Development and validation of a conventional in vitro total fertilization failure prediction model

Abstract

Background: Conventional total fertilization failure (TFF) is a challenging problem for clinicians. The predictive model developed in this study aims to predict the individual probability of conventional in vitro total fertilization failure.

Methods: The prediction model was developed based on 1635 patients who underwent first-attempt in vitro fertilization (IVF) cycles from January 2018 to January 2020. Total fertilization failure and normal fertilization occurred in 218 and 1417 cycles, respectively. Multivariate logistic regression analyses were used to develop the prediction model. Performance of our model was evaluated using calibration (Hosmer-Lemeshow test) and discrimination (area under the receiver operating characteristic curve [AUC]).

Results: Thirteen risk factors for TFF were included in the prediction model, as follows: female age; female body mass index; infertility duration; number of oocytes retrieved; stimulation protocol; infertility etiology; infertility diagnosis; male age; sperm concentration; total sperm motility; normal sperm morphology percentage; swim-up sperm motility; and swim-up sperm concentration. The AUC of our model was 0.815 (95% CI: 0.783-0.846), indicating satisfactory discrimination performance.

Conclusion: Considering female and male factors (especially sperm parameters), we established a model that predicts the probability of TFF in conventional IVF procedures that will be helpful in the laboratory supporting IVF to facilitate physicians in determining optimal treatment.

Keywords: Conventional IVF; Predictive model; Sperm parameters; Total fertilization failure.