Higher probability of live-birth in high, but not normal, responders after first frozen-embryo transfer in a freeze-only cycle strategy compared to fresh-embryo transfer: a meta-analysis.

STUDY QUESTION

Does the outcome of the comparison of live birth rates between the first frozen embryo transfer (ET) (in a freeze-only cycles strategy, i.e. frozen ET group) and a fresh embryo transfer (fresh ET group) differ considering the type of ovarian response?

SUMMARY ANSWER

Α significantly higher probability of live birth is present in high, but not normal, responders, after the first frozen ET in a freeze-only cycle strategy as compared to a fresh ET.

WHAT IS KNOWN ALREADY

It has been hypothesised that freezing all good embryos in a fresh in-vitro fertilisation (IVF) cycle and deferring embryo transfer in subsequent cycles may provide a more physiological endometrial environment for embryo implantation when compared to a fresh ET. However, currently, three relevant meta-analyses have been published with conflicting results, while none of them has taken into consideration the type of ovarian response. Recently, the publication of additional, large relevant randomised controlled trials (RCTs) in patients with different types of ovarian response makes possible the comparative evaluation of the first frozen ET (in a freeze-only cycle strategy) versus fresh ET, considering the type of ovarian response.

STUDY DESIGN, SIZE, DURATION: A systematic review and meta-analysis was performed aiming to identify RCTs comparing the first frozen ET (in a freeze-only cycle strategy) to a fresh ET. The main outcome was live birth, while secondary outcomes included ongoing pregnancy, clinical pregnancy, moderate/severe ovarian hyperstimulation syndrome (OHSS) and miscarriage.

PARTICIPANTS/MATERIALS, SETTING, METHODS: We identified eight eligible RCTs, including 5265 patients, which evaluated the first frozen ET in a freeze-only cycle strategy versus a fresh ET either in high responders (n = 4) or in normal responders (n = 4). No relevant RCTs were present in poor responders. Meta-analysis of weighted data using fixed and random effects model was performed. Results are reported as relative risk (RR) with 95% confidence interval (CI).

MAIN RESULTS AND THE ROLE OF CHANCE

Eligible RCTs were published between 2011 and 2018. Four RCTs (n = 3255 patients) compared the first frozen ET (in a freeze-only cycle strategy) to a fresh ET in normal responders and four RCTs (n = 2010 patients) did the comparison in high responders. In high responders, a significantly higher probability of live birth was observed in the frozen ET group when compared with the fresh ET group (RR: 1.18, 95% CI: 1.06-1.31; fixed effects model; heterogeneity: I2 = 0%; three studies; n = 3398 patients). However the probability of live birth was not significantly different between the frozen ET group and the fresh ET group in normal responders (RR: 1.13, 95% CI: 0.90-1.41; random effects model; heterogeneity: I2 = 77%; three studies; n = 1608 patients). The risk of moderate/severe OHSS was significantly lower in the frozen ET group when compared with the fresh ET group both in high (RR: 0.19, 95% CI: 0.10-0.37; fixed effects model; heterogeneity: not applicable; a single study; n = 1508 patients) and normal responders (RR: 0.39, 95% CI: 0.19-0.80; fixed effects model; heterogeneity: I2 = 0%; two studies; n = 2939 patients).

LIMITATIONS, REASONS FOR CAUTION: Considerable heterogeneity was present among the studies, regarding ovarian stimulation protocols and the triggering signal used for inducing final oocyte maturation as well as the cryopreservation methods, while the quality of evidence was poor for the live birth rate in high responders. Moreover, the analysis did not apply a standard for determining 'high' or 'normal' responders since the type of ovarian response followed the characterisation of populations as reported by the authors of the eligible studies.

WIDER IMPLICATIONS OF THE FINDINGS

A freeze-only cycle strategy should be the preferred option in high responders since it enhances the probability of live birth, while reducing the chance of moderate/severe OHSS. In normal responders, the same strategy could be applied, in the interest of patient safety or clinic convenience, without compromising the chances of live birth.

STUDY FUNDING/COMPETING INTEREST(S): No external funding was used and there were no competing interests.

PROSPERO REGISTRATION NUMBER

PROSPERO registration number: CRD42018099389.