The Establishmentand Evaluation of Nomogram Models to Predict Pregnancy Outcomes of Ivf/ Icsi Fresh Cycle Treatment.
X. Yuan,Y. Yuan,Q. Wang,Y. Xu,J. Gao,Y. Li,C. Zhou
DOI: https://doi.org/10.1016/j.fertnstert.2017.07.1009
IF: 6.7
2017-01-01
Fertility and Sterility
Abstract:To identify independent risk factors of IVF/ICSI pregnancy outcomes and establish Nomogram models to predict IVF/ICSI prognosis. Retrospective study. 11453 fresh IVF/ICSI cycles from Jan 2009 to Dec 2013 were recruited in First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China. Univariate analyses were applied to 54 clinical/laboratory variables to explore relative factors influencing IVF/ICSI outcomes. Logistics regression was utilized to identify independent predictors of IVF/ICSI prognosis. Four Nomogram models to predict clinical pregnancy, live birth, miscarriage and ectopic pregnancy rate were subsequently established and values by C-index and Decision Curve Analysis. Of the 11453 fresh cycles, mean maternal age was 32.6 years. For infertility diagnosis, 32.1% of tubal factors, 21.8% of male factors, 5.8% of endometriosis, 34.1% of unexplained infertility, and 6.2% of multiple factors. Overall clinical pregnancy rate was 44.0% (5039/11453). Among the 5039 pregnant cycles, 4059 of them had thorough follow-ups, with live birth rate 81.6% (3314/4059), miscarriage rate 14.7%, and ectopic pregnancy rate 3.6%. Logistic regression model to predict IVF/ICSI clinical pregnancy rate with C-index 0.78 (95% CI, 0.77-0.79) contained 10 prognostic factors (p< 0.05), including maternal age, infertility diagnosis, endometrial thickness (EMT)/Estradiol/progesterone on hCG day, No. of oocytes retrieved, No. of two-pronucleus (2PN) zygote, No. of oocytes fertilized by ICSI, No. of usable embryos, and OHSS. Logistic regression model to predict live birth rate with C-index 0.65 (95% CI, 0.63-0.67), contained 7 prognostic factors (p< 0.05) including maternal age, infertility diagnosis, parity, No. of treatment cycles, EMT on hCG day, No. of 2PN zygote and OHSS. Similarly, the Logistic regression model to predic miscarriage rate with C-index 0.72 (95% CI, 0.70-0.75) contained 5 prognostic factors (p< 0.05) including maternal age, infertility diagnosis, parity, No. of treatment cycles, and OHSS. Meanwhile, the Logistic regression model to predict ectopic pregnancy rate with C-index 0.75 (95% CI, 0.70-0.79), contained 3 prognostic factors (p< 0.05) including infertility diagnosis, EMT and progesterone on hCG day. Four Nomogram models established subsequently with satisfying predictive calibration. Our study well presented clinical and epidemiological characteristics of infertility in Chinese population, and revealed independent prognostic factors of IVF/ICSI fresh cycles, and established four Nomogram models to predict clinical pregnancy, live birth, miscarriage and ectopic pregnancy rate of IVF/ICSI fresh cycle, making them adequate and scientific tools for both patients and physicians.