Abstract Study question Do the associations between age, previous attempts, frozen and day-5 transfers relative to LBR and mLBR differ between national registries (UK vs France)? Summary answer Both registries agree age negatively correlates with LBR and mLBR; disagree on effects of frozen transfers, previous attempts and day-5 transfers on LBR and mLBR. What is known already Due to the high risks of multiple pregnancy, the focus of ART success was shifted from having a high number of live birth rate (LBR) to one full-term, healthy baby and a low multiple live birth rate (mLBR). Elective single embryo transfer (eSET) has been an effective strategy for reducing the risk of mLBR but in most guidelines, it is recommended for young patients (<35 years) only. There is extensive evidence from large-scale studies characterising how factors, seen in eSET policies, affect LBR. However, only a few studies focus on mLBR. Study design, size, duration A retrospective analysis of two national registry datasets, HFEA (UK) (N = 442042; 2014-2018) and ABM (France) (N = 305142; 2014-2018), evaluating how age (<35 yrs, 35-37 yrs, 38-40 yrs, >41 yrs), fresh/frozen embryo transfer (ET), previous attempts (0, 1, 2, 3, 4, 5 attempts), and embryo stage (day-5 or day-3 transfers) impact LBR and mLBR following eSET (control group) and multiple embryo transfer (MET) (treatment group). Participants/materials, setting, methods Statistical analyses of the datasets were conducted using adjusted two-sided odds ratios from chi-squared tests with 95% confidence intervals (α = 0.05) with Bonferroni correction. Main results and the role of chance Two registries agreed on the negative correlation between age and LBR (HFEA: eSET OR:1.00-0.79-0.73-0.57, p < 0.001; MET OR:1.00-0.85-0.74-0.55, p < 0.001; ABM: eSET OR:1.00-0.75-0.71-0.52, p < 0.001; MET OR:1.00-0.80-0.69-0.50, p < 0.001) and mLBR following MET (HFEA OR:1.00-0.79-0.70-0.71, p < 0.001; ABM OR:1.00-0.74-0.74-0.59, p < 0.001). mLBR following eSET remained constant with increasing age in both datasets (HFEA OR:1.00-0.91-0.92-1.34, NS; ABM OR:1.00-0.87-1.02-1.13, NS). LBR was significantly higher for fresh vs frozen eSET in both datasets (HFEA OR:1.36, p < 0.001; ABM OR:1.83, p < 0.001). However, HFEA showed no significant difference in LBR between fresh and frozen MET (OR:0.98, NS), whereas LBR in fresh MET was significantly higher in ABM (OR:1.28, p < 0.001). There was disagreement on fresh vs frozen ET mLBR (HFEA: eSET OR:1.10, NS; MET OR:0.97, NS; ABM: eSET OR:0.81, p < 0.05; MET OR:1.36, p < 0.001). Registry results differed regarding the impact of previous attempts on LBR and mLBR.Day-5 ETs had significantly higher LBR compared to day-3 (HFEA eSET OR:1.29, p < 0.001; MET OR:1.37, p < 0.001; ABM: eSET OR:1.50, p < 0.001; MET OR:1.35, p < 0.001). mLBR was significantly higher following day-5 MET (HFEA OR:1.59, p < 0.001; ABM OR:1.27, p < 0.001) but only significantly higher following eSET in the French registry (HFEA OR:0.64, NS; ABM OR:1.40, p < 0.01). Limitations, reasons for caution Limitations included selection bias due to the retrospective design of the study, as a result, the populations include predominantly caucasian, European populations. The conclusions drawn relate to the practices from particular countries, so it is important to evaluate multiple, more diverse datasets. Wider implications of the findings It is encouraging that both registries agreed on the effects of age, whilst remaining factors were not generalizable and require further assessment. This study demonstrated that even large, national registries contain inherent biases and policymakers should not draw policies based on one dataset, especially when prospective trial is not feasible. Trial registration number N/A
Read full abstract