Background
Controlled ovarian stimulation (COS) was reported to have enhanced the incidence of ovarian hyperstimulation syndrome (OHSS). Besides, a number of epidemiological and population-based studies have suggested that COS followed by fresh transfer may result in increased risk of perinatal outcomes in pregnancies [
1‐
6]. Compared with the in vitro fertilization/ intracytoplasmic sperm injection (IVF/ICSI) cycles which need complex stimulation protocols to gain multiple follicular growth, FET are simpler and safer, with only one aim of preparing a receptive endometrium. The superfluous viable embryos were cryopreserved, and would be transferred in the succeeding cycles. The accumulative success rate could be improved after one ovarian stimulation and retrieval cycle. Therefore, along with the refinement of laboratory techniques, the proportion of FET has increased [
7] dramatically since the first baby after the frozen-thawed embryo transfer (FET) cycle was born in 1984 [
8,
9].
With the wide use of FET, there were concerns about the negative effect of cryopreservation on the health of children born. A number of observational studies [
10‐
12] suggested that both obstetric and perinatal outcomes after FET are similar to those after IVF/ICSI cycles, whereas other investigators indicated that they were even better than those of fresh IVF/ICSI cycles [
13,
14]. These observational studies suggested that there was no significant difference in pivotal perinatal outcome because of the flaws in design of the study. Regrettably, there was no randomized controlled trial which compared obstetric and perinatal outcomes after FET cycles with those after fresh IVF/ICSI cycles.
Because these obstetric and perinatal outcomes were mostly resulting from multiple gestations, owing to multiple ETs, it is reasonable to compare the obstetric and perinatal outcomes in single embryo transfer (SET) cycles as they have the advantage of reducing multiple pregnancy rates. An early meta-analysis and systematic review [
1] compared the obstetric and perinatal outcomes after FET with singleton pregnancy with those after fresh IVF cycles, and concluded that pregnancies after FET cycles may have a better result with regard to the obstetric and perinatal outcomes. Although ten outcome measurements were evaluated, the studies and samples were relatively small.
Hence, it is urgent to perform a meta-analysis with larger samples to compare the obstetric and perinatal outcomes in singletons pregnancy after FET and IVF/ICSI. The present study includes additional three studies and aims to evaluate which one has a better perinatal outcome of singleton pregnancy, IVF/ICSI or FET.
Methods
Literature identification
MEDLINE, Google Scholar and the Cochrane Library were searched from inception until July 2015. The keywords were used to search relative studies: one including terms on obstetric or perinatal outcomes (obstetric outcome, obstetric complication, perinatal outcome, perinatal complication), the other one about reproductive techniques (in vitro fertilization, intracytoplasmic sperm injection, frozen embryo transfer, IVF, ICSI, FET). We combined these subsets with “AND” to get a number of publications associated with our analysis. Papers published in non-English were excluded. The papers were reviewed by two investigators independently, and a third author was needed when there was a disagreement.
Study selection and data extraction
Studies which compared the obstetric or perinatal outcomes in singleton pregnancy after IVF/ICSI vs. FET were selected. The primary outcome of interest was preterm birth and/or low birth weight and/or still birth and/or perinatal mortality and/or cesarean section. For studies to be eligible, 2 × 2 tables were used for outcome data extraction. We also recorded the treatment type, number of cycles and number of obstetric/perinatal complications. If necessary, we would contact research author to clarify the data. The quality of the observational studies was assessed by Newcastle-Ottawa Quality Assessment Scales [
15]. The quality of the publications included was evaluated by two reviewers, and a third reviewer was needed when there was any disagreement about inclusion.
Statistical analysis
Meta-analysis was attempted wherever appropriate. The data of each study was extracted in 2 × 2 tables. Odds Risks (ORs) and risk differences with 95 % Confidence Intervals (CI) was used to describe the dichotomous outcomes of each study. Forest plots were used to evaluate the heterogeneity of the exposure effects graphically and l2 was implied to assess the heterogeneity between studies. A sensitivity analysis was performed by altering the fixed-to-random effect analysis in the event of moderate heterogeneity (l2 > 50 %). A P-value of ≤ 0.10 rather than the conventional level of ≤ 0.05 was used to determine statistical significance because the X
2 test for heterogeneity has low power in a meta-analysis especially the study had a small sample size. RevMan 5.0 (Cochrane Collaboration, Oxford, UK) was implied for statistical analyses.
Discussions
So far, only one meta-analysis and systematic review [
1] has compared the perinatal and obstetric outcomes in singleton pregnancy after IVF/ICSI and FET. To our knowledge, the present meta-analysis is the largest in regard to sample size with 94,472 IVF/ICSI cycles and 32,439 FET cycles. In the present meta-analysis, 11, 12, 6, 6, 5 studies were included to compare the risk of preterm birth, LBW, still birth, perinatal mortality, and cesarean section respectively. Our results indicated that singleton pregnancy after FET was at a decreased risk of LBW and preterm birth; however compared with IVF/ICSI cycles, singleton pregnancy after FET has a higher risk of cesarean section, which was in accordance with the former meta-analysis by Maheshwari et al. [
1]. On the contrary, we found there was no significant difference in the risk of perinatal mortality. Additionally, the present study assessed the still birth for the first time, and the result demonstrated that there was no significant difference in the risk of still birth after the IVF/ICSI and the FET cycles.
The present meta-analysis suggested that the risks of preterm birth and LBW were decreased in singleton pregnancy subsequent to FET. Many other studies also found that there was lower risk of preterm birth, very preterm birth, low birth weight, small for gestational age, and perinatal mortality in FET pregnancies [
16‐
19]. Another research did not find significant difference in the birth weights and preterm birth rates between singleton FET pregnancies and singleton spontaneous conceptions [
20]. As preterm birth always accompanies with LBW, these two outcomes are related. A review by Evans et al. also concluded that FET was associated with reduced risk of ovarian hyperstimulation syndrome and improved outcomes for both mother and baby [
21]. Why FET cycles have better outcomes compared with fresh ET is still not clear. The possible explanations may be as follows:
Firstly, FET involves in mini-stimulation or even no stimulation for ovarian. The endometrium was in the state of physiological condition, which may have a positive influence not only on the endometrial receptivity and early implantation but also on placentation and subsequent fetal growth [
4,
21]. Two comparative studies also found that births from FET have a better perinatal outcome and a similar neonatal and birth outcome compared with fresh ET [
1,
20], which confirmed the above assumptions.
Secondly, ovarian stimulation with a supraphysiologic hormone level in fresh embryo transfer cycles has negative effect on endometrial receptivity and embryos development, and results in the asynchronism between the embryo and endometrium which have detrimental effect on the development of embryo. Other studies suggested that ovarian hyper-stimulation during fresh cycles change angiogenesis of endometrium and embryo imbed [
22‐
24].
Thirdly, the process of FET involving embryo cryopreservation and embryo thawing would weed out poor quality embryos, and permit top quality embryos to survive, leading to a better clinical outcome [
25]. In fresh IVF/ICSI cycles, embryos in normal morphology with less development potential are more likely transferred.
In the present study, we also revealed that there was similar risk of the perinatal mortality and still birth in singleton pregnancy between IVF/ICSI and FET. Whereas, the previous study done by Maheshwari et al. [
1] believed that a lower risk of perinatal mortality in singleton pregnancy was associated with FET. The difference in results may be because of the difference between studies included. Besides, the rate of cesarean section in pregnancies subsequent to FET was higher than that after IVF/ICSI. The possible reason may be that women undergoing FET were more likely to have previous cesarean sections compared with women undergoing fresh embryo transfer. Besides, pregnant women after FET may have attempted many times and conceived finally, and they considered the cesarean as a safer way to deliver and preferred to choose cesarean section.
The limitations of the present study embody the integral defect of studies included: variation in design, exclusion & inclusion criteria, definition of outcomes, methodological differences, small number of study subjects, imprecise information on drug exposures, and lack of adjustment for meaningful confounders. However, it was impossible for us to adjust for some confounders due to lack of individual patient data.
Notwithstanding these limitations, the present meta-analysis and systematic review provides a valuable summary of the results of published studies. From what we have discussed above, singleton pregnancy after FET has a lower risk of preterm birth, LBW than that after fresh IVF/ICSI-ET cycles, and has a similar risk of perinatal mortality and still birth with that after IVF/ICSI. With the improvement of cryopreservation facilities and techniques, elective cryopreservation for later use may be recommended. But the clinical and cost effectiveness of the elective cryopreservation as well as acceptability of infertile couple should be evaluated before this strategy applied into clinical practice.