Predicting the unexpected total fertilization failure in conventional in vitro fertilization cycles: What is the role of semen quality?

Xiaoming Jiang; Jiali Cai; Lanlan Liu; Zhenfang Liu; Jinhua Chen; Chao Yang; Kaijie Chen; Xiaolian Yang; Jie Geng; Caihui Ma; Shuiyan Lian; Li Xu; Jianzhi Ren

doi:10.3389/fcell.2023.1133512

Predicting the unexpected total fertilization failure in conventional in vitro fertilization cycles: What is the role of semen quality?

Front Cell Dev Biol. 2023 Feb 23:11:1133512. doi: 10.3389/fcell.2023.1133512. eCollection 2023.

Authors

Xiaoming Jiang^{1

2}, Jiali Cai^{1

2}, Lanlan Liu^{1

2}, Zhenfang Liu¹, Jinhua Chen¹, Chao Yang¹, Kaijie Chen¹, Xiaolian Yang¹, Jie Geng¹, Caihui Ma¹, Shuiyan Lian¹, Li Xu¹, Jianzhi Ren¹

Affiliations

¹ Reproductive Medicine Center, Xiamen University Affiliated Chenggong Hospital, Xiamen, Fujian, China.
² School of Medicine, Xiamen University, Xiamen, Fujian, China.

Abstract

Background: Male and female gametes factors might contribute to the total fertilization failure (TFF). In first in vitro fertilization (IVF) cycles, decision-making of insemination protocol was mainly based on semen quality for the contribution of female clinical characteristics to TFF remained obscure. The purpose of the study was to evaluate the role of semen quality in predicting unexpected TFF. Methods: A single-center retrospective cohort analysis was performed on 19539 cycles between 2013 and 2021. Two algorithms, a Least Absolute Shrinkage and Selection Operator (LASSO) and an Extreme Gradient Boosting (Xgboost) were used to create models with cycle characteristics parameters. By including semen parameters or not, the contribution of semen parameters to the performance of the models was evaluated. The area under the curve (AUC), the calibration, and the net reclassification index (NRI) were used to evaluate the performance of the models. Results: The prevalence of TFF were .07 (95%CI:0.07-0.08), and .08 (95%CI:0.07-0.09) respectively in the development and validation group. Including all characteristics, with the models of LASSO and Xgboost, TFF was predicted with the AUCs of .74 (95%CI:0.72-0.77) and .75 (95%CI:0.72-0.77) in the validation group. The AUCs with models of LASSO and Xgboost without semen parameters were .72 (95%CI:0.69-0.74) and .73 (95%CI:0.7-0.75). The models of LASSO and Xgboost with semen parameters only gave the AUCs of .58 (95%CI:0.55-0.61) and .57 (95%CI:0.55-0.6). For the overall validation cohort, the event NRI values were -5.20 for the LASSO model and -.71 for the Xgboost while the non-event NRI values were 10.40 for LASSO model and 0.64 for Xgboost. In the subgroup of poor responders, the prevalence was .21 (95%CI:0.18-0.24). With refitted models of LASSO and Xgboost, the AUCs were .72 (95%CI:0.67-0.77) and .69 (95%CI:0.65-0.74) respectively. Conclusion: In unselected patients, semen parameters contribute to limited value in predicting TFF. However, oocyte yield is an important predictor for TFF and the prevalence of TFF in poor responders was high. Because reasonable predicting power for TFF could be achieved in poor responders, it may warrant further study to prevent TFF in these patients.

Keywords: oocyte yield; poor responder; prediction model; semen parameters; total fertilization failure.

Grants and funding

This work was supported by the National Natural Science Foundation of China (grant number 22176159); the Xiamen medical advantage subspecialty construction project (grant number 2018296) and the Special Fund for Clinical and Scientific Research of Chinese Medical Association (grant number 18010360765).