Abstract Background This study aimed to analyze the clinical data within one year of the patients who underwent hysteroscopic endometrial polypectomy with levonorgestrel-releasing intrauterine system (LNG-IUS) insertion, and to identify the risk factors affecting the duration of irregular uterine bleeding pattern (IUBP), trying to build models predicting this period and evaluate the performance of the models, in order to shorten the IUBP duration through intervention. Methods Clinical data were collected from 245 patients who underwent hysteroscopic endometrial polypectomy and LNG-IUS placement at our hospital between January 2018 and December 2022. The data, recorded within one year after LNG-IUS insertion, included age, the number of polyps, preoperative endometrial thickness, maximum polyp diameter, the timing of LNG-IUS placement, LNG-IUS expulsion and migration, coital bleeding, pelvic pain, amenorrhea, irregular uterine bleeding pattern, and polyp recurrence. The differences of adverse reactions were analyzed by age and LNG-IUS placement timing stratified. Logistic regression model and Generalized Additive Model (GAM) were established to predict the duration of irregular uterine bleeding after LNG-IUS insertion, and attempts were made to explore the relationship between the variables. Results Age was the independent risk factor for amenorrhea after LNG-IUS insertion, the patients with aged ≥ 40 years were more likely to experience amenorrhea (P < 0.05). No significant difference was observed between patients who had LNG-IUS placement at surgery immediately and those who had LNG-IUS placement within 3 months postoperatively regarding the incidence of adverse reactions. Preoperative endometrial thickness and maximum polyp diameter were the independent risk factors for the duration of irregular uterine bleeding after LNG-IUS insertion. Comprised preoperative endometrial thickness and maximum polyp diameter identified using logistic regression model and GAM model could predict the duration of IUBP effectively. The tenfold cross-validation showed that the GAM model (AUC 0.906) had slightly better predictive power than the bivariate logistic regression model (AUC 0.902). Conclusions Hysteroscopic polypectomy combined with LNG-IUS insertion was an effective measure to treat endometrial polyps. LNG-IUS intraoperative placement did not increase the incidence of adverse reactions. Preoperative endometrial thickness and maximum polyp diameter play a key role in valuable prediction for the duration of irregular uterine bleeding after LNG-IUS insertion.
Zhao et al. (Mon,) studied this question.