and surgical failure; and (4) women with severe acute and chronic disorders who could not withstand hysteroscopy.

Data on sociodemographics, reproductive history, previous surgical history, bimanual pelvic examination findings, ultrasound results, operative notes, and surgical findings were acquired from medical records. Variables included age; gravidity and parity; menopause duration; the duration of an IUD in situ; previous vaginal delivery; abortion times; previous surgical history (including Cesarean section, LEEP [loop electrosurgical excision procedure] myomectomy, IUD removal failure); other complaints (including uterine bleeding and intrauterine fluid); cervical atrophy; uterine anomalies (including septate uterus, fibroids, and adenomyosis); postabortion insertion or postpartum insertion within 6 months; uterine position; IUD shape; uterine volume; endometrial thickness; and the location of the embedded IUD. Intrauterine device placement immediately after dilation and curettage was considered postabortion insertion. Transvaginal ultrasonography was used to calculate the uterine volume using the following formula: volume = 0.5233 × D1 (longitudinal dimension) × D2 (anteroposterior dimension) × D3 (transverse dimension) cm³.²⁰ Before surgery, speculum and bimanual pelvic examinations were performed to assess the cervix, uterus, and IUD string. In addition, ultrasonography was performed to assess the intrauterine condition of the IUD. If necessary, HSG was performed to understand the relationship between the IUD and the endometrial and myometrial layers.

A gynecologist with substantial experience with the procedure used a hysteroscope (Olympus, Tokyo, Japan) to perform hysteroscopic IUD retrieval. With an empty bladder, the participant was examined in the lithotomy position. After general anesthesia, the uterine cavity was dilated with glycine or saline solution at a flow rate of 200 to 300 mL/min and a pressure of 13.3 to 16.0 kPa. If the cervical adhesions were dense, the cervix was dilated gradually with Hegar's dilator to no. 7.5 under ultrasound guidance. When the device was located, the IUD was carefully extracted using Alligator forceps or a ring hook. If the IUD fragments were small, the grasping forceps were inserted through a metal sleeve to grasp and remove the fragments. If the uterine cavity adhesions were dense, miniature scissors were used to break the adhesions and restore the anatomy, followed by IUD retrieval. If necessary, misoprostol or other prostaglandins were used for cervical ripening.

Statistical analysis

All statistical analyses were conducted using IBM SPSS (version 29; SPSS, Chicago, IL). χ², Fisher exact, and independent t tests were performed. Calculations of crude odds ratios (OR) and 95% CI were conducted using logistic regression analysis. For uterine volume, a continuous variable, OR indicated an association between a 1-cm³ difference in uterine volume and the odds of IUD embedment. Odds ratio indicated the association of each one abortion difference with the odds of IUD embedment. To ascertain significant independent predictors of IUD embedment, variables having a link to IUD embedment (P < 0.2) were added in multivariable logistic regression. A P value of 0.05 was deemed statistically significant.

RESULTS

There were 760 postmenopausal women who visited the family planning clinic for the removal of a retained IUD. Of the 760 participants, 29 were excluded from the study; 6 women had low-lying IUD, 2 had perforations, 2 experienced removal failure due to dense vaginal and cervical adhesions, 3 had an IUD in the endocervical canal, and 16 had a fractured residual IUD without embedment (Fig. 1). Ultimately, 731 women were included in this study. Of these women, 430 had a normally located IUD, while 301 had an embedded IUD.

FIG. 1.

Flowchart of the study. IUD, intrauterine device.

Table 1 shows the baseline demographic characteristics and surgical outcomes of participants in the embedded and control groups. The menopause duration among the women in the embedded and control groups ranged from 1 to 40 years (mean, 7.7 y) and 1 to 38 years (mean, 8.2 y), respectively. Intrauterine devices were in situ for an average of 28.5 years (9-52 y) and 28.7 years (5.5-50 y) in the embedded and control groups, respectively. There were no differences in age, parity, menopausal duration, years with an IUD in situ, or uterine position between the embedded and control groups. There were no differences in endometrial thickness, uterine bleeding, intrauterine fluid, or history of cervical LEEP or myomectomy. We found that women in the embedded group had more abortions than those in the control group (1.3 ± 0.88 vs 1.2 ± 1.0, P = 0.022). No IUD were placed at the time of delivery. The uterine volumes of the embedded group were significantly smaller than those of the control group (22.7 cm³ [95% CI, 20.9-24.5] vs 27.9 cm³ [95% CI, 24.8-30.9], P = 0.007). Furthermore, factors that were found to be associated with IUD embedment in the bivariate analysis were prior IUD removal failure (P < 0.001), cervical atrophy (P = 0.01), and IUD shape (P = 0.004).

Table 1.

Baseline demographic factors and surgical outcomes of the embedded and control groups

Parameter	Control (n = 430)	Embedded (n = 301)	P
Age, mean (SD), y	58.9 (6.1)	58.5 (5.9)	0.32
95% CI	58.3-59.5	57.8-59.1
Nulliparous, n (%)	4 (0.9)	3 (1.0)	0.99
Multiparous, n (%)	426 (99.1)	298 (99.0)
Gravida, mean (SD)	2.4 (1.2)	2.4 (1.0)	0.112
95% CI	2.25-2.48	2.31-2.54
Prior vaginal delivery, n (%)	303 (70.5)	212 (70.4)	0.997
Abortion time, mean (SD), n (%)	1.2 (1.0)	1.3 (0.88)	0.022
Abortion ≥3, n (%)	33 (7.7)	26 (8.6)	0.64
Menopause duration, mean (SD), y	8.2 (6.5)	7.7 (5.9)	0.564
95% CI	7.5-8.8	7.0-8.4
Menopause duration >1 y, n (%)	374 (87.0)	267 (88.7)	0.484
Duration IUD in situ, mean (SD), y	28.7 (7.3)	28.5 (7.0)	0.675
Range	5.5, 50.0	9, 52
95% CI	28-29.4	27.7-29.3
Previous surgery, n (%)
Prior Cesarean section	123 (28.6)	86 (28.6)	0.992
Prior cervical excision (LEEP)	6 (1.4)	4 (1.3)	0.939
Prior myomectomy	2 (0.5)	2 (0.7)	0.719
Prior IUD removal failure	22 (5.1)	43 (14.3)	<0.001
Other complaints, n (%)
Uterine bleeding	28 (6.5)	20 (6.6)	0.943
Intrauterine fluid	103 (24.0)	74 (24.6)	0.85
Cervical atrophy, n (%)	223 (51.9)	185 (61.5)	0.01
Uterine anomaly, n (%)
Septate uterus	0	1 (0.3)	0.41
Fibroids	99 (23.0)	57 (18.9)	0.184
Adenomyosis	0	1 (0.3)	0.41
Postabortion insertion, n (%)	78 (18.1)	52 (17.3)	0.76
Postpartum insertion within 6 mo, n (%)	6 (1.4)	4 (1.3)	0.939
Uterine position, n (%)			0.89
Anteverted	245 (57.0)	166 (55.1)
Neutral	31 (7.2)	23 (7.6)
Retroverted	154 (35.8)	112 (37.2)
IUD shapes, n (%)			0.004
Ring	335 (77.9)	216 (71.8)	0.058
V	46 (10.7)	57 (18.9)	0.002
γ	16 (3.7)	4 (1.3)	0.051
T	14 (3.3)	15 (5.0)	0.239
Others (including merina)	19 (4.4)	9 (3.0)	0.322
Uterine volume, cm3	27.9 (24.8-30.9)	22.7 (20.9-24.5)	0.007
Endometrial thickness, mm	3.09 (1.7)	3.1 (1.5)	0.926
Location of embedded IUD
Fundal	—	64
Corner	—	5
Cervical	—	33
Lateral	—	184
Submucosal fibroid	—	2
Unclear	—	13

95%CI, 95% confidence intervals; IUD, intrauterine device; LEEP, loop electrosurgical excision procedure; SD, standard deviation.

In the multivariable logistic regression analysis (Table 2), risk factors independently associated with IUD embedment were prior IUD removal failure (adjusted odds ratio [aOR], 3.29; 95% CI, 1.90-5.71) and V-shaped IUD (aOR, 2.05; 95% CI, 1.33–3.16). Moreover, uterine volume was a protective factor; as the uterine volume increased by 1 cm³, the embedment risk decreased by 1% (aOR, 0.99; 95% CI, 0.982-0.999).

Table 2.

Multivariable logistic regression analysis for independent risk factors associated with IUD embedment

	aOR (95% CI)	P
Abortion times, per 1 abortion	1.145 (0.978–1.341)	0.092
Cervical atrophy
No	Reference	—
Yes	1.33 (0.973–1.817)	0.074
Prior IUD removal failure
No	Reference	—
Yes	3.29 (1.90–5.71)	<0.001
Ring-shaped IUD	Reference	—
V-shaped IUD	2.05 (1.33–3.16)	0.001
γ-shaped IUD	0.419 (0.136–1.29)	0.130
T-shaped IUD	1.615 (0.743–3.509)	0.226
Other Shapes	0.676 (0.290–1.577)	0.365
Uterine volume, per 1 cm3	0.99 (0.982–0.999)	0.024

aOR, adjusted odds ratio; 95%CI, 95% confidence intervals; IUD, intrauterine device.

When comparing the two largest fibroids in the two groups, there were no statistically significant differences in the location, diameter, or type of fibroids (P > 0.05) between the embedded and control groups (Table 3). Among the 731 women, none experienced secondary IUD perforation or other complications, such as bladder or rectal injuries. Hysteroscopy and laparoscopy were performed simultaneously in one woman in the embedded group because the IUD was found to be embedded into a 6-cm submucosal fibroid upon hysteroscopy. A histopathological diagnosis was obtained for all 731 women; one woman was diagnosed with endometrial cancer and was started on appropriate treatment.

Table 3.

Comparison of the fibroid characteristics between the embedded and control groups

Fibroid characteristics	Control (n = 430)	Embedded (n = 301)	P
Mean no. of fibroids (95% CI)	0.3 (0.24-0.36)	0.22 (0.17-0.28)	0.156
	Fibroid 1 mm (95% CI)	Fibroid 1 mm (95% CI)
Mean fibroid size	6.6 (5.1-7.9)	4.7 (3.4-6.0)	0.154
Fibroid location	Fibroid 1 n (%) n = 99	Fibroid 1 n (%) n = 57	0.386
Anterior	52 (53)	22 (39)	0.093
Posterior	27 (27)	16 (29)	0.915
Fundal	4 (4)	3 (5)	0.722
Lateral	14 (14)	14 (25)	0.102
Cervical	2 (2)	1 (2)	>0.99
None	331 (77)	244 (81)
Fibroid type	Fibroid 1 n (%) n = 99	Fibroid 1 n (%) n = 57	0.327
Intramural	84 (85)	45 (79)	0.384
Subserosal	11 (11)	11 (19)	0.157
Submucosal	4 (4)	1 (2)	0.653
Pendunculated	0	0	—

95%CI, 95% confidence intervals; Fibroid 1, the largest fibroid of each group.

DISCUSSION

In this study, we found a cumulative IUD embedment rate of 41.2% among women undergoing hysteroscopy. In addition, there was a statistically significant increase in the presenting V-shaped IUD and previous IUD removal failure in women with embedded IUDs versus controls. Furthermore, women with embedded IUDs had smaller uterine volumes than those with normally placed IUDs. There was no statistically significant difference among the groups regarding the size, location, and type of fibroids.

A partially embedded IUD could signify migration or transfer to an extrauterine location and thus create a potential pathway for infection. The participants could seek medical attention for stomach discomfort, unusual bleeding, pregnancy, or the inability to feel the cervical marker.²¹ Intrauterine device removal in postmenopausal women is difficult because of decreased levels of circulating estrogen, a smaller uterus, and cervical atrophy. Thus, the IUD should be removed 6 to 12 months after menopause. The biggest challenge of hysteroscopy, especially among postmenopausal women, is placing the hysteroscope sheath into the cervical canal, specifically, the internal os. The forceps can grab a section of the device when placed through the metallic sleeve holding the hysteroscope; the device provides minimal resistance when being removed. The need for laparoscopy depends on preoperative radiological and gynecological examinations. Women with cervical atrophy were treated with misoprostol for cervical ripening before hysteroscopy at our center.²² The benefit of IUD removal via hysteroscopy in postmenopausal women is that the IUD position in relation to the uterus can be precisely analyzed. If deemed safe, the IUD can be removed with a lower chance of subsequent perforation or complications. In addition, hysteroscopy aids in determining the need for additional laparoscopy to ensure a safe approach for removal.¹⁴ At our facility, hysteroscopic IUD removal with Alligator forceps or a ring hook was attempted in the operating room for all postmenopausal women. The results showed that the IUD was inserted into the lateral uterine cavity in 61.1% of the participants. Most removals were not difficult and did not require any specialized procedures. Moreover, no other complications were observed. In one women in this series, hysterography and laparoscopy were performed concurrently because the IUD was embedded in a submucosal fibroid. Thus, our study confirmed the value and safety of hysteroscopic IUD removal in postmenopausal women.

This was the first study to show that prior IUD removal failure must be considered as a risk factor for IUD embedment. Forty-three women (14.3%) in the embedded group and 22 (5.1%) in the control group reported previous IUD removal failure. Furthermore, five women (2 in the embedded group and 3 in the control group) had more than one episode of IUD removal failure in the clinic. Therefore, hysteroscopy is recommended for women who have experienced IUD removal failure instead of undergoing dilation and curettage.

We also found that V-shaped IUDs may be related to a greater risk of IUD embedment, even though the most frequently removed IUD in the study were ring shaped (75.4%). This may be indicative of the fact that V-shaped IUDs have been used since the 1980s and hence have a larger incidence in the population; thus, their intrinsic quality might contribute to the increased risk of embedment in the Chinese population. However, Gerkowicz et al¹⁸ and Merki-Feld et al²³ found that American women in the reproductive age group frequently had malpositioned ParaGard IUD.

Furthermore, we demonstrated that women with embedded IUDs had a smaller uterine volume than those with a normally placed IUD. Shipp et al¹⁹ previously demonstrated that embedded IUDs were associated with a reduced endometrial cavity width of the fundus. We observed no discernible differences in the gravidity, delivery methods, or years with an IUD in situ between the two groups. While others have demonstrated that a retroflexed uterus is linked to a greater prevalence of IUD malalignment, the association between uterine positioning and embedment was not significant in our study.¹⁸

Studies have linked early postpartum and immediate postabortion insertions to a greater chance of expulsion; however, no such studies have been conducted to identify IUD embedment in women with such insertions.^24,25 Immediate postabortion IUD insertion was not identified as a risk factor for embedment in our study, demonstrating that immediate postabortion IUD insertion was safe over a long period. However, the trimester in which the abortion occurred was not considered. The safety of IUD implantation within half a year after delivery is evidenced by the fact that there was no link between IUD embedment and IUD implantation 6 months after delivery. However, the timing of IUD placement in relation to nursing care or delivery is unknown. Similar to the study by Braaten et al,²⁶ no significant correlations between IUD embedment and fibroids were identified in our study. However, Gerkowicz et al¹⁸ demonstrated that the presence of more fibroids and submucosal fibroids was associated with IUD malpositioning.

To our knowledge, no studies have examined the prevalence of IUD embedment in postmenopausal women undergoing hysteroscopy. The current study is the largest to evaluate parameters related to IUD embedment. Because the study was conducted at a single center, the potential for misevaluation was reduced by having only two clinicians carry out the examinations. Our results must be interpreted within the context of the study design. This was a retrospective study that was exploratory in nature and constrained by participant characteristics noted during the relevant period. A multicenter study with a larger sample size is required to obtain more definitive results.

CONCLUSIONS

Prior IUD removal failure, V-shaped IUD, and smaller uterine volumes were independent risk factors for IUD embedment in postmenopausal women.