As gynecologic providers continue to provide intrauterine devices (IUDs) for long-acting reversable contraception (LARC) with insertion of those devices, the associated complication rate needs to be well-understood, so that patients can make properly informed shared healthcare decisions. There appears to be IUD embedment into the uterine wall that can occur over time, which can cause its fracture when retrieving it with grasp of its string. The authors previously published a case series [1] that described situations in which fractured IUDs were encountered after attempts at their removal occurred. Given that patients continue to be referred to the Director of Gynecologic Sonography at our institution for assistance in managing such complications when they are encountered, this case series was continued, and the authors believe it important to report these findings at this time.

The cases of these IUD fractures are listed in Table 1, revealing some relevant patient and IUD characteristics. Since these findings are retrospective in nature, and not prospectively collected, the proportion of IUD placement that this complication represents can only be estimated. For the population of patients described in this case series, we were able to approximate the number of IUD insertions, by our provider group practicing at multiple sites. The IUD insertions which were performed over the 7-year duration of this case series (2013 – 2019) are listed in Table 2, for the different types of devices, for this provider group, consisting of 18 different attending physicians. While some of the IUD insertions were performed by resident physicians, only 2 of the cases listed in Table 1 had an IUD inserted by a resident. None of the patients presented with complaints of pain at the time when IUD fracture was identified. There were also no visual or histologic evidence of inflammation or infection in the cases for which hysteroscopy was performed.

Table 1: List of fractured IUDs encountered.

Table 2: IUD insertions and fractures, by type, over seven years (2013 – 2019).

Whether an IUD is a non-hormone containing copper device (e.g., ParaGard^®) or a levonorgestrel intrauterine system (LNG-IUS, e.g., Mirena^®), the IUD term is applied. Each of these cases were identified as fractured IUDs, found at the time of removal from the uterus. The IUD fractures appear to have occurred after being embedded into the uterine wall, fracturing when the usual effort was exerted for the IUD removal (i.e., pulling on the string). It is presumed that embedment of the IUD into the uterine wall must have occurred, to result in the fractured IUD which was found. This phenomenon may be similar to the process of IUD perforation, alternatively referred to as “migration”,[2] in that it may occur over time, and is not necessarily temporally related to the IUD insertion. Figure 1 shows this with the use of 2D ultrasound, and Figure 2 illustrates this with the application of three-dimensional transvaginal sonography (3DTVS). Figure 3 shows a hysteroscopic view of an embedded IUD arm, at the time of its hysteroscopic removal.

As more young nulliparous women are using this LARC method [3], the number of larger frame devices being used, along with the duration of their potential use (5 years vs. 10 years), may be a relevant factor leading to the complication described in this case series. The size difference between a ParaGard^® or Mirena^® (having a diameter of 32 mm) and a Skyla^® or Kyleena^® (smaller frame size, having a diameter of 28 mm) may be clinically important. A difference in the intrauterine interostial diameter also exists between a nulliparous and a multiparous woman [4], which may also be clinically relevant.

While embedment of an IUD into the uterine wall may not commonly occur, this case series describes its approximate incidence of occurrence. Though this embedment cannot be anticipated, there may be factors relevant to its occurrence [5]. The patient and IUD features which are listed in Table 1, may represent identifiable conditions worthy of note. The duration of IUD use, as well as the IUD type, appears to be especially pertinent. The counting of the IUD insertions among all of the physicians who would typically refer for imaging of such complications, as well as the patients who underwent hysteroscopy for resolving this complication, yielded approximately 1000 IUD insertions per year over the course of 7 years. Naturally, the number of IUD insertions by this provider group, may not exactly correlate with the entire number of IUDs which were in place during the course of this case series, with some resulting in this complication. Nonetheless, in spite of the temporal discordance, these numbers may represent a close approximation of the risk of this phenomenon. It should be additionally noted that there may be limited continuation rates of IUD users [6], making an estimate of the true prevalence of IUD fracture to have limited accuracy in this dataset. Though it may be more ideal to present the described phenomenon in a prospectively followed investigation, that is unlikely to ever be practically performed. Thus, this case series is presented, for its valuable clinical information, to inform the patient (and physician) population. The low IUD complication rate that has been reported [7] may not be consistent with the experience in this case series, as we found the prevalence of this phenomenon (i.e., embedment and IUD fracture) to be about 1 per 500 IUD insertions for any IUD type, and specifically, slightly more than 1 per 100 ParaGard IUD insertions. The calculated relative risks of this particular IUD complication in our experience are described in Table 2, and this phenomenon may occur more commonly than what has been previously reported [8]. This complication may also be replicated by others who have the capability of this type of surveillance with sonography in skilled hands. The information contained in this present case series should be considered when obtaining informed consent to a patient considering this LARC method.

Chris Kabir, MS – for assistance with the statistical analyses performed.

None of the authors have any financial conflicts of interest.

Elliot M. Levine, MD: Conducted the extensive supporting literature review and data analysis.

Carlos M. Fernandez, MD: Performed the sonography for each described case in data set.

Marie Cabiya, MD: Instituted the original case report and identified the clinical problem described.

Imaan Ansari, MD: Resident physician who helped organize data in presented table.

Leah Delfinado, MD: Assistance in writing and editing manuscript.