Preimplantation genetic testing and frozen embryo transfer synergistically decrease very pre-term birth in patients undergoing in vitro fertilization with elective single embryo transfer

Abstract

Purpose

To study the effects of frozen embryo transfer (FET) and FET post-PGT on pre-term and very pre-term births in patients undergoing in vitro fertilization (IVF).

Materials and methods

A study was conducted using the SART National Summary Report from 2014 to 2017. Cycle inclusion criteria were eSET, fresh embryo transfers (ET), frozen embryo transfers without PGT (FET), and frozen embryo transfers with PGT (FET/PGT). Exclusion criteria were use of gestational carriers and donor eggs. Pregnancy outcomes included live births and gestational age at birth.

Results

A total of 161,550 eSETs were analyzed for the effect of FET and FET/PGT on IVF outcome and pre-term births including 43,618 ET, 58,812 FET, and 59,120 FET/PGT cycles. Live birth rates in patients with FET/PGT were significantly higher than those in ET (52.9% vs 46.4%, P < 0.0001) and FET (52.9% vs 43.1%, P < 0.0001). Patients with FET had a significantly lower live birth rate compared with that of ET (43.1% vs 46.4%, P < 0.0001). Both FET and FET/PGT significantly decreased total pre-term births compared with ET (10.8% and 10.5% vs 11.5%, P < 0.05 and < 0.001). FET/PGT significantly reduced very pre-term births when compared with ET and FET (1.5% vs 2.0%, P < 0.0001 and 1.5% vs 1.9%, P = 0.0002).

Conclusion

This study demonstrates that PGT significantly improves IVF outcome. Moreover, patients undergoing FET/PGT had significantly decreased total pre-term births. More importantly, patients with FET/PGT had significantly lower very pre-term births.

Introduction

Pre-term birth is defined by the International Federation of Gynecology and Obstetrics (FIGO) and the World Health Organization (WHO) as birth of an infant prior to 37 weeks and 0 days gestation. Pre-term birth rates in the USA declined from 2007 to 2014, likely secondary to decreases in births to teens and young mothers [1]. In the following years, however, pre-term birth rates have risen, and accounted for nearly 10% of all births in 2017, despite a 7% decline in the teen birth rate [2]. Pre-term birth not only increases perinatal morbidity and mortality for both child and mother, but also negatively affects the long-term health of the infant [3, 4]. Infants delivered pre-term may have impaired neurologic development with associated increases in risk for psychiatric disorders in adulthood, as well as a higher risk for developing type 2 diabetes, hypertension, and cardiovascular disease in later life [4–6]. Multiple risk factors have been linked to pre-term birth, including age, ethnicity, and socioeconomic status [7]. Pre-term birth has been associated with, or is a sequela of, pre-term pre-labor rupture of membranes (PPROM), maternal or fetal infections such as chorioamnionitis, gestational and pre-gestational diabetes, pre-eclampsia and its associated spectrum of disorders, and other causes [8, 9].

Many studies have demonstrated that compared to the general population, patients who conceive through assisted reproductive technologies experience significantly higher rates of pre-term births, which may be associated with multiple gestations, higher average maternal age, suboptimal implantation environment due to supra-physiologic estrogen levels during IVF stimulation, and donor egg usage [10–13]. Increased prevalence of pre-term birth and low birth weights have also been found among singletons conceived through IVF [13, 14]. It is not known which component of IVF contributes most significantly to adverse perinatal outcomes. Frozen embryo transfer (FET) has become increasingly more common in IVF treatment, most commonly in association with preimplantation genetic testing (PGT), and in some instances to decreased chances of ovarian hyperstimulation syndrome. While some studies have suggested that FET may be associated with improved placentation, reduced pre-term births, and decreased rates of low birthweight infants [14, 15], these findings have not been affirmed in recent randomized controlled studies of fresh vs frozen embryo transfers [16–18].

PGT has become a significant component of contemporary IVF treatment. The number of IVF cycles with PGT has more than doubled from 2014 to 2017 based on SART data (Figure 2). Studies have shown that PGT for aneuploidy (PGT-A) significantly increased implantation and live birth rates, as well as led to reduced miscarriage rates and time-in-treatment to achieve a healthy live birth [19–21]. Studies on the effects of PGT on neonatal outcome are relatively rare. Zhang et al. [22] reported an association between trophectoderm biopsy and pre-eclampsia, but found no adverse effect on neonatal outcomes. Bay et al. [23] also reported that IVF pregnancies with PGT had comparable placenta previa, pre-term birth, and neonatal outcomes with pregnancies after IVF without PGT. One clinical trial did show that PGT reduced pre-term birth, low birthweight, and neonatal intensive care unit (NICU) admission, but the study did not separate singletons from twins, which may have compromised its conclusions [24]. All these are single-center studies with relatively limited patient numbers. In this study, we used large cohort SART data from 2014 to 2017 to determine the effects of FET and FET/PGT on pre-term and very pre-term births in patients with eSET.

Fig. 2.

Trends of ART cycle numbers

Materials and methods

In the USA, a majority of IVF programs report outcomes to SART, which then transfer these registry data to the National Assisted Reproductive Technology Surveillance System (NASS). IVF outcomes are then published annually on the SART.org website as the National Summary Report. We used the National Summary Report data from 2014 to 2017 for this study. Inclusion criteria were eSET, ET, FET, and FET/PGT. Exclusion criteria were use of gestational carriers, donor eggs, and donated embryos. The cohort flow diagram of the study is illustrated in Figure 1. In total, 901,721 cycles were reported from 2014 to 2017. After exclusion of gestational carriers and transfers with multiple embryos, 195,570 transfers were by eSET on day 5/6. After exclusion of donor eggs and donated embryos, a total of 161,550 transfers were by eSET with autologous eggs and were included in the study for final analyses. Among them, 43,618 cycles were for ET, 58,812 for FET, and 59,120 for FET/PGT. Key pregnancy outcomes included live births and gestational age at birth (term, ≥37 weeks; pre-term, ≥ 32 weeks to <37 weeks; and very pre-term, <32 weeks). The main outcomes were described as percentages. Chi-square test was used for statistical analyses between categorical variables. Two-tailed P values of <0.05 were considered statistically significant.

Fig. 1.

Study design flow diagram

Results

A total of 161,550 eSETs were analyzed for the effect of FET and FET/PGT on IVF outcome and pre-term births including 43,618 ETs, 58,812 FETs, and 59,120 FET/PGTs. Primary outcomes are summarized in Table 1. The live birth rate in patients with FET/PGT was significantly higher than those among patients in ET (52.9% vs 46.4%, P < 0.0001) and FET (52.9% vs 43.1%, P < 0.0001) groups. Live birth rate in patients with FET was significantly lower than that of ET (43.1% vs 46.4%, P < 0.0001). The same patterns were found with regard to implantation rates (Table 1).

Table 1.

IVF outcome comparison among ET, FET, and FET/PGT

	ET	FET	FET/PGT
# of transfers	43,618	58,812	59,120
Singletons (%)	98.2	98.6	98.6
Twins (%)	1.7	1.4	1.4
Triplets or more (%)	0.01	0.02	0.01
Live births (%)	46.4	43.1	52.9
Total pre-term birth (<37 weeks, %)	11.5	10.8	10.5
Pre-term birth (32–37 weeks, %)	9.5	8.9	9.0
Very pre-term birth (<32 weeks, %)	2.0	1.9	1.5
Clinical pregnancies (%)	54.8	53.6	61.9
Miscarriage rate (%)	13.5	18.0	13.0
Implantation rate (%)	52.8	50.1	58.9
	P values
Summary of statistical analyses	FET vs ET	FET/PGT vs ET	FET/PGT vs FET
Singletons (%)	0.006	0.0008	>0.05
Twins (%)	0.008	0.002	>0.05
Triplets or more (%)	>0.05	>0.05	>0.05
Live births (%)	<0.0001	<0.0001	<0.0001
Total pre-term birth (<37 weeks, %)	<0.05	0.001	>0.05
Pre-term birth (32–37 weeks, %)	>0.05	>0.05	>0.05
Very pre-term birth (<32 weeks, %)	>0.05	<0.0001*	0.0002**
Clinical pregnancies (%)	<0.05	<0.0001	<0.0001
Miscarriage rate (%)	<0.0001	>0.05	<0.0001
Implantation rate (%)	<0.0001	<0.0001	<0.0001

*FET/PGT vs ET: P < 0.0001, RR 0.767, 95% CI 0.673–0.875

**FET/PGT vs FET: P = 0.0002, RR 0.787, 95% CI 0.694–0.891

Total pre-term birth rates in patients with FET or FET/PGT were significantly lower compared with that in the patients with ET (10.8% and 10.5% vs 11.5%, P < 0.05 and P =0.001, Table 1), and no statistical difference was found between FET and FET/PGT (Table 1). There were no statistically significant differences among ET, FET, and FET/PGT in terms of pre-term births (P > 0.05); however, patients with FET/PGT had very significantly lower very pre-term births when compared with ET and FET (1.5% vs 2.0%, P < 0.0001 and 1.5% vs 1.9%, P = 0.0002, Table 1).

Figure 3 shows the pattern of term, pre-term, and very pre-term births among different age groups. Patients with FET and FET/PGT had higher term births in all age groups when compared with ET (Figure 3a). Compared with FET, patients with FET/PGT had higher term births in all age groups except for the 35–37 age group (Figure 3a). Term birth in the FET/PGT (89.8%, n = 25,117) was significantly higher than that of the ET group (88.5%, n = 31,903, P = 0.0007) at the < 35 age group. There was no statistical difference in all other age groups. Patients with FET had lower pre-term births compared with ET except for the >40 age group (Figure 3b). Patients with FET/PGT had consistently lower pre-term births than those with ET in all age groups, but only FET/PGT vs ET had statistical difference at the <35 age group (8.7% vs 9.5%, P <0.05, Figure 3b). There was no statistical difference in all other groups. Figure 3c showed that patients with FET/PGT had lower very pre-term births in all age groups when compared with ET and FET, though the decrease was only statistically significant in the <35 age group (FET/PGT vs ET: P = 0.002, RR 0.89, 95% CI 0.83–0.95; FET/PGT vs FET: P = 0.001, RR 0.89, 95% CI 0.83–0.95).

Fig. 3.

Comparison of term (a), pre-term (b), and very pre-term (c) births among ET, FET, and FET/PGT in different age groups. a Percentage of term births in <35 age group: FET/PGT (n = 25,117, 89.8%) vs ET (n = 31,903, 88.5%), P = 0.0007; FET/PGT (n = 25,117, 89.8%) vs FET (n = 35,477, 89.1%), P = 0.08. There was no statistical difference among the other age groups. b Percentage of pre-term births in <35 age group: FET/PGT (n = 25,117, 8.7%) vs ET (n = 31,903, 9.5%), P < 0.05. There was no statistical difference among the other age groups. c Percentage of very pre-term births in <35 age group: FET/PGT (n = 25,117, 1.6%) vs ET (n = 31,903, 2.0%), P = 0.002, RR = 0.89, 95% CI 0.83–0.95; FET/PGT (n = 25,117, 1.6%) vs FET (n = 35,477, 2.0%), P = 0.001, RR = 0.89, 95% CI 0.83–0.95. There was no statistical difference among the other age groups

Discussion

In this study, we used SART annual report data with the largest available patient cohort to compare pre-term and very pre-term births in IVF patients after ET, FET, and FET/PGT with eSET. We found that both FET and FET/PGT were associated with decreased total pre-term births compared with ET. Of even greater clinical significance was the finding that FET/PGT was associated with significantly decreased rates of very pre-term births, a finding that was consistent among all age groups.

Though PGT has become one of the most important tools in the IVF treatment arsenal, and the number of IVF cycles with PGT has increased significantly in recent years (Figure 2), there are relatively few reports on subsequent neonatal outcomes and on the effect of PGT on pre-term births. Most of the aforementioned studies are single-center retrospective studies with limited sample size [22–25]. In a randomized study, Forman et al. [24] reported that patients undergoing embryo transfer after PGT had decreased risk of low neonatal birthweight, pre-term birth, and NICU admission. One major limitation of that study, however, was the lack of consistency in the number of embryos transferred between PGT and the control group. There were more twin births within the control group, which may have been the cause of increased pre-term births. Zhang et al. [22] found that trophectoderm biopsy and subsequent FET were associated with a higher rate pre-eclampsia, but found no statistically significant differences in pre-term births. Other studies showed that IVF pregnancies with PGT were associated with increased risk of placenta previa, delivery via cesarean section, and pre-term births when compared with spontaneously conceived pregnancies, but these increased risks were comparable to other IVF pregnancies [23].

Maheshwar et al. [14] reported that the relative risk of delivering at <37 weeks gestation was 0.84 in singleton pregnancies after FET compared with those after ET, with an absolute risk reduction of 2%. In this study, we found 0.7% decrease in births at <37 weeks gestation after FET compared with that after ET (10.8% vs 11.5%, P < 0.05, RR = 0.95). Our study also found that FET/PGT was associated with an absolute risk reduction of 1% in birth at <37 weeks gestation compared with ET (10.5% vs 11.5%, P < 0.001, RR = 0.92). Though there was no statistically significant difference in births at <37 weeks between FET/PGT and FET, the rate of very pre-term birth was significantly lower in the FET/PGT group compared with those of FET (1.5% vs 1.9%, P = 0.0002, RR = 0.787) and ET (1.5% vs 2.0%, P < 0.0001, RR = 0.767, Table 1). This finding suggests that the effect of PGT on pre-term births was independent from that of FET, or that there may be an additive effect on the beneficial effects of FET. The mechanism by which FET/PGT decreases pre-term births is not clear. Twin rates in both FET and FET/PGT were significantly lower than that in the ET group even with eSET, which may be one reason for higher pre-term births in the ET groups. On the other hand, however, very pre-term births in the FET/PGT group were significantly lower than that in the FET group despite similar twin birth rates (Table 1), suggesting that the use of PGT may have a direct effect on the rate of pre-term births. One hypothesis is that PGT may exclude some mosaic embryos or embryos with minor DNA deletions from transfer. In IVF cycles which did not utilize PGT, these embryos may have been selected based on normal morphology, transferred, implanted successfully, and ultimately led to live birth, but potentially with a greater rate of pre-term births. Studies have shown that mosaicism is predominantly confined to the placenta as opposed to true fetal mosaicism, with confined placental mosaicism rates as high as 89% [26, 27]. As the placenta plays a critical role in fetal development and the maintenance of a healthy pregnancy, a significant degree of mosaicism, and potentially abnormal function within the placenta, may be a cause of decreased placental viability and subsequently of pre-term birth. Additional risk factors for pre-term birth such as intra-uterine growth restriction may also be a consequence of increased placental mosaicism [26]. This study demonstrates decreased very pre-term births among all age groups, but there was a statistically significant decrease only in the <35 age group (Figure 3). Kahraman et al. (unpublished data) found significantly higher proportion of mosaicism in that age group, and a recent study demonstrated that the impacts of mosaic traits (level and type) on the pregnancy/live birth and miscarriage were different [28], which further supports our hypothesis that reduction of mosaic embryo transfers after PGT may be one of the mechanisms by which FET/PGT reduced very pre-term births.

A small randomized trial has shown that freezing all embryos in a fresh IVF cycle followed by FET in subsequent cycles improved pregnancy rates [29]. Chen et al. [16] have shown that elective FET resulted in a higher live birth rate than ET among anovulatory women with the polycystic ovary syndrome. On the other hand, elective FET among infertile women in ovulatory women with infertility did not result in significantly higher live birth rates than ET [17]. The results of this study show that both live birth and embryo implantation rates were significantly lower in FET compared with those in ET, and the miscarriage rate in the FET was significantly higher than that in the ET (Table 1). FET/PGT, however, significantly increased implantation and live birth rates compared to the ET and FET groups. The lower live birth rate in the FET group compared with ET may have resulted from the increased miscarriage rate. Increased miscarriage may stem from a variety of reasons such as poor blastocyst quality in the FET group, possibly due to damage during the freezing/thawing process, or may be related to poorer embryo quality de novo. The improved live birth and implantation rates from the FET/PGT group suggest a greater likelihood that the increased miscarriage risk in the FET may not be related to blastocyst damage during the freezing/thawing process, but more likely from lower embryo quality.

The strength of this study is the use of large cohort data, covering the majority of IVF programs in USA. Use of such large cycle numbers derived from national data also reduces many of the limitations of previous single-center studies that explored similar outcomes. The inclusion of only eSET cycles also helped reduce the potential confounding effect of multi-gestation pregnancies on pre-term births, which was a significant limitation of previous studies [22–25]. The weakness of the study is that the data were collected from the SART Summary Report, so neonatal outcomes such as birth weight were not available. Other factors which may affect the risk of pre-term birth, such as patient BMI or etiology of infertility, were also not available. Other limitations are that the data were extrapolated from the SART Summary Report; the cohort size we utilized in this study significantly minimizes the chance that any error encountered from our extrapolated data would have impacted our findings given their statistical significance. Additionally, differences between PGT methodologies such as aCGH or NGS could have an impact on the PGT-A diagnosis, but data is not available to make such distinctions in our study.

In conclusion, PGT has developed into one of the most important aspects of IVF treatment. This study using large cohort SART data demonstrates that PGT significantly improves IVF outcome. Moreover, this study shows that patients undergoing PGT accompanied with subsequent FET had reduced pre-term as well as significantly reduced very pre-term births. Patients seeking infertility treatment should be counseled regarding the lower incidence of very pre-term birth associated with PGT. With a trend toward decreasing cost over time, it is reasonable to consider adoption PGT with FET as the standard of care in IVF treatment.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethics approval

Not applicable

Consent to participate

Not applicable

Consent for publication

Not applicable