Uterine transverse diameter and risk of LNG-IUS malposition in adenomyosis patients: a threshold effect analysis from a retrospective cohort study

Out of the initial 104 patients (Fig.  1 ), two were excluded following postoperative pathology confirmation of endometrial cancer. Three additional patients were excluded due to early LNG-IUS removal resulting from intolerable vaginal spotting. The final analysis included 99 patients, representing 95.2% of the original cohort. Fig. 1 The flowchart of study. In figure, we showed the procedures of subjects selection The flowchart of study. In figure, we showed the procedures of subjects selection Based on preoperative uterine transverse diameter, 99 patients were stratified into low-value ( n  = 29), middle-value ( n  = 33), and high-value ( n  = 37) tertiles. Comparative analysis of demographic characteristics across these groups revealed no statistically significant differences in most demographic parameters, obstetric history, and clinical manifestations (Table  1 ). A notable finding was the significant variation in intrauterine device (IUD) malposition rates among the three groups (3.45% vs. 6.06% vs. 27.03%, P  = 0.006), demonstrating a progressive increase in malposition risk with increasing uterine transverse diameter. Table 1 Baseline characteristics of participants stratified by preoperative uterine transverse diameter tertiles Characteristics Low tertile ( n  = 29) Middle tertile ( n  = 33) High tertile ( n  = 37) P -value P -value* Demographic and clinical parameters  Age (years) 41.79 ± 5.01 41.09 ± 6.65 42.00 ± 4.98 0.782 0.913  BMI (kg/m²) 22.24 ± 2.84 22.69 ± 2.62 23.54 ± 3.77 0.233 0.210  CA125 (U/mL) 23.39 ± 13.05 33.11 ± 33.20 31.38 ± 24.37 0.281 0.432  Preoperative hemoglobin (g/L) 117.34 ± 19.41 118.36 ± 19.36 116.65 ± 17.79 0.930 0.926 IUD position, n (%) 0.006† -  Normal 28 (96.55) 31 (93.94) 27 (72.97)  Abnormal 1 (3.45) 2 (6.06) 10 (27.03) Obstetric history  Gravidity, n (%) 0.753 -   0 0 (0.00) 2 (6.06) 1 (2.70)   1 2 (6.90) 3 (9.09) 5 (13.51)   2 9 (31.03) 7 (21.21) 11 (29.73)   3 10 (34.48) 11 (33.33) 10 (27.03)   ≥ 4 8 (27.59) 10 (30.30) 10 (27.03)  Parity, n (%) 0.446 -   0 0 (0.00) 2 (6.06) 1 (2.70)   1 11 (37.93) 13 (39.39) 9 (24.32)   2 15 (51.72) 14 (42.42) 25 (67.57)   ≥ 3 3 (10.35) 4 (12.12) 2 (5.40) Previous abortion, n (%) 21 (72.41) 22 (66.67) 18 (48.65) 0.110 - Previous cesarean section, n (%) 11 (37.93) 12 (36.36) 15 (40.54) 0.936 - Clinical manifestations  Retroflexed uterus, n (%) 13 (44.83) 14 (42.42) 18 (48.65) 0.870 -  Menorrhagia, n (%) 23 (79.31) 26 (78.79) 30 (81.08) 0.969 - Dysmenorrhea VAS score, n (%) 0.781 -  0 13 (44.83) 14 (42.42) 12 (32.43)  1–3 11 (37.93) 11 (33.33) 13 (35.13)  4–6 5 (17.24) 8 (24.24) 8 (21.62)  7–10 0 (0.00) 0 (0.00) 4 (10.81) Data are presented as mean ± SD or n (%) BMI body mass index, IUD intrauterine device, VAS visual analog scale, CA125 cancer antigen 125 P -value was calculated using one-way ANOVA for continuous variables and Chi-square test or Fisher’s exact test for categorical variables P -value* represents adjusted P -value after controlling for confounding factors †Statistically significant ( P  < 0.05) Baseline characteristics of participants stratified by preoperative uterine transverse diameter tertiles Data are presented as mean ± SD or n (%) BMI body mass index, IUD intrauterine device, VAS visual analog scale, CA125 cancer antigen 125 P -value was calculated using one-way ANOVA for continuous variables and Chi-square test or Fisher’s exact test for categorical variables P -value* represents adjusted P -value after controlling for confounding factors †Statistically significant ( P  < 0.05) In the unadjusted Cox regression analysis, each 1-cm increment in uterine transverse diameter was associated with a 17% increased risk of IUD malposition (HR = 1.17, 95% CI: 0.88–1.57). After adjusting for age, this association did not show significant changes and remained stable (HR = 1.17, 95% CI: 0.88–1.56, P  = 0.290). After a thorough adjustment for potential confounding factors, the association became significantly stronger (HR = 1.49, 95% CI: 0.98–2.25, P  = 0.060) (Table  2 ). Table 2 Multivariate Cox regression analysis for the association between preoperative uterine transverse diameter and IUD malposition Variables Crude model Model 1 Model 2 Uterine transverse diameter 1.17 (0.88–1.57) 1.17 (0.88–1.56) 1.49 (0.98–2.25) P value 0.286 0.290 0.060 Data are presented as hazard ratio (95% confidence interval) Model 1: Adjusted for age Model 2: Adjusted for age, body mass index, gravidity, parity, history of abortion, history of cesarean section, uterine position, menorrhagia, dysmenorrhea VAS score, serum CA125, and preoperative hemoglobin level Multivariate Cox regression analysis for the association between preoperative uterine transverse diameter and IUD malposition Data are presented as hazard ratio (95% confidence interval) Model 1: Adjusted for age Model 2: Adjusted for age, body mass index, gravidity, parity, history of abortion, history of cesarean section, uterine position, menorrhagia, dysmenorrhea VAS score, serum CA125, and preoperative hemoglobin level Piecewise Cox regression analysis (Table  3 ) revealed a statistically significant non-linear relationship between uterine transverse diameter and IUD malposition (non-linearity test P  = 0.018). A distinct threshold effect was observed at 5.4 cm: In cases where the uterine transverse diameter is below the threshold of 5.4 cm, each 1-cm increase in this measurement corresponds to a 74% reduction in the risk of malposition. (HR = 0.26, 95% CI: 0.07–0.99, P  = 0.049). When measurements exceed the 5.4 cm threshold, risk of malposition rises by 112% with each additional centimeter (HR = 2.12, 95% CI: 1.28–3.49, P  = 0.003). As depicted in Fig.  2 , the smoothing curve reveals a non-linear relationship, showing that IUD placement accuracy diminishes significantly once the uterine transverse diameter surpasses 5.4 cm. Table 3 Association between preoperative uterine transverse diameter and IUD malposition using linear and piecewise linear regression models Models and segments HR (95% CI) P value Model I (Linear) 1.48 (0.98–2.25) 0.063 Model II (Piecewise) Threshold value 5.4 cm Below threshold (< 5.4 cm) 0.26 (0.07–0.99) 0.049* Above threshold (≥ 5.4 cm) 2.12 (1.28–3.49) 0.003** Log-likelihood ratio test† 0.018* Models were adjusted for age, body mass index, gravidity, parity, history of cesarean section, dysmenorrhea severity, preoperative hemoglobin level, and other potential confounders HR hazard ratio, CI confidence interval †Log-likelihood ratio test comparing Model II versus Model I * P  < 0.05, ** P  < 0.01 Association between preoperative uterine transverse diameter and IUD malposition using linear and piecewise linear regression models Models were adjusted for age, body mass index, gravidity, parity, history of cesarean section, dysmenorrhea severity, preoperative hemoglobin level, and other potential confounders HR hazard ratio, CI confidence interval †Log-likelihood ratio test comparing Model II versus Model I * P  < 0.05, ** P  < 0.01 Fig. 2 Association between preoperative uterine transverse diameter and probability of normal IUD positioning Association between preoperative uterine transverse diameter and probability of normal IUD positioning The smoothing curve illustrates the non-linear relationship between preoperative uterine transverse diameter and the probability of normal IUD positioning. Solid dots represent the fitted curve, and open circles indicate the 95% confidence intervals. The x-axis shows the preoperative uterine transverse diameter in units, and the y-axis represents the probability of normal IUD position (1.0 = normal position, 0.0 = malposition). A clear threshold effect is observed at 5.4 units of uterine diameter, after which the probability of normal IUD positioning decreases substantially. The analysis was adjusted for age, body mass index, gravidity, parity, history of abortion, history of cesarean section, uterine position, menorrhagia, dysmenorrhea VAS score, serum CA125 level, and preoperative hemoglobin level. Subgroup analysis (Table  4 ) revealed significant effect modification by dysmenorrhea severity in the association between uterine transverse diameter and IUD malposition (interaction P  = 0.022). In patients with severe dysmenorrhea (VAS score ≥ 7), a stronger association was observed (HR = 3.43, 95% CI: 1.32–8.89, P  = 0.011). In patients with mild to moderate dysmenorrhea, as indicated by a Visual Analog Scale (VAS) score of less than 7, no significant association was found (HR = 1.01, 95% CI: 0.61–1.65, P  = 0.976). Additionally, the stratified analyses revealed no significant effect modification based on hemoglobin levels or history of cesarean section, with interaction P-values of 0.504 and 0.630, respectively. Table 4 Subgroup analyses of the association between preoperative uterine transverse diameter and IUD malposition Subgroup N HR (95% CI) P value P for interaction History of cesarean section 0.630  Yes 38 1.07 (0.49–2.34) 0.856  No 61 1.48 (0.93–2.37) 0.101 Dysmenorrhea VAS score 0.022*  110 64 1.84 (1.05–3.21) 0.033* Adjusted for age, body mass index, gravidity, parity, history of abortion, uterine position (anteversion/retroversion), serum CA125 level, and menorrhagia * P  < 0.05 Subgroup analyses of the association between preoperative uterine transverse diameter and IUD malposition Adjusted for age, body mass index, gravidity, parity, history of abortion, uterine position (anteversion/retroversion), serum CA125 level, and menorrhagia * P  < 0.05

This study followed the STROBE guidelines for designing and reporting a retrospective cohort study. The study included patients with adenomyosis. These patients visited Fujian Maternity and Child Health Hospital from May 2021 to May 2023. The diagnosis of adenomyosis was made through comprehensive clinical evaluation combining patient symptoms and transvaginal ultrasound following the updated MUSA criteria (Morphological Uterus Sonographic Assessment). Diagnosis required the presence of at least two of the following features: junctional zone irregularity, asymmetric myometrial thickening, myometrial cysts, hyperechoic islands, echogenic subendometrial lines and buds, or fan-shaped shadowing [ 9 ]. All ultrasound examinations were performed by experienced sonographers with specialized training in gynecological ultrasound. All patients were treated by senior gynecologists from our team, following a consistent clinical protocol of direct LNG-IUS insertion without hormonal pretreatment, in accordance with current guidelines recommending LNG-IUS as first-line therapy for adenomyosis. No patients had received GnRH agonists, dienogest, or other hormonal therapies within 3 months prior to insertion. Inclusion criteria: Premenopausal adult females. Symptomatic adenomyosis diagnosed by transvaginal ultrasonography. Underwent hysteroscopic levonorgestrel-releasing intrauterine system (LNG-IUS) insertion. Possessed complete clinical records and follow-up documentation. Premenopausal adult females. Symptomatic adenomyosis diagnosed by transvaginal ultrasonography. Underwent hysteroscopic levonorgestrel-releasing intrauterine system (LNG-IUS) insertion. Possessed complete clinical records and follow-up documentation. Exclusion criteria: Concurrent uterine myoma (diameter > 3 cm). Previous uterine surgery history (excluding cesarean section). Concomitant uterine malformation. Severe systemic diseases. Mental or cognitive disorders. Premature LNG-IUS removal due to postoperative complications. Concurrent uterine myoma (diameter > 3 cm). Previous uterine surgery history (excluding cesarean section). Concomitant uterine malformation. Severe systemic diseases. Mental or cognitive disorders. Premature LNG-IUS removal due to postoperative complications. The insertion procedure was timed 3 to 7 days after menstruation ended, or following cessation of bleeding in cases of prolonged menstrual flow. We excluded pregnancy and active pelvic inflammatory disease through preoperative screening. Our team’s senior physicians, each with 15 + years of hysteroscopic surgical experience, conducted all procedures. It should be emphasized that hysteroscopic guidance is not routinely required for LNG-IUS insertion in clinical practice; this approach was specifically chosen for research standardization to ensure accurate device placement and systematic cavity evaluation in our study population. Patients underwent intravenous compound anesthesia and were placed in the bladder lithotomy position. We employed a 0° rigid hysteroscope system (Karl Storz GmbH, Tuttlingen, Germany) with normal saline as the distension medium at a pressure of 80–100 mmHg. After sequential cervical dilation up to Hegar 7, the uterine cavity was systematically examined for research purposes. In clinical practice, such systematic hysteroscopic evaluation should be considered primarily when there is clinical suspicion of endometrial pathology (such as abnormal uterine bleeding, especially in our study population with adenomyosis where the majority of patients presented with menorrhagia), rather than as a routine procedure for all LNG-IUS candidates .Any pathological lesions were addressed, and endometrial curettage was performed as part of the research protocol. The LNG-IUS was then methodically inserted to the uterine fundus using a dedicated inserter. Direct hysteroscopic visualization confirmed that both arms of the device were fully deployed and the system was properly positioned at the fundus. The threads were subsequently trimmed to extend approximately 2 cm beyond the external cervical os. The uterine transverse diameter was measured using a Voluson E10 transvaginal ultrasound (GE Healthcare). A sonographer with at 15 years of experience conducted standardized measurements of the uterine transverse diameter during the early follicular phase (days 7–10) of the menstrual cycle [ 10 ]. Measurements were taken within four weeks before the LNG-IUS insertion. The primary outcome variable was LNG-IUS malposition, defined by two criteria [ 11 ]: Downward Displacement: This occurs when the upper edge of the intrauterine device is located more than 20 mm from the uterine fundus. Device Expulsion: Characterized by partial device location within the cervical canal or complete vaginal expulsion. Downward Displacement: This occurs when the upper edge of the intrauterine device is located more than 20 mm from the uterine fundus. Device Expulsion: Characterized by partial device location within the cervical canal or complete vaginal expulsion. All patients followed a standardized follow-up schedule: ultrasound examinations at 1, 3, 6, and 12 months post-LNG-IUS insertion, with subsequent follow-ups every 6 months. Follow-up was discontinued if the device was displaced or expelled. The last follow-up appointment was scheduled for October 2024. Follow-up duration ranged from 3 to 30 months (median 25 months). Covariate selection was based on previously reported potential influencing factors, including age, body mass index, gravidity, parity, abortion history, cesarean section history, uterine position, menorrhagia, dysmenorrhea VAS score, serum CA125 level, and preoperative hemoglobin level. Clinical data were collected using the hospital’s electronic medical record system, and data accuracy was verified by two individuals. No missing data were present, and complete case analysis was employed. Statistical analysis was performed using R software version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were expressed as mean ± standard deviation, with between-group comparisons using one-way ANOVA. Categorical variables were reported as counts and percentages, with chi-square or Fisher’s exact tests used for between-group comparisons as appropriate. Cox proportional hazards regression models were used to assess the association between preoperative uterine transverse diameter and IUD malposition. Three models were constructed: an unadjusted model, Model 1 (age-adjusted), and Model 2 (fully adjusted). In our statistical analysis, we applied both standard linear regression and piecewise linear regression models to explore potential non-linear relationships in our dataset. To determine the optimal threshold point for the piecewise regression analysis, we implemented the likelihood profile method. Subsequently, log-likelihood ratio tests were performed to assess whether the piecewise linear model offered significant improvement over the conventional linear approach. Stratified analyses were conducted based on predefined clinical characteristics to assess effect modification: cesarean section history, dysmenorrhea severity (VAS score 110 g/L). Likelihood ratio tests were used to evaluate interaction effects. Results were expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). All statistical tests were two-sided, with P  < 0.05 considered statistically significant. For threshold effect studies in survival analysis, we adhered to the statistical principle requiring at least 10 events per covariate to ensure stability and reliability of Cox regression models. This study employed a consecutive cohort design, including all patients with adenomyosis who met inclusion criteria at our center from May 2021 to May 2023. This approach offers several advantages: (1) avoidance of selection bias, ensuring representativeness of findings; (2) maximization of rare event observation capacity; (3) provision of real-world clinical data. Based on literature reports, the incidence of LNG-IUS malposition in adenomyosis patients ranges from 10 to 20%. The consecutive cohort design ensures observation of sufficient event numbers for survival analysis and threshold effect identification. We planned to employ piecewise linear regression to identify non-linear relationships, which is particularly effective for exploring threshold effects. The study protocol received approval from the Ethics Committee of Fujian Maternal and Child Health Hospital, with the approval number 2025KY022. As this research is retrospective in nature, all patient data were anonymized, and no personal identifying information was collected. In line with the Declaration of Helsinki, informed consent was waived. The study was conducted in strict accordance with medical ethical guidelines, ensuring the confidentiality and security of patient data throughout the research process.

This retrospective cohort study of 99 adenomyosis patients revealed a significant non-linear association between uterine transverse diameter and LNG-IUS malposition. A critical threshold of 5.4 cm was identified: beyond this point, each 1-cm increase in uterine transverse diameter was associated with a 112% increased risk of LNG-IUS malposition. Notably, this association was more pronounced among patients experiencing severe dysmenorrhea.

Our investigation examined how uterine transverse diameter relates to levonorgestrel-releasing intrauterine system (LNG-IUS) malposition in adenomyosis patients. We tracked 99 subjects for 3–30 months, focusing on transverse diameter as a key morphological marker. Findings revealed a notable non-linear correlation between transverse diameter and LNG-IUS malposition, with a critical threshold at 5.4 cm. Beyond this point, each additional centimeter corresponded to a 112% heightened malposition risk (HR = 2.12, 95% CI: 1.28–3.49). In recent years, studies on LNG-IUS for adenomyosis have expanded our understanding of factors affecting treatment efficacy [ 12 – 14 ] We compared our findings with published literature to emphasize potential clinical applications. Chen and colleagues previously identified enlarged uterine volume as a contributing factor [ 5 ]. Building on this observation, we applied non-linear analytical techniques to investigate whether uterine transverse diameter might offer predictive value. Our retrospective design certainly limits definitive conclusions, but may help clinicians identify suitable LNG-IUS candidates. Previous investigators examined general relationships between uterine dimensions and LNG-IUS malposition in adenomyosis patients [ 15 ]. Our approach differs in its focus on targeted measurement parameters, allowing for more granular analysis of these correlations. While acknowledging the inherent limitations of retrospective data, we attempted to establish usable clinical reference points. Notably, we observed patterns linking dysmenorrhea severity with LNG-IUS expulsion risk - findings generally aligning with Chen’s observations. After adjusting for various covariates, severe dysmenorrhea maintained its association with increased risk of LNG-IUS malposition. The threshold we identified for uterine transverse diameter could serve as one element in clinical risk assessment, particularly for patients with severe dysmenorrhea. These women may benefit from thorough preoperative ultrasound evaluation. When measurements exceed this threshold, clinicians might consider closer follow-up or complementary approaches such as GnRH agonists, High-Intensity Focused Ultrasound, or LNG-IUS suture fixation techniques [ 4 , 16 ]. We must emphasize that prospective validation is needed before firmly establishing such protocols. By focusing on transverse diameter as a morphological marker, we hope to contribute to existing risk stratification frameworks. These preliminary findings may inform the development of personalized treatment protocols. Importantly, LNG-IUS treatment may be particularly advantageous in young women with early-stage adenomyosis when the disease has not yet progressed and uterine dimensions remain within normal limits. Early intervention in this population may help prevent disease progression, preserve reproductive function, and optimize long-term treatment outcomes. We attempted to minimize the methodological limitations inherent to retrospective studies through strict inclusion/exclusion criteria and standardized measurements. Our statistical approach employed Cox regression with piecewise linear regression, identifying a potential 5.4 cm threshold for uterine transverse diameter. The relationship between dysmenorrhea severity (quantified using VAS scores) and device malposition provided intriguing associations. While we adjusted for known confounders including age, BMI, obstetric history, and laboratory parameters, residual confounding remains an unavoidable limitation in retrospective analyses. This research is scientifically novel due to its innovative use of uterine transverse diameter as an independent morphological indicator. Additionally, it employs non-linear statistical analysis, providing a more precise risk assessment framework for clinical practice. These insights could optimize personalized treatment strategies in adenomyosis patients and establish a theoretical foundation for future investigations. Despite the study’s important clinical implications, several limitations should be noted. As a single-center retrospective observational study, the external validity of the results may be limited. The small sample size ( n  = 99) may restrict statistical power, especially in stratified analyses and when exploring non-linear relationships, as some subgroups may have insufficient participants to identify significant differences. Our exclusion criteria, particularly for patients with larger uterine fibroids (> 3 cm) and those with previous uterine surgery history (excluding cesarean sections), may limit the applicability of results to these populations. Regarding operational aspects, all patients underwent LNG-IUS insertion under hysteroscopic direct visualization (note: hysteroscopy is not routinely required for IUS insertion, but was chosen in this study to ensure optimal placement and address coexisting endometrial lesions). In our study population with adenomyosis (mean age 41.79 ± 5.01 years), there were additional considerations that supported systematic hysteroscopic evaluation: first, adenomyosis patients may have an increased risk of concurrent endometrial pathology; second, the majority of our patients (approximately 80%) presented with menorrhagia, which warrants endometrial evaluation according to current guidelines [ 17 ]. While hysteroscopy offers advantages in direct observation, combined diagnosis and treatment [ 18 ], and ensuring accurate LNG-IUS placement, this approach differs from commonly used blind insertion or ultrasound-guided insertion methods. Consequently, caution is warranted when applying these findings to populations utilizing different insertion techniques. An important limitation of our study is the exclusive use of hysteroscopic insertion with cervical dilation up to Hegar 7. While this approach enabled direct visualization of proper device placement and concurrent treatment of intrauterine pathology, cervical dilation may potentially increase the risk of subsequent IUS expulsion compared to blind insertion techniques. This procedural approach may limit the generalizability of our findings to clinical settings where blind insertion or ultrasound-guided placement is routinely performed. Future studies should investigate whether our identified threshold effect remains consistent across different insertion techniques. Furthermore, since our study subjects were exclusively from the Fujian region of China, geographical and racial differences should be considered when extrapolating these findings. An important consideration in interpreting our findings is that our study population had a mean age exceeding 40 years, which may limit the applicability of our 5.4 cm threshold to younger patients. Recent evidence suggests that adenomyosis has an earlier onset than previously recognized, with younger patients often presenting with smaller uterine dimensions and potentially different disease characteristics. Young women may have smaller transverse diameters and different uterine compliance, potentially affecting the relationship between uterine dimensions and device stability. Future research should specifically investigate whether our identified threshold remains applicable in younger populations and whether age-specific thresholds might be more clinically relevant. Despite adjusting for multiple known confounding factors, unmeasured confounders may still exist, and potential selection bias cannot be completely ruled out. Finally, the observational nature of our study only allows us to examine associations between uterine transverse diameter and LNG-IUS malposition without establishing definitive causation. The systematic follow-up protocol established in our study (ultrasound examinations at 1, 3, 6, and 12 months post-insertion, followed by every 6 months thereafter) provides a regular monitoring protocol for adenomyosis patients. Regular ultrasound surveillance enables early detection of device malposition, assessment of symptom improvement, and identification of disease progression, particularly crucial for patients with larger uterine dimensions who are at higher risk for device complications. However, it must be acknowledged that adenomyosis, as a chronic progressive disease, typically requires monitoring until menopause (usually necessitating several years of long-term follow-up), while our follow-up period of only 3–30 months (median 25 months) is insufficient for comprehensive evaluation of long-term efficacy and safety of LNG-IUS. Our study provides only preliminary insights into early temporal patterns of device malposition. Longer-term follow-up studies are critically needed to determine the durability of threshold effects, assess long-term device stability, and monitor disease progression. Future research should prioritize multi-center, large-sample prospective studies to enhance external validity. In-depth exploration of potential physiological mechanisms underlying uterine morphological changes and intrauterine device positioning stability is needed. Additionally, integrating interdisciplinary approaches from imaging, endocrinology, and biomechanics could lead to more comprehensive clinical risk prediction models, ultimately improving therapeutic outcomes for adenomyosis patients.

Adenomyosis is a prevalent gynecological disorder characterized by the invasion of endometrial glands and stroma into the uterine myometrium [ 1 ]. Recent epidemiological studies reveal that the global prevalence of adenomyosis among women of reproductive age is approximately 15%−20%, with an increasing trend [ 2 ]. Adenomyosis not only significantly impacts patients’ quality of life but also imposes substantial physical and psychological burdens. Adenomyosis patients benefit from several therapeutic approaches, with the levonorgestrel-releasing intrauterine system (LNG-IUS) now established as first-line therapy due to its unique combination of simple insertion technique, complete reversibility, and well-documented clinical efficacy [ 3 ]. Despite these advantages, clinicians frequently encounter a significant challenge in LNG-IUS management—device malposition, which manifests either as gradual downward displacement within the uterine cavity or complete expulsion. This complication occurs at concerning rates, with published clinical data documenting expulsion frequencies ranging from 20 to 30% in adenomyosis cases [ 4 ]. Such positioning failures not only compromise therapeutic efficacy but also expose patients to unintended pregnancy risks, underscoring the clinical importance of identifying reliable predictors of LNG-IUS stability in this patient population. Uterine morphology is a crucial factor affecting the stability of intrauterine devices (IUDs). Uterine transverse diameter, as an essential indicator of uterine morphology, reflects the degree of lateral uterine expansion. Existing research has demonstrated a significant statistical association between uterine volume and LNG-IUS expulsion [ 5 ]. Notably, the maximum transverse diameter of the uterine cavity has been confirmed to be closely correlated with IUD stability [ 6 ]. The 2019 clinical practice guidelines from the French National College of Obstetricians and Gynecologists highlight that a large transverse diameter of the uterine cavity is a risk factor for IUD expulsion [ 7 ]. These findings suggest that uterine transverse diameter may serve as a potential predictor of LNG-IUS malposition. Existing research has begun to explore how uterine morphology affects the stability of intrauterine device positioning [ 8 ]. However, systematic studies on the link between uterine transverse diameter and LNG-IUS malposition are still limited. Previous studies primarily focused on the association between uterine volume and intrauterine device expulsion, lacking precise analysis of specific morphological indicators. Furthermore, most existing research relies on cross-sectional or case-control studies, and there is a lack of prospective or well-designed cohort studies to confirm potential causal relationships. Although some studies indicate that changes in uterine morphology might affect the stability of intrauterine device positioning, the exact mechanisms and clinical implications still need further investigation. Based on limitations in existing research, we proposed that increased uterine transverse diameter might be associated with a higher risk of LNG-IUS malposition in adenomyosis patients. To explore this relationship, we retrospectively analyzed data from a cohort of adenomyosis patients, focusing on the connection between uterine transverse diameter measurements and LNG-IUS malposition events. This study aims to provide clinicians with practical evidence for risk assessment before LNG-IUS insertion, helping to develop treatment plans better suited to individual patients.