Occurrence of smooth endoplasmic reticulum aggregates in metaphase II oocytes: relationship with stimulation protocols and outcome of ICSI and IVF cycles

STUDY QUESTION: Is there any association between the appearance of smooth endoplasmic reticulum aggregates (SERa) in oocytes and ovarian stimulation, embryological, clinical and neonatal outcomes of ICSI and IVF cycles? SUMMARY ANSWER: A suboptimal prolonged ovarian stimulation is detrimental to oocytes by inducing the occurrence of SERa, which reduces the reproductive potential of oocytes. WHAT IS KNOWN ALREADY: Controlled ovarian stimulation recruits oocytes of different qualities. Based on current evidence, it was agreed that non-homogeneous cytoplasm may represent the normal variability among oocytes rather than a dysmorphism with developmental significance. The only exception is the appearance of SERa within the ooplasm. Owing to the lack of univocal evidence in this literature about the safety of injecting oocytes with SERa and the mechanism responsible for the occurrence of SERa, this topic is still a matter of debate. STUDY DESIGN, SIZE, DURATION: A retrospective, longitudinal cohort study performed at a tertiary level public infertility center. We included 1662 cycles (180 SERa+ and 1482 SERa-) from 1129 women (age: 20--44 years) who underwent IVF/ICSI treatments in 2012-2019. The SERa+ cycles had at least one SERa+ oocyte in the oocyte cohort. The SERa- cycles had morphologically unaffected oocytes. PARTICIPANTS/MATERIALS, SETTING, METHODS: We collected stimulation data and embryological, clinical, neonatal outcomes of SERa- and SERa+ cycles and oocytes MAIN RESULTS AND THE ROLE OF CHANCE: Overall, 347 out of 12 436 metaphase II oocytes (2.8%) were affected by SER. We performed only 12 transfers involving at least one SERa+ embryo. Stimulation length (P = 0.002), serum progesterone (P = 0.004) and follicle size (P = 0.046) at trigger, number of retrieved (P = 0.004) and metaphase II (P = 0.0001) oocytes were significantly higher in SERa+ than SERa- cycles. Fertilization rate was significantly (P < 0.0001) reduced in SERa+ cycles and oocytes compared to SERa- counterparts. Embryos of SERa+ cycles had a lower blastocyst formation rate compared to embryos of SERa- cycles (P = 0.059). Statistical analysis according to a generalized estimating equation model performed at patient level demonstrated that the duration of ovarian stimulation was predictive of SERa+ oocytes appearance. The clinical success of SERa+ cycles was lower than SERa- cycles, although no differences in neonatal birthweights or malformations were recorded in sibling unaffected oocytes of SERa+ cycles. LIMITATIONS, REASONS FOR CAUTION: Given that SERa+ oocytes were discarded in our center for years and transfers of embryos originating from affected oocytes were generally avoided, clinical outcomes of SERa+ cycles are largely attributable to the transfer of embryos derived from unaffected oocytes of SERa+ cycles and we did not have data about newborns from affected oocytes, since none of the transfers involving SERa+ embryos resulted in a progressive clinical pregnancy. WIDER IMPLICATIONS OF THE FINDINGS: For the first time, we speculate that the late-follicular phase elevated serum progesterone caused by a suboptimal prolonged ovarian stimulation may be detrimental to the oocytes by inducing the occurrence of SERa, resulting in negative effects on their reproductive potential. This raises the question of whether some stimulation regimens could be worse than others and a change in stimulation protocol would reduce the possibility of producing oocytes with suboptimal maturation. In particular, our data highlight the importance of correct timing of the trigger in order to maximize oocyte collection, not only in terms of numerosity but also their reproductive potential.