Abstract
OBJECTIVE: The study aimed to develop a prediction model to assess the risk of subsequent pregnancy loss in patients with recurrent pregnancy loss (RPL), based on preconception hematologic and biochemical parameters. METHODS: This multicenter retrospective study included RPL patients from 2020 to 2022 and collected their preconception routine blood and liver/kidney function test results. Patients were assigned to either training or validation set. The outcome was pregnancy loss before 24 weeks of gestation. Candidate predictors were selected using backward method and the Akaike Information Criterion. A multivariable logistic regression model was constructed, and a nomogram was developed to estimate the risk of pregnancy loss. The model's discrimination was evaluated using the area under the receiver operator characteristic curve (AUC), and calibration was assessed using the Hosmer-Lemeshow test. Decision curve analysis (DCA) was performed to evaluate the clinical utility of the model. RESULTS: A total of 645 RPL patients were enrolled, with 351 in the training set and 294 in the validation set. Higher preconception levels of creatinine and indirect bilirubin were associated with a lower risk of pregnancy loss, while higher levels of uric acid, mean corpuscular volume, red blood cell distribution width-coefficient of variation, mean corpuscular hemoglobin concentration, red blood cell count, and lymphocyte count were associated with increased risk. The final model included 11 variables: eight parameters, in addition to maternal age, number of previous live births, and number of previous pregnancy losses. The model demonstrated good discrimination in the training set (AUC = 0.721) and good calibration (Hosmer-Lemeshow P = 0.183). External validation confirmed good performance, with an AUC of 0.740. The DCA indicated a positive net benefit of the model. CONCLUSIONS: The nomogram developed in this study, incorporating preconception hematologic and biochemical parameters, demonstrated good predictive performance for pregnancy loss in RPL patients and may serve as a useful tool for individualized risk assessment.