Ovarian stimulation for freeze-all IVF cycles: a systematic review

BACKGROUND: Freeze-all IVF cycles are becoming increasingly prevalent for a variety of clinical indications. However, the actual treatment objectives and preferred treatment regimens for freeze-all cycles have not been clearly established. OBJECTIVE AND RATIONALE: We aimed to conduct a systematic review of all aspects of ovarian stimulation for freeze-all cycles. SEARCH METHODS: A comprehensive search in Medline, Embase and The Cochrane Library was performed. The search strategy included keywords related to freeze-all, cycle segmentation, cumulative live birth rate, preimplantation genetic diagnosis, preimplantation genetic testing for aneuploidy, fertility preservation, oocyte donation and frozen-thawed embryo transfer. We included relevant studies published in English from 2000 to 2018. OUTCOMES: Our search generated 3292 records. Overall, 69 articles were included in the final review. Good-quality evidence indicates that in freeze-all cycles the cumulative live birth rate increases as the number of oocytes retrieved increases. Although the risk of severe ovarian hyperstimulation syndrome (OHSS) is virtually eliminated in freeze-all cycles, there are certain risks associated with retrieval of large oocyte cohorts. Therefore, ovarian stimulation should be planned to yield between 15 and 20 oocytes. The early follicular phase is currently the preferred starting point for ovarian stimulation, although luteal phase stimulation can be used if necessary. The improved safety associated with the GnRH antagonist regimen makes it the regimen of choice for ovarian stimulation in freeze-all cycles. Ovulation triggering with a GnRH agonist almost completely eliminates the risk of OHSS without affecting oocyte and embryo quality and is therefore the trigger of choice. The addition of low-dose hCG in a dual trigger has been suggested to improve oocyte and embryo quality, but further research in freeze-all cycles is required. Moderate-quality evidence indicates that in freeze-all cycles, a moderate delay of 2-3 days in ovulation triggering may result in the retrieval of an increased number of mature oocytes without impairing the pregnancy rate. There are no high-quality studies evaluating the effects of sustained supraphysiological estradiol (E2) levels on the safety and efficacy of freeze-all cycles. However, no significant adverse effects have been described. There is conflicting evidence regarding the effect of late follicular progesterone elevation in freeze-all cycles. WIDER IMPLICATIONS: Ovarian stimulation for freeze-all cycles is different in many aspects from conventional stimulation for fresh IVF cycles. Optimisation of ovarian stimulation for freeze-all cycles should result in enhanced treatment safety along with improved cumulative live birth rates and should become the focus of future studies.