Single embryo transfer by Day 3 time-lapse selection versus Day 5 conventional morphological selection: a randomized, open-label, non-inferiority trial

STUDY QUESTION: Does single cleavage-stage (Day 3) embryo transfer using a time-lapse (TL) hierarchical classification model achieve comparable ongoing pregnancy rates (OPR) to single blastocyst (Day 5) transfer by conventional morphological (CM) selection? SUMMARY ANSWER: Day 3 single embryo transfer (SET) with a hierarchical classification model had a significantly lower OPR compared with Day 5 SET with CM selection. WHAT IS KNOWN ALREADY: Cleavage-stage SET is an alternative to blastocyst SET. Time-lapse imaging assists better embryo selection, based on studies of pregnancy outcomes when adding time-lapse imaging to CM selection at the cleavage or blastocyst stage. STUDY DESIGN, SIZE, DURATION: This single-centre, randomized, open-label, active-controlled, non-inferiority study included 600 women between October 2015 and April 2017. PARTICIPANTS/MATERIALS, SETTING, METHODS: Eligible patients were Chinese females, aged <= 36 years, who were undergoing their first or second fresh IVF cycle using their own oocytes, and who had FSH levels <= 12 IU/mL on Day 3 of the cycle and 10 or more oocytes retrieved. Patients who had underlying uterine conditions, oocyte donation, recurrent pregnancy loss, abnormal oocytes or <6 normally fertilized embryos (2PN) were excluded from the study participation. Patients were randomized 1:1 to either the cleavage-stage SET with a time-lapse hierarchical classification model for selection (D3 + TL) or blastocyst SET with CM selection (D5 + CM). All normally fertilized zygotes were cultured in Primo Vision. The study was conducted at a tertiary IVF centre (CITIC-Xiangya) and OPR was the primary outcome. MAIN RESULTS AND THE ROLE OF CHANCE: A total of 600 patients were randomized to the two groups, among which 585 (D3 + TL = 290, D5 + CM = 295) were included in the Modified-intention-to-treat (mITT) population and 517 (D3 + TL = 261, D5 + CM = 256) were included in the PP population. In the per protocol (PP) population, OPR was significantly lower in the D3 group (59.4%, 155/261) than in the D5 group (68.4%, 175/256) (difference: -9.0%, 95% CI: -17.1%, -0.7%, P = 0.03). Analysis in mITT population showed a marginally significant difference in the OPR between the D3 + TL and D5 + CM groups (56.6 versus 64.1%, difference: -7.5%, 95% CI: -15.4%, 0.4%, P = 0.06). The D3 + TL group resulted in a markedly lower implantation rate than the D5 + CM group (64.4 versus 77.0%; P = 0.002) in the PP analysis, however, the early miscarriage rate did not significantly differ between the two groups. LIMITATIONS, REASONS FOR CAUTION: The study lacked a direct comparison between time-lapse and CM selections at cleavage-stage SET and was statistically underpowered to detect non-inferiority. The subjects eligibility criteria favouring women with a good prognosis for IVF weakened the generalizability of the results. WIDER IMPLICATIONS OF THE FINDINGS: The OPR from Day 3 cleavage-stage SET using hierarchical classification time-lapse selection was significantly lower compared with that from Day 5 blastocyst SET using conventional morphology, yet it appeared to be clinically acceptable in women underwent IVF.