STUDY QUESTIONWhat is the cumulative live birth rate following a ‘freeze-all’ strategy compared with a ‘fresh-transfer’ strategy?SUMMARY ANSWERThe ‘freeze-all’ strategy resulted in a similar cumulative live birth rate as the ‘fresh-transfer’ strategy among high responders (>15 oocytes retrieved) but did not benefit normal (10–15 oocytes) and suboptimal responders (<10 oocytes).WHAT IS KNOWN ALREADYFrozen-thawed embryo transfer is associated with a decreased risk of adverse obstetric and perinatal outcomes compared with fresh embryo transfer. It is unclear whether the ‘freeze-all’ strategy should be offered to all women undergoing ART treatment.STUDY DESIGN, SIZE, DURATIONA population-based retrospective cohort study using data collected by the Victorian Assisted Reproductive Treatment Authority. This study included 14 331 women undergoing their first stimulated ART cycle with at least one oocyte fertilised between 1 July 2009 and 30 June 2014 in Victoria, Australia. Demographic characteristics, type of ART procedures and resulting pregnancy and birth outcomes were recorded for the stimulated cycle and associated thaw cycles until 30 June 2016, or until a live birth was achieved, or until all embryos from the stimulated cycle had been used.PARTICIPANTS/MATERIALS, SETTING, METHODSWomen were grouped by whether they had undergone the ‘freeze-all’ strategy (n = 1028) where all embryos were cryopreserved for future transfer, or the ‘fresh-transfer’ strategy (n = 13 303) where selected embryo(s) were transferred in the stimulated cycle, and remaining embryo(s) were cryopreserved for future use. A discrete-time survival model was used to evaluate the cumulative live birth rate following ‘freeze-all’ and ‘fresh-transfer’ strategy.MAIN RESULTS AND THE ROLE OF CHANCEA total of 1028 women undergoing ‘freeze-all’ strategy and 13 303 women undergoing ‘fresh-transfer’ strategy had 1788 and 22 334 embryo transfer cycles resulting in 452 and 5126 live births, respectively. Most women (61.3%) in the ‘freeze-all’ group had more than 15 oocytes retrieved in the stimulated cycle compared with 18.1% of women in the ‘fresh-transfer’ group (P < 0.001). For high responders (>15 oocytes), the cumulative live birth rate in the ‘freeze-all’ group was similar to the ‘fresh-transfer’ group (56.8% vs. 56.2%, adjusted hazard ratio (AHR) 0.90, 95% CI 0.77–1.04). However, the likelihood of a live birth was lower in the ‘freeze-all’ group compared with the ‘fresh-transfer’ group among normal responders (10–15 oocytes) (33.2% vs. 46.3%, AHR 0.62, 95% CI 0.46–0.83) and suboptimal responders (<10 oocytes) (14.6% vs. 28.0%, AHR 0.67, 95% CI 0.14–1.01). During the minimum follow-up time of 2 years, 34.1%, 24.4% and 8.4% of suboptimal, normal and high responders, respectively, in the ‘freeze-all’ group did not return for any embryo transfer after the stimulated cycle, whereas all women in the ‘fresh-transfer’ group had at least one embryo transferred in the stimulated cycle.LIMITATIONS REASONS FOR CAUTIONA limitation of this population-based study is the lack of information available on clinic-specific protocols for the ‘freeze-all’ strategy and the potential impact of these on outcomes. Data were not available on whether the ‘freeze-all’ strategy was used to prevent ovarian hyperstimulation syndrome (OHSS).WIDER IMPLICATIONS OF THE FINDINGSThis study presents population-based evidence on clinical efficacy associated with a ‘freeze-all’ and ‘fresh-transfer’ strategy. The ‘freeze-all’ strategy may benefit some subgroups of patients, including women who are high responders and those who are at risk of OHSS, but should not be offered universally. Clinicians should consider the potential impact of electively deferring embryo transfer on treatment discontinuation in choosing the optimal embryo transfer strategy for couples undergoing ART treatment.STUDY FUNDING/COMPETING INTEREST(S)No specific funding was received to undertake this study. There is no conflict of interest, except that M.B. is a shareholder in Genea Ltd.