LAUSR

Can cut-off values for oocyte maturation rate (OMR), fertilization rate (FR) and blastocyst formation rate (BFR) be established to define OZEMA (Oocyte/Zygote/Embryo Maturation arrest) phenotypes? This study is the first to develop quantile regression plots for OMR, FR and BFR in relation to age whereby age-specific cut-offs can be determined. Despite the high success rate of intracytoplasmic sperm injection (ICSI) cycles, outcomes may be compromised by oocyte maturation arrest (low OMR), fertilization failure (low FR), or embryo developmental arrest (low BFR). In 2023, the Online Mendelian Inheritance in Man database categorized these three phenotypes under the term Oocyte/Zygote/Embryo Maturation Arrest (OZEMA). However, data-based cut-off values for OMR, FR, and BFR are currently lacking. This retrospective cohort study analyzed data from 9,499 participants aged 18 to 45 years, who underwent at least one ICSI cycle between July 2015 and September 2024 in our tertiary center. Oocyte donors were excluded from OMR calculations. For FR/BFR calculations, cycles involving donor sperm, assisted oocyte activation, or vitrified-warmed oocytes were excluded. This resulted in 16,994 cycles (9,499 patients) for OMR analysis and 14,116 cycles (7,982 patients) for FR/BFR analysis. A quantile regression model with a constrained nonlinear term for age (monotone decreasing/convex) was fit for OMR, FR and BFR in relation to age. Subsequently, age-specific cut-off values were determined on the 0.05/0.10 quantile. Before applying this model, the minimum number of cumulus-oocyte-complexes (COCs) for OMR, ICSI inseminated oocytes for FR, or normally fertilized (2PN) oocytes for BFR was calculated to achieve a 90%/95% confidence interval (CI) halfwidth of at most 0.15, with 50% probability. When plotting this constrained quantile regression model for the 0.05 quantile, cut-off values for each age could be determined. For example, for ages 25, 30, 35 and 40 years, following values were determined: (1) For OMR: 44% at 25 years, 39% at 30 years, 34% at 35 years, 29% at 40 years; (2) For FR: 45% at 25 years, 34% at 30 years, 24% at 35 years, 13% at 40 years; (3) For BFR: 0% at 25 years, 0% at 30 years, 0% at 35 years, and 0% at 40 years. In addition, 0.10 quantiles were also calculated for BFR: 29% at 25 years, 19% at 30 years, 8% at 35 years, and 0% at 40 years. However, these cut-off values can only be applied if a minimum number of COCs for OMR, ICSI inseminated oocytes for FR, or 2PN fertilized oocytes for BFR are retrieved. Using the asymptotic normal approximation of the sampling distribution of a proportion, a minimum cumulative number of 31 (90% CI) or 43 (95% CI) COCs, ICSI inseminated oocytes, and 2PN fertilized oocytes should be retrieved to achieve a CI halfwidth of at most 0.15. The 0.05 or 0.10 quantile was based on clinical experience. In future research, machine learning may be applied to discover data clustering. Furthermore, as a high minimum cumulative number of 31 or 43 oocytes/zygotes is required, further finetuning of the analysis with additional cycle data could potentially reduce these numbers. Defining OZEMA phenotypes is crucial for advancing diagnosis (e.g. genetic screening panels), prognosis, and therapeutic strategies for these patients. Furthermore, the established cut-offs in this study provide a critical tool for standardizing comparisons across studies investigating these phenotypes. Yes