Biochemical pregnancy loss is unrelated to embryo stage of development and euploidy at transfer: evidences from 2452 warming cycles

Study question: Is biochemical pregnancy loss (BPL) decreased in cryopreserved euploid blastocyst transfers after preimplantation genetic diagnosis for comprehensive chromosomal testing (PGD-A) compared to untested cleavagestage embryos and blastocysts? Summary answer: Biochemical losses have similar incidence in PGD-A and untested embryo/blastocyst warming transfers suggesting no causative relationship with duration of in-vitro culture and chromosome aneuploidies. What is known already: Biochemical pregnancy loss, defined as -hCG levels decline after a short period of positivity, has not known underlying causes. Embryonic chromosome anomalies have been advocated as the main cause. However, no direct evidences have been reported showing that biochemical losses occur as a consequence of a primary aneuploidy in the embryo. Blastocyst stage PGD-A has been indicated as an efficient approach to accurately identify euploid embryos. The comparison of pregnancy outcomes between warming IVF cycles with or without PGD-A represents a powerful approach to directly investigate causative relationships between chromosome aneuploidies, duration of in vitro culture and pregnancy losses. Study design, size, duration: A retrospective single center cohort study including all consecutive embryo transfer cycles from 2008 to 2015 was performed. Only warming cycles were analysed to avoid the potential for confounding due to hormonal imbalances deriving from ovarian stimulation. Furthermore, cleavage and blastocyst stage transfers were also evaluated separately. Three different groups were thus analysed: cleavage stage transfers without PGD-A (group A), blastocyst stage transfers without PGD-A (Group B) and blastocyst stage transfer with PGD-A (Group C). Participants/materials, setting, methods: -hCG was always measured on days 16th after ovulation/progesterone’s administration. BPL was defined as -hCG levels ≥10IU/L in 1occasions, but not sustained and in the absence of ultrasonographic intrauterine/extrauterine pregnancy. For group A and B, the number of embryos transferred was decided on the basis of medical and biological data, while for group C elective single euploid embryo transfers were performed. Logistic regression was used to test association between the 3groups and other cycle’s/patient’s covariates with BPL. Main results and the role of chance: 2452 warming cycles involving 3241 embryos (mean per transfer 1.32 0.56) were performed in natural (74.4%) or hormone replacement cycles (25.6%) on an endometrium ≥7mm and with a level TSH ≤2.5mU/L. Mean female age at treatment was 36.25 3.9, 35.4 3.9, 37.4 3.5 for group A, B and C, respectively (NS). 1062 (43.3%;95%CI:41.4– 45.3) positive pregnancy tests were recorded, 217/637 (34.1% 95%CI:30.4–37.9; mean number of embryos transferred 1.84 0.67) following warming cleavage stage embryos, 412/1026 (40.2; 95%CI:37.1–43.2, mean number of embryos transferred 1.3 0.45) following warming blastocyst stage embryos and 433/789 (54.9%; 95%CI:51.3–58.4, mean number of embryos transferred 1.0 0.13) following euploid blastocyst transfers. Overall, 161 biochemical pregnancy losses per positive -hCG value were observed (15.2%; 95%CI 13.0– 17.5). Biochemical pregnancy loss rate was 19.3% (42/217;95%CI 14.3–25.2), 13.8% (57/412; 95%CI 10.6–17.5), 14.3% (62/433; 95%CI 11.2–18.0) for cleavage stage, blastocyst and euploid blastocyst, respectively. Logistic regression analysis showed no significant association for all covariates tested to biochemical loss incidence. In particular, blastocyst stage with PGD-A did not modify the risk (OR 0.98; 95%CI 0.67–1.45; p 0.93) as well as the stage of transfer, cleavage or blastocyst, showed no effect (OR 1.36; 95%CI 0.83–2.22; p 0.18). Female age, male age, sperm factor, stimulation protocol, endometrial preparation protocol, coagulation factors, autoimmunity factors, were all unrelated to the risk for biochemical losses. Limitations, reasons for caution: BPL was evaluated only from day 16 onwards. More information about the incidence of subclinical pregnancy losses and related causative factors could derive from earlier evaluations. The retrospective nature of the study is also a reason for caution. Embryo’s mosaicism could not be ruled out from the analysis. Wider implications of the findings: This study provides evidences that biochemical losses are not due to aneuploidies. Furthermore, no association with extended in vitro culture and other patient and cycle’s characteristics were observed. Future studies looking at other genetic (non-chromosomal) and uterine factors are needed to unravel mechanisms underlying this relatively frequent adverse event reproduction. Trial registration number: none.