Nomogram model to predict the probability of spontaneous abortion in human assisted reproductive technology

Abstract

Abstract Background: Spontaneous abortion is one of the most common complications of pregnancy in assisted reproductive technology (ART), it has become a thorny issue that poses a threat to pregnant women and their unborn children. Prevention is more important than treatment as for spontaneous abortion, Unfortunately, there is no accurate model to predict and prevent its occurrence. Methods: This retrospective cohort study evaluated clinical data from 15,723 patients successfully conceived who underwent IVF/ICSI in the first people’s hospital of shangqiu and the affiliated hospital of zhengzhou university from december 2015 to February 2021. Patients were divided into an "abortion group" and a "non-abortion group" according to the outcome of abortion. We further performed a multi-factor and single-factor logistic regression analysis to determine the independent factors, then we establish nomogram model to avoid the occurrence of spontaneous abortion in ART patients based on the independent influencing factors. Results: In this study, 15,723 patients had successfully conceived with ART, of which 2,336 patients (14.86%) had abortion, and 13,387 patients (85.14%) did not. Logistic regression analysis showed that female age (OR 1.050; 95% CI 1.029 ~ 1.061; P<0.001), male age (OR 1.102; 95% CI 1.086 ~ 1.115; P<0.001), follicle-stimulating hormone (OR 1.044; 95% CI 1.015 ~ 1.069; P<0.001), anti-Mullerian hormone (OR 0.891; 95% CI 0.862 ~ 0.925; P<0.001) were risk factors for spontaneous abortion. Then a nomogram model of abortion occurrence was constructed according to the above factors, and the AUC of the modeling group is 0.746 (95%CI 0.707-0.784), the AUC of the verification group is 0.718 (95%CI 0.999-0.737). Conclusions: Our study found that female age, male age, follicle-stimulating hormone, anti-mullerian hormone were independent risk factors for spontaneous abortion, and the Nomogram model is an effective, simple and intuitive clinical model for predicting the probability of spontaneous abortion in ART.