LOW FERTILIZATION: statistical definition and clinical relevance

A reduced fertilization rate is a frustrating condition for patients and embryologists. This study aimed develop a statistical metric, LOW FERTILIZATION, to determine whether a distinct pathological entity underlies reduced fertilization or if it merely reflects statistical variability.

An indirect argument inspired by proof by contradiction was developed using a statistical model based on two assumptions: (1) fertilization occurs with same probability (P) for all oocytes with P based on Istanbul Consensus benchmarks, (2) no underlying clinical condition affects the couple. With these assumptions, oocytes fertilization in a single cycle follows a binomial distribution, with LOW FERTILIZATION defined as outcome occurring with an expected probability below 5% under the null hypothesis. Afterwards, we applied findings to a real-world population evaluating the frequency of LOW FERTILIZATION, predictive factors, and recurrence risk.

We provided a table indicating the maximum number of fertilized oocytes to define low fertilization, alongside R code, to facilitate reproducibility. For real-world data in 2653 women, LOW FERTILIZATION was observed in 118 cases (4.5%, 95%CI: 3.7-5.3%), occurring in 3.6% of cIVF and 5.2% of ICSI cycles. Severe male factor infertility was associated with LOW FERTILIZATION in ICSI (adjusted OR = 2.60, 95% CI: 1.37-4.73). The adjusted OR of recurrence of LOW FERTILIZATION was 3.70 (95%CI: 1.40-8.71), rising to 7.26 (95%CI: 2.48-19.93) for couples undergoing ICSI in both cycles.

The LOW FERTILIZATION metric demonstrates a clinical condition, although relatively rare, associated with reduced fertilization. It also offers a valuable tool to guide future research investigating factors associated with this condition.