Abstract
Objectives
To evaluate the relationship between intrauterine device (IUD) malposition and patient‐reported symptoms, identify risk factors associated with malposition, and assess the reliability of string checks compared with ultrasound findings.
Methods
A retrospective chart review was conducted at West Virginia University between January 2021 and December 2023. Patients with malpositioned IUDs identified by pelvic ultrasound were included. Data collected included demographics, body mass index (BMI), IUD type, provider type, anatomical abnormalities, symptom profiles, and string check findings. Descriptive statistics were used to summarize the data, and associations with types of IUD malposition were assessed in the statistical analysis.
Results
A total of 175 patients with malpositioned IUDs were identified. Approximately half of the patients were symptomatic (53.1%), most commonly reporting pelvic pain or abnormal uterine bleeding, while 46.9% were asymptomatic. Low intrauterine placement was the most frequent malposition pattern. String visualization was common but did not reliably exclude malposition. No demographic, device, or provider characteristics were significantly associated with specific malposition types.
Conclusions
Nearly half of patients with malpositioned IUDs were asymptomatic, and visible strings were frequently present, highlighting the limitations of symptom assessment and string checks. Ultrasound plays a central role in evaluating IUD position when clinical concern exists.
Keywords: abnormal uterine bleeding, contraception, intrauterine device, malposition, ultrasound
Abbreviations
- AIUM
American Institute of Ultrasound
- BMI
body mass index
- CPT
current procedural terminology
- IUD
intrauterine device
Intrauterine devices (IUDs) are among the most effective forms of reversible contraception, with failure rates below 1% and high long‐term continuation and patient satisfaction. As IUD utilization has increased, complications related to device position have become increasingly relevant in clinical practice. IUD malposition—commonly defined as nonfundal location, low intrauterine or cervical positioning, or myometrial embedment—has been reported in approximately 2–10% of insertions, with wide variability depending on timing of evaluation and diagnostic modality. 1 , 2 , 3 Prior studies suggest that malposition may be associated with pelvic pain, abnormal uterine bleeding, expulsion, and potentially reduced contraceptive efficacy, particularly for copper‐containing devices; however, other investigations report asymptomatic malposition detected incidentally on imaging, leading to uncertainty regarding clinical significance and optimal management. 2 , 3 , 4
Pelvic ultrasound remains the gold standard for assessment of IUD position and enables detection of nonfundal placement and embedment that may not be appreciated on physical examination or string assessment. 3 , 4 , 5 In routine practice, follow‐up commonly relies on patient‐ or clinician‐performed string checks, despite limited evidence regarding their accuracy in predicting appropriate device position. Additionally, data evaluating patient‐, device‐, and provider‐related factors associated with IUD malposition are limited, and few studies have simultaneously examined symptom presentation, string check findings, and ultrasound‐confirmed malposition within the same cohort. 2 , 4 , 6 The purpose of this study was to evaluate symptom patterns among patients with ultrasound‐confirmed IUD malposition, characterize malposition types, and assess the reliability of string checks compared with imaging findings.
Methods
This is a retrospective, referral‐based cohort study. The study population included all patients who underwent transvaginal ultrasound with a primary or secondary indication of IUD evaluation at West Virginia University clinics between January 2021 and December 2023.
This study was approved by the Internal Review Board of West Virginia University and included a HIPAA waiver due to the retrospective nature of the study. Studies were identified by a query of our electronic medical records (EPIC) using the current procedural terminology (CPT) code 76830 (some patients also had a CPT 76586) and described that the IUD was mispositioned.
Patients were categorized as “Symptomatic” if they reported pelvic pain or abnormal bleeding, and “Asymptomatic” if the ultrasound was prompted by non‐pain indications (eg, missing strings, incidental finding during fibroid evaluation).
Adequate positioning of the IUD was defined with the following criteria: the IUD shaft midline within the endometrial cavity with arms fully deployed at 90° and the top portion of the IUD should be no lower than 5 mm from the fundus of the cavity. The distance between the superior aspect of the IUD and the uterine fundus was measured in millimeters when available from the radiology report or determined on image review.
For the purposes of this study, a fundal IUD was defined as a device located ≤5 mm from the uterine fundus. IUDs positioned ≥6 mm from the fundus were classified as low‐lying or displaced. This threshold was selected to identify even subtle deviations from fundal placement that may be associated with clinical symptoms, including pelvic pain and abnormal uterine bleeding, while still maintaining an intrauterine location.
Mispositioning was defined in different categories including, embedded in the myometrium, perforation, low sitting (lower uterine segment or endocervical), misaligned and expulsion. For being embedded in the myometrium, we either consider partial or full penetration within the myometrium. A “low sitting IUD” was defined if the top of the IUD was lower than 5 mm from the fundus of the uterus. Penetration was defined as an IUD that was not found within the uterus, and it was found in the pelvic or abdominal cavity. Expulsion was defined as the absence of IUD in the uterus, abdomen or pelvis (findings corroborated by x‐rays or computed tomography) and patient witnessing the expulsion of the device.
Data from the IUD placement was retrieved via electronic medical records. The IUD was inserted by either physicians (attendings or resident physicians under the supervision of a faculty member) or advance practitioners (nurse practitioners or physician's assistant). All the following types of IUD were included in the study (Paragard, CooperSurgical, Inc., Trumbull, CT), levonorgestrel‐releasing 52 mg IUD (Mirena, Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ), levonorgestrel‐releasing 52 mg IUD (Liletta, Odyssea Pharma, SPRL, Belgium), levonorgestrel‐releasing 13.5 mg IUD (Skyla, Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ), and levonorgestrel‐releasing 19.5 mg IUD (Kyleena, Bayer HealthCare Pharmaceuticals Inc., Whippany, NJ).
All the ultrasounds were performed by sonographers certified by the Registry of Diagnostic Medical Sonography and all sites were accredited by the American Institute of Ultrasound (AIUM). A routine of 2D and 3D ultrasound following the AIUM guidelines was performed using GE Voluson E8 and E10 ultrasound machines. All the readings were provided by 4 board‐certified obstetricians and gynecologists.
Demographic data was collected via electronic medical records, including age, race, body mass index (BMI), obstetrical history, type of IUD, if a string check was done after the IUD placement, interval between symptoms and IUD insertion. Data from IUD insertion was also collected including correlation with pregnancy. Postpartum IUDs were considered as IUD placed either post placenta or within the first 6 weeks postpartum. Patient's symptoms were recorded including pelvic pain, discomfort, abnormal uterine bleeding or expulsion of the device.
Our primary outcome was to determine the most common symptoms in patients with a mispositioned IUD.
Secondary outcomes include assessing associations between IUD mispositioning and patient characteristics, including type of IUD, anatomical/structural abnormalities, BMI of the patients, and provider type.
Tertiary outcomes include determining the proportion of patients who had a normal string check and had a mispositioned IUD.
Statistical Analysis
Descriptive statistics were used to summarize the characteristics of patients. Categorical data were described using contingency tables with counts and percentages. Continuous variables were summarized using mean with standard deviation. Chi‐square test and Fisher's exact test were used to assess the independence between 2 categorical variables. Wilcoxon rank sum test or t‐tests were used to assess the difference in continuous variables, including age, gravida, BMI, uterine length, uterine width, and uterine height. Statistical inferences were based on 2‐sided tests at a significance level of p < .05. Statistical analyses were carried out using SAS 9.4 (SAS Institute, Cary, NC) and R (version 3.6.3, R Foundation, Vienna, Austria).
Results
One hundred and seventy‐five subjects were identified in the study, and 1 patient had incomplete data and was excluded from the analysis. The strings of the IUD were visualized in 116 patients (66.7%); 13 (7.5%) had no strings visualized on exam, and 45 (25.9%) had no string check.
Demographic characteristics, including age, race, BMI, prior cesarean section, and gravidity, were similar across groups, with no statistically significant differences (all p > .05); this is represented in Table 1. Notably, patients with non‐visualized strings tended to have a higher BMI (61.5% with BMI ≥30) compared to those with string visualization (45.7%) or no string check (64.4%), though this did not reach statistical significance (p = .08).
Table 1.
Demographic and Uterine Characteristics Stratified by IUD String Visualization Status
| Overall (N = 175) | IUD String Check | p‐Value | |||
|---|---|---|---|---|---|
| String Visualized (N = 117, 66.9%) | Strings Not Visualized (N = 13, 7.4%) | No String Check Done (N = 45, 25.7%) | |||
| Race | 1.00 | ||||
| Others | 12 (6.8%) | 8 (6.8%) | 1 (7.6%) | 3 (6.6%) | |
| White | 163 (93.1%) | 108 (92.3%) | 12 (92.3%) | 42 (93.3%) | |
| Prior C‐section | .98 | ||||
| No prior C‐section | 121 (69.1%) | 82 (70%) | 9 (69.2%) | 30 (66.6%) | |
| C‐section = 1, 2, or 3 | 54 (30.9%) | 35 (30%) | 4 (30.7%) | 15 (33.3%) | |
| BMI | .08 | ||||
| ≤29 | 84 (48%) | 63 (54.9%) | 5 (38.4%) | 16 (35.5%) | |
| ≥30 | 91 (52%) | 54 (46.1%) | 8 (61.5%) | 29 (64.4%) | |
| Age | .62 | ||||
| Mean (SD) | 29.1 (8.4) | 29.5 (8.6) | 27.4 (6.3) | 28.5 (8.7) | |
| Gravida | .63 | ||||
| Mean (SD) | 1.7 (1.7) | 1.7 (1.7) | 1.3 (1.3) | 1.8 (2) | |
| Any previous cervical procedures | 1.00 | ||||
| Yes | 5 (2.9%) | 5 (4.3%) | 0 (0%) | 1 (2.2%) | |
| No | 170 (97.1%) | 112 (95.7%) | 13 (100%) | 44 (97.7%) | |
| Uterine length | .82 | ||||
| Mean (SD) | 7.79 (1.53) | 7.77 (1.6) | 8.05 (1.2) | 7.76 (1.2) | |
| Uterine width | .47 | ||||
| Mean (SD) | 4.42 (1.0) | 4.50 (0.9) | 4.32 (1.0) | 4.29 (1.0) | |
| Uterine height | .08 | ||||
| Mean (SD) | 4.19 (1.0) | 4.30 (1.0) | 3.69 (0.7) | 4.07 (0.9) | |
String visualization refers to pelvic examination performed at the time of presentation prompting ultrasound evaluation.
Data from the IUD is represented in Table 2. The most common IUD was Mirena IUD (57.23%) and the most common indication for the IUD placement was contraception (89%). Most of the patients were diagnosed with either mispositioned IUD or complication of IUD at 6 months post insertion (69.23%).
Table 2.
IUD and Procedural Characteristics at the Time of Insertion Stratified by String Visualization Status
| Overall (N = 175) | IUD Information | p‐Value | |||
|---|---|---|---|---|---|
| String Visualized (N = 117, 66.9%) | Strings Not Visualized (N = 13, 7.4%) | No String Check Done (N = 45, 25.7%) | |||
| Cervix dilated at insertion | 1.00 | ||||
| Yes | 9 (5.1%) | 7 (5.9%) | 0 (0%) | 2 (4.44%) | |
| No | 166 (94.9%) | 110 (94.1%) | 13 (100%) | 43 (95.56%) | |
| Indication for the IUD | .37 | ||||
| Abnormal uterine bleeding | 20 (11.4%) | 13 (11.1%) | 0 (0%) | 7 (15.5%) | |
| Contraception | 155 (88.5%) | 104 (88.9%) | 13 (100%) | 38 (84.4%) | |
| Type of IUD | .12 | ||||
| Kylena | 10 (5.8%) | 5 (4.3%) | 1 (7.6%) | 4 (8.8%) | |
| Liletta | 17 (9.8%) | 9 (7.7%) | 0 (0%) | 8 (17.7%) | |
| Mirena | 100 (57.1%) | 66 (56.5%) | 8 (61.5%) | 26 (57.7%) | |
| Paraguard | 43 (24.5%) | 34 (29%) | 3 (23%) | 6 (13.3%) | |
| Skyla | 5 (2.8%) | 3 (2.5%) | 1 (7.6%) | 1 (2.2%) | |
| Prior IUD | .27 | ||||
| Yes | 36 (20%) | 20 (17%) | 4 (30.7%) | 11 (24.4%) | |
| No | 140 (80%) | 97 (83%) | 9 (69.2%) | 34 (75.5%) | |
| Was another IUD inserted | .83 | ||||
| Yes | 36 (20.6%) | 25 (21.4%) | 3 (23%) | 8 (17.8%) | |
| No | 139 (79.4%) | 92 (78.6%) | 10 (77%) | 37 (82.2%) | |
| Type of provider who inserted the IUD | .11 | ||||
| APP | 63 (36%) | 42 (35.9%) | 8 (61.6%) | 13 (28.9%) | |
| MD | 112 (64%) | 75 (64.1%) | 5 (38.4%) | 32 (71.1%) | |
The relationship between IUD malpositioning and symptoms is explored in Table 3. In the analysis of 174 patients with confirmed IUD mispositioning, 53.1% were symptomatic. Pain, irregular bleeding, heavy flow, expulsion, and difficulty with removal were evaluated as individual symptoms. Pain was reported in 54.6% of patients with myometrial embedment, 55.1% of patients with low‐positioned IUDs, and 77.8% of those with uterine perforation (p = .07). Expulsion of the IUD was primarily self‐reported in 72.2% of patients diagnosed with expulsion, with none of the embedded or perforated cases reporting this (p < .0001). Difficulty with removal in clinic was significantly more common among patients with embedded (36.4%) and malrotated (25.0%) IUDs (p = .0001) as expected.
Table 3.
Type of IUD Malposition and Associated Clinical Presentation
| Variable | Overall (N = 175) | Type of IUD Malpositioning | p‐Value | ||||
|---|---|---|---|---|---|---|---|
| Embedded in the Myometrium (N = 33, 18.9%) | Expulsion (N = 18, 10.2%) | Low in the Uterine Cavity (N = 78, 44.6%) | Perforation (N = 18, 10.3%) | Malrotated (N = 28, 16%) | |||
| Was a string check done after the IUD insertion | .57 | ||||||
| No | 50 (28.5%) | 10 (30.3%) | 5 (27.7%) | 19 (24.4%) | 8 (44.4%) | 8 (28.6%) | |
| Yes | 125 (71.4%) | 23 (69.7%) | 13 (72.3%) | 59 (75.6%) | 10 (55.6%) | 20 (71.4%) | |
| Patient was symptomatic | .07 | ||||||
| Yes | 93 (53.1%) | 18 (54.5%) | 6 (33.3%) | 43 (55.1%) | 14 (77.8%) | 12 (42.9%) | |
| No | 82 (46.8%) | 15 (45.5%) | 12 (66.7%) | 35 (44.9%) | 4 (22.2%) | 16 (57.1%) | |
| Patient complained of irregular bleeding | .13 | ||||||
| Yes | 47 (26.8%) | 6 (18.1%) | 2 (11.2%) | 27 (34.6%) | 3 (16.7%) | 9 (32.1%) | |
| No | 128 (73.1%) | 27 (81.9%) | 16 (88.8%) | 51 (65.4%) | 15 (83.3%) | 19 (67.9%) | |
| Was the patient complaining of heavy flow | .36 | ||||||
| Yes | 22 (12.5%) | 5 (15.1%) | 0 (0%) | 11 (14.1%) | 1 (5.5%) | 5 (17.9%) | |
| No | 153 (87.4%) | 28 (84.9%) | 18 (100%) | 67 (85.9%) | 17 (94.4%) | 23 (82.1%) | |
| Patient noticed the expulsion of the device | <.001 | ||||||
| Yes | 16 (9.1%) | 0 (0%) | 13 (72.3%) | 2 (2.5%) | 0 (0%) | 1 (3.6%) | |
| No | 159 (90.8%) | 33 (100%) | 5 (27.7%) | 76 (97.5%) | 18 (100%) | 27 (96.4%) | |
| The device was unable to be removed in clinic | <.001 | ||||||
| Yes | 26 (14.8%) | 12 (36.3%) | 0 (0%) | 4 (5.1%) | 3 (16.7%) | 7 (25%) | |
| No | 149 (85.1%) | 21 (63.7%) | 18 (100%) | 74 (94.9%) | 15 (83.3%) | 21 (75%) | |
| Strings visualized on exam | .31 | ||||||
| Not visualized | 40 (22.9%) | 6 (18.1%) | 2 (11.2%) | 17 (21.8%) | 7 (38.9%) | 8 (28.6%) | |
| Visualized | 135 (77.1%) | 27 (81.9%) | 16 (88.8%) | 61 (78.2%) | 11 (61.1%) | 20 (71.4%) | |
Note: Bold showed statistical significance.
Low uterine cavity positioning was the most frequent mispositioning type (44.6%), followed by myometrial embedment (18.9%), malrotation (16.0%), perforation (10.3%), and expulsion (10.3%). Symptom presentation varied by mispositioning type, with significant associations for expulsion and removal difficulty. When evaluating the degree of displacement, the vast majority of malpositioned IUDs in our cohort were located ≥20 mm below the fundus. Only 2 patients (1.1%) demonstrated a fundal‐to‐IUD distance between 5 and 20 mm. Given this distribution, the 5 mm threshold effectively captured nearly all clinically relevant displacements in this study population.
Discussion
In this retrospective cohort of 175 patients with ultrasound‐confirmed IUD malposition, we identified a substantial discordance between clinical presentation and imaging findings. Nearly half of patients (46.9%) were asymptomatic despite documented malposition, underscoring the limited diagnostic value of symptoms alone in excluding abnormal device position. This finding is consistent with prior ultrasound‐based studies demonstrating that malposition may be incidentally detected and may not reliably correlate with patient‐reported symptoms. 1 , 3 Together, these data support maintaining a low threshold for pelvic ultrasound when IUD position is clinically relevant.
Although a large proportion of patients were asymptomatic, symptom patterns varied by malposition subtype. Pelvic pain, abnormal uterine bleeding, and expulsion were associated with specific malposition categories, emphasizing that not all malpositions carry equivalent clinical implications. Difficulty with IUD removal clustered with myometrial embedment and malrotation, highlighting the importance of defining malposition subtype on imaging to guide counseling and management. Prior literature similarly describes a spectrum of malposition and migration patterns detectable on imaging, each with distinct clinical considerations. 1 , 3
Low intrauterine placement was the most common malposition pattern in this cohort, consistent with previous reports identifying the lower uterine segment as a frequent location for displaced IUDs. 1 , 3 Evidence suggests that low‐placed IUDs may have altered contraceptive performance and may be associated with higher rates of expulsion, depending on device type. 7 These findings reinforce the clinical importance of confirming fundal placement and support the role of ultrasound in evaluating suspected malposition.
String visualization was documented in a majority of patients; however, visible strings did not exclude malposition. All patients with visible strings had ultrasound‐confirmed malposition, demonstrating the limited sensitivity of string checks for confirming correct device location. These findings align with prior studies showing that string presence does not reliably correlate with fundal placement and that physical examination alone may fail to detect displacement or embedment. 1 , 8 When localization is uncertain or removal is anticipated, imaging remains essential, with ultrasound as the preferred initial modality and radiography as a complementary tool when ultrasound findings are inconclusive. 6
No demographic, device‐related, or provider‐related characteristics were significantly associated with specific malposition types in this cohort. While prior population‐based studies have identified risk factors for serious complications such as uterine perforation, these predictors may not directly translate to nonperforating malposition patterns detected on ultrasound. 5 , 9 The absence of strong predictors may reflect a true lack of association or limited power within subgroup analyses.
This study has limitations. The retrospective design and inclusion of only patients with ultrasound‐confirmed malposition introduce selection bias and preclude estimation of malposition incidence. Additionally, documentation variability may affect symptom classification. While some clinical guidelines suggest a 20‐mm threshold to define malposition primarily to ensure contraceptive efficacy, our data indicates that most symptomatic displacements already exceed this distance. By utilizing a 5‐mm threshold, we were able to include the small subset of patients with minor displacement who nonetheless presented with symptoms. However, we acknowledge that for asymptomatic patients, a more conservative 20‐mm cutoff may be appropriate to avoid unnecessary intervention. Our findings are subject to selection bias, as the study denominator includes only those patients whose clinical presentation warranted sonographic evaluation. Consequently, the rates of displacement reported here likely overestimate the prevalence in the general population but accurately reflect the diagnostic yield in a clinical “at‐risk” cohort. Strengths include consistent use of pelvic ultrasound interpreted by board‐certified obstetrician–gynecologists and standardized sonographic criteria for IUD positioning, consistent with prior validation studies. 1 , 8
Conclusion
In this retrospective ultrasound‐based study of 175 patients with IUD malposition, nearly half of affected patients were asymptomatic, and visible strings were frequently present, demonstrating that absence of symptoms and presence of strings cannot reliably exclude malposition. These findings argue against reliance on symptom status or string checks alone as reassurance of correct IUD positioning. Pelvic ultrasound plays a central role in evaluating IUD location and should be considered when symptoms occur, when string findings are abnormal, when placement is uncertain, or when difficult removal is anticipated. Future studies are needed to define optimal imaging strategies and identify patient populations that may benefit from targeted post‐placement ultrasound evaluation.
The authors acknowledge the sonographers and clinical staff of West Virginia University for their contributions to patient care and data acquisition. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number 5U54GM104942‐08. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Dr. Dueñas‐García is a consultant for Applied Medical. Relationship is not relevant to this publication.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Benacerraf BR, Shipp TD, Lyons JG, Bromley B. Ultrasound evaluation of intrauterine devices. Obstet Gynecol 2009; 114:585–590.19701039 [Google Scholar]
- 2. Braaten KP, Goldberg AB. Malpositioned intrauterine devices: when you should intervene (and when you should not). Obstet Gynecol 2012; 119:206–210.22183241 [Google Scholar]
- 3. Shipp TD, Bromley B, Benacerraf BR. The significance of nonfundal intrauterine device placement detected on sonography. J Ultrasound Med 2010; 29:1779–1782.21098850 [Google Scholar]
- 4. Moschos E, Twickler DM. Intrauterine devices: management of complications. Ultrasound Q 2011; 27:81–89.21606790 [Google Scholar]
- 5. Kaislasuo J, Suhonen S, Gissler M, Lähteenmäki P, Heikinheimo O. Intrauterine contraception: incidence and factors associated with uterine perforation—a population‐based study. Hum Reprod 2012; 27:2658–2663. [DOI] [PubMed] [Google Scholar]
- 6. Bozdag G, Yumru AE, Yildiz S, et al. Comparison of transvaginal ultrasonography and pelvic X‐ray for localization of intrauterine contraceptive devices. Contraception 2020; 101:202–206. [Google Scholar]
- 7. Turok DK, Jacobson JC, Dermish AI, et al. Performance of low‐placed intrauterine devices for contraception. Obstet Gynecol 2014; 123:1061–1069. [Google Scholar]
- 8. Moschos E, Twickler DM. Diagnostic accuracy of sonography in the detection of intrauterine contraceptive devices. J Ultrasound Med 2011; 30:1499–1505.22039022 [Google Scholar]
- 9. Kaislasuo J, Suhonen S, Gissler M, Lähteenmäki P, Heikinheimo O. Uterine perforation caused by intrauterine devices: clinical course and treatment. Hum Reprod 2013; 28:1546–1551. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.