P–410 Comparison of the incidence of miscarriage rates in frozen versus fresh embryo transfers in IVF/ICSI cycles: a meta-analysis of Randomized Controlled Trials (RCT)

Study question Is the incidence of miscarriage rates increased in frozen embryo transfer (ET) compared with fresh ET? Summary answer The frozen ET is not associated with an increased risk of miscarriage in comparison with fresh ET. What is known already Nowadays, the freeze-all policy has gained great popularity because it minimizes OHSS and it seems to improve the reproductive outcomes. This increase was possible thanks to progress in vitrification procedures. Nevertheless, some recent evidence suggests more caution on the wide use of elective frozen ET in a general IVF population for its possible effect on some pregnancy-related complications. Endometrial preparation schemes for frozen ET include natural cycles or, more frequently, artificial cycles in which the endometrium is prepared with exogenous steroids. In this last case, the corpus luteum is absent, but its demise could be correlated with some pregnancy-related complications. Study design, size, duration A meta-analysis, based on PubMed, Cochrane CENTRAL, EMBASE, was conducted to estimate and compare the miscarriage rate (MR) in frozen versus fresh ET. Following PICOS, inclusion criteria were: Population, patients undergoing IVF/ICSI; Intervention, frozen ET; Comparison, fresh ET; Outcome, MR (primary one); clinical pregnancy rate (CPR), ongoing pregnancy rate (OPR) and live birth rates (LBR) (secondary ones). Study design: RCT. Participants/materials, setting, methods Electronic and manual search, conducted from 1990 to 2020, yielded 269 studies. Two researchers reviewed the studies independently, excluding 247 and 14 studies after the first and second screening. The outcome data from the 8 included studies were combined using a Mantel–Haenszel model and applying the random effects models. The dichotomous data results of each study were expressed as risk ratio (RR) with 95% confidence intervals (CI). Heterogeneity was evaluated using the I2 statistic. Main results and the role of chance The eight studies comprise 6,934 participants, 3,450 undergoing frozen ET and 3,484 fresh ET (Ferraretti et al., 1999; Shapiro et al., 2011a; Shapiro et al.; 2011b; Chen et al., 2016; Shi et al., 2018; Le et al., 2018; Wei et al., 2019; Stormlund et al., 2020). Preliminary analyses excluded interaction between covariates. RR of miscarriage ranged from 0.29 (Chen et al., 2016) to 3.24 (Ferraretti et al., 1999). The level of heterogeneity (I2) was 48% and it was considered as moderate. MR resulted not significantly different in frozen [279/1,601 pregnancies] compared with fresh ET [308/1,458 pregnancies] (RR = 0.727, 95%CI=0.43–1.25, p = 0.25). By analysing the 5 trials in which the endometrium was prepared with steroids (Ferraretti et al., 1999; Shapiro et al., 2011a; Shapiro et al., 2011b; Chen et al., 2016; Le et al., 2018), MR resulted not significantly different in frozen [126/720] versus fresh ET [160/682] (RR = 0.605, 95%CI 0.22–1.66; p = 0.33; I2=43%). For the secondary outcomes, CPR (RR = 1.14; 95%CI=0.97–1.35; 0.56 vs 0.52 for frozen and fresh ET, respectively), OPR (RR = 1.12; 95%CI=0.95–1.33; 0.49 vs 0.46 for frozen and fresh ET, respectively) and LBR (RR = 0.93; 95%CI=0.60–1.44; 0.49 vs 0.47 for frozen and fresh ET, respectively) were not significantly different between the two groups. Limitations, reasons for caution The MR was not the primary outcome of the included RCTs. This could have determined a sample size not calibrated for the analysis of the primary outcome of the present meta-analysis. Wider implications of the findings: This meta-analysis indicates that MR was similar between frozen and fresh ET. The endometrial preparation, in artificial cycles without corpus luteum, does not seem to influence the good course of pregnancy. Based on these results, the choice of a fresh ET should not derive by the presumed reduction of MR. Trial registration number Not applicable