AUTHOR=Wang Jinming , Li Dan , Yeung Yang , Guo Zhenglong , Huang Hongke , Wang Li , Shi Weili , Huang Jianmei , Yang Wenke , Ren Yanxin , Liao Shixiu , Hao Yibin 

TITLE=Risk prediction for recurrent pregnancy loss based on routine inspections in the first trimester of pregnancy, a retrospective study in China

JOURNAL=Frontiers in Medicine

VOLUME=Volume 12 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2025.1476722

DOI=10.3389/fmed.2025.1476722

ISSN=2296-858X

ABSTRACT=BackgroundRecurrent pregnancy loss (RPL) is one of the most common pregnancy complications in obstetrics and gynecology, and the incidence rate of RPL is about 2%. To establish a risk prediction model for recurrent pregnancy loss based on routine inspections in the first trimester of pregnancy.Materials and methodsA total of 3,010 women at Henan Provincial People’s Hospital between January 2019 and December 2023. 810 women at Shangqiu Maternal and Child Health Hospital between January 2021 and April 2024. There were 523 women in the training set, 282 women in the testing set, and 229 women in the external validation set. Twelve routine inspections in the first trimester of pregnancy (4 ~ 12 weeks) were collected including thyroid-stimulating hormone (TSH), free triiodothyronine (FT3), free thyroxine thyroid (FT4), thyroxine (TT4), total triiodothyronine (TT3), peroxidase antibody (TPO-Ab), thyroid globulin antibody (TG-Ab), 25-hydroxyvitamin D (25-(OH) D), ferritin (Ferr), Homocysteine (Hcy), vitamin B12 (VitB12), folic acid (FA). Logistic regression analysis was used to establish a risk prediction model based on training set. Receiver operating characteristic (ROC) curve and decision curve analysis (DCA) were employed to evaluate the performance of prediction model on testing set and external validation set.ResultsChi-square test results for each single characteristic indicated that, TPO-Ab (p = 0.005), TG-Ab (p < 0.001), 25-(OH) D (p < 0.001), Hcy (p < 0.001) and FA (p < 0.001) were closely related to RPL. The prediction accuracy of the logistic regression model on the testing set was 71.28%, and area under ROC curve was 0.766. The prediction accuracy of the model on external validation set was 69.87%, and area under ROC curve was 0.759. Calibration curve and DCA curves of testing set and external validation set indicated that the model had good clinical value.ConclusionTPO-Ab, TG-Ab, 25-(OH) D, Hcy and FA may be closely related to the occurrence and development of RPL. The model only requires routine inspections in the first trimester of pregnancy to effectively indicate high-risk groups of RPL before the first miscarriage, making it convenient for clinical application and implementation.