Impact of trophectoderm biopsy on obstetric and perinatal outcomes following frozen–thawed embryo transfer cycles

Abstract

STUDY QUESTION

Does trophectoderm biopsy for preimplantation genetic testing (PGT) increase the risk of obstetric or perinatal complications in frozen–thawed embryo transfer (FET) cycles?

SUMMARY ANSWER

Trophectoderm biopsy may increase the risk of hypertensive disorders of pregnancy (HDP) in pregnancies following FET cycles.

WHAT IS KNOWN ALREADY

Trophectoderm biopsy has replaced blastomere biopsy as the standard of care to procure cells for PGT analysis. Recently, there has been concern that trophectoderm biopsy may adversely impact obstetric and perinatal outcomes. Previous studies examining this question are limited by use of inappropriate control groups, small sample size or reporting on data that no longer reflects current IVF practice.

STUDY DESIGN, SIZE, DURATION

This was a retrospective cohort study conducted at a single university-affiliated fertility center. A total of 756 patients who underwent FET with transfer of previously vitrified blastocysts that had either trophectoderm biopsy or were unbiopsied and resulted in a singleton live birth between 2013 and 2019 were included.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Obstetric and perinatal outcomes for patients aged 20–44 years who underwent FET with transfer of previously vitrified blastocysts that were either biopsied (n = 241) or unbiopsied (n = 515) were analyzed. Primary outcome was odds of placentation disorders including HDP and rate of fetal growth restriction (FGR). Binary logistic regression was performed to control for potential covariates.

MAIN RESULTS AND THE ROLE OF CHANCE

The biopsy group was significantly older, had fewer anovulatory patients, was more often nulliparous and had fewer embryos transferred compared to the unbiopsied group. After controlling for potential covariates, the probability of developing HDP was significantly higher in the biopsy group compared with unbiopsied group (adjusted odds ratio (aOR) 1.943, 95% CI 1.072–3.521; P = 0.029).There was no significant difference between groups in the probability of placenta previa or placenta accreta. There was also no significant difference in the rate of FGR (aOR 1.397; 95% CI, 0.815–2.395; P = 0.224) or the proportion of low (aOR 0.603; 95% CI, 0.336–1.084; P = 0.091) or very low (aOR 2.948; 95% CI, 0.613–14.177; P = 0.177) birthweight infants comparing biopsied to unbiopsied groups.

LIMITATIONS, REASON FOR CAUTION

This was a retrospective study performed at a single fertility center, which may limit the generalizability of our findings.

WIDER IMPLICATIONS OF THE FINDINGS

Trophectoderm biopsy may increase the risk of HDP in FET cycles, however, a prospective multicenter randomized trial should be performed to confirm these findings.

STUDY FUNDING/COMPETING INTEREST(S)

No specific funding was obtained for this study. The authors declare no conflict of interest.

TRIAL REGISTRATION NUMBER

NA.

preimplantation genetic testing, frozen–thawed embryo transfer, ART, perinatal complications, obstetric complications, trophectoderm biopsy

Introduction

There is growing concern that particular aspects of IVF protocols may increase the risk of obstetric and perinatal complications, most notably the risk of hypertensive disorders of pregnancy (HDP), which refers to a spectrum of hypertensive disorders in pregnancy ranging from gestational hypertension to eclampsia (Helmerhorst et al., 2004; Pandey et al., 2012; Tandberg et al., 2015; Luke, 2017; American College of Obstetricians and Gynecologists, 2018; Ernstad et al., 2019; Saito et al., 2019; Sunderam et al., 2019; Luke et al., 2020). Previous studies showed higher implantation and ongoing pregnancy rates after trophectoderm biopsy compared with cleavage stage blastomere biopsy for preimplantation genetic testing (PGT; Fritz, 2008; Scott et al., 2013). This finding suggested that trophectoderm biopsy was safer than blastomere biopsy and it consequently replaced blastomere biopsy as standard of care. However, there have been concerns raised recently regarding the potential detrimental effect of trophectoderm biopsy on outcomes (Paulson, 2019; Zhang et al., 2019).

While trophectoderm biopsy does not seem to adversely impact the developmental and the implantation potential of an embryo at the outset, it is possible that its deleterious effects may not manifest until later in pregnancy. Given that a sample of 5–10 cells is removed from the trophectoderm, which ultimately gives rise to the placenta, it is plausible that its disruption via biopsy to procure a sample for genetic analysis may damage the embryo and result in abnormalities of placentation. This could conceivably manifest in conditions such as HDP, placenta previa, placenta accreta or fetal growth restriction (FGR), all of which are associated with significant morbidity and mortality (American College of Obstetricians and Gynecologists, 2013, 2018, 2019; American College of Obstetricians and Gynecologists and Society for the Maternal-Fetal Medicine, 2018).

Previous studies exploring the question of whether trophectoderm biopsy negatively impacts obstetric and perinatal outcomes have produced mixed results that are either difficult to interpret or limited by the selection of control group and/or not applicable to current IVF practice (Desmyttere et al., 2012; Eldar-Geva et al., 2014; Forman et al., 2014; Bay et al., 2016; Jing et al., 2016). For example, one prior study reported a higher rate of HDP after trophectoderm biopsy in frozen–thawed embryo transfer (FET) cycles compared to blastomere biopsy in fresh cycles (Jing et al., 2016). This result is limited by the use of an inappropriate control group given that there is overwhelming evidence of higher rates of HDP after frozen compared to fresh cycles (Chen et al., 2016; Coates et al., 2017; Maheshwari et al., 2018). Other studies have included either only blastomere biopsies (Desmyttere et al., 2012; Eldar-Geva et al., 2014; Bay et al., 2016) or multiple births in their cohort (Forman et al., 2014).

Nevertheless, a recent study showed a higher incidence of HDP after trophectoderm biopsy following both fresh and FET cycles. However, these differences were no longer significant when the data were limited to only FET cycles, likely due to their small sample size (Zhang et al., 2019). The discrepancy in findings described above warrants further investigation as it has been estimated that PGT is being used in as much as 40% of all IVF cycles in the USA (Munné, 2018) and biopsied blastocysts are vitrified to be transferred in FET cycles owing to the logistical challenges of having results early enough for fresh transfer.

Consequently, our aim was to determine whether the risk of obstetric and perinatal complications, particularly disorders of abnormal placentation such as HDP and FGR, were higher after trophectoderm biopsy and blastocyst transfer in FET cycles compared to unbiopsied blastocysts. We sought to address the shortcomings of the aforementioned studies by only including FET cycles with transfer of vitrified–thawed blastocysts, which were biopsied or unbiopsied and resulted in a singleton live birth, to reflect modern practice.

Materials and methods

Study design

This was a retrospective cohort study performed at a single university-affiliated fertility center. Our objective was to assess whether there were differences in obstetric and perinatal outcomes in pregnancies resulting in singleton live births conceived from FET cycles that involved transfer of blastocyst(s) that had undergone trophectoderm biopsy for PGT compared to FET cycles with transfer of an unbiopsied blastocyst(s).

FET cycles performed between March 2013 and January 2019 were screened for inclusion. Only the first FET cycle per patient who was <45 years old resulting in a singleton live birth was included. Thus, a patient may have undergone prior FET cycles that did not result in pregnancy or livebirth, which were not included in this study. For both groups, some patients had undergone prior fresh embryo transfer, which may or may not have resulted in a live birth. However, as we describe later, we controlled for parity in our statistical analysis. Furthermore, only cycles in which vitrified–thawed blastocyst(s) derived from autologous oocytes were transferred in a natural or programmed cycle and resulted in a singleton live birth were included. Donor oocyte cycles, FET cycles resulting in multiple births, cycles in which cleavage stage embryos were transferred or those in which the embryos that were transferred had been cryopreserved using slow freezing techniques were excluded. Modified natural FET cycles using letrozole were also excluded. This study was approved by our university’s institutional review board and because it was a retrospective chart review a waiver for written consent was granted.

Treatment protocol

IVF stimulation protocols were selected based on individual patient factors and physician preference. Gonadotrophin dosing was also chosen based on individual patient parameters and adjusted according to patient response. Details regarding the protocols have been previously described (Diluigi et al., 2011; Johnston-MacAnanny et al., 2011). When ≥3 follicles reached ≥17 mm in diameter, the patients were instructed to inject either hCG 3300–10 000 IU s.c. (Pregnyl, Merck, Kenilworth, NJ, USA; Novarel, Ferring Pharmaceuticals, Parispanny, NJ, USA). GnRH agonist 1 mg (leuprolide acetate, Abbott Laboratories, Chicago, IL, USA) or a combination of the two medications for final oocyte maturation. Transvaginal ultrasound-guided oocyte retrieval was performed 35 h after trigger injection. Oocytes were fertilized with either ICSI or conventional insemination.

Only good-quality blastocysts according to Gardner criteria (3BB or higher) underwent trophectoderm biopsy (Gardner and Schoolcraft, 1999). If PGT was being utilized, the zona pellucida of the developing embryo was breached on Day 3 using multiple 150 µs laser pulses (constant 0.9 J) with either a Zilos or Lykos medical grade laser (Hamilton Thorne, Beverly, MA, USA). Five to ten cells were biopsied from the trophectoderm of expanding and fully expanded blastocysts. Following biopsy, the blastocysts were vitrified and the biopsy samples were sent to an outside facility for genetic analysis. Genetic analysis was performed with either array comparative genomic hybridization or next generation sequencing.

Similarly, for unbiopsied embryos, only good-quality blastocysts according to Gardner criteria (3BB or higher) (Gardner and Schoolcraft, 1999) were vitrified on Day 5 or 6. Vitrification was performed using the same method for both biopsied and unbiopsied blastocysts. Prior to vitrification, a laser pulse of 300 μs (constant 0.9 J) was applied to collapse the blastocoel using either a Zilos or Lykos medical grade laser (Hamilton Thorne, Beverly, MA, USA). The embryos were cryopreserved using a commercial vitrification kit (Cat #90133, Fujifilm, Santa Ana, CA, USA). Prior to embryo transfer, all blastocysts were warmed using a commercial warming kit (Cat #90137, Fujifilm, Santa Ana, CA, USA). Embryos were cultured for 1–2 h before transfer to allow for re-expansion.

FET was performed in either a natural or programmed cycle, as previously described (Kaye et al., 2018), based on the patient’s ovulatory status and physician preference. For both groups, we followed The American Society for Reproductive Medicine (ASRM) and the Society of Assisted Reproductive Technology (SART) committee opinion on guidance on number of embryos to transfer (ASRM & SART Committee Opinion, 2017). In terms of protocols, in brief, programmed cycles began with downregulation with a GnRH agonist in the luteal phase of the preceding cycle followed by increasing doses of oral estradiol (Estrace, Allergan, Madison, NJ, USA) or transdermal estradiol (Vivelle, Novartis, Cambridge, MA, USA) following menses. I.M. progesterone (Watson Pharmaceuticals, Copiague, NY, USA) was started when endometrial thickness reached ∼8 mm. FET was performed on the sixth day of I.M. progesterone. In a natural FET cycle, transvaginal ultrasound was performed on patients on cycle Day #10 to ensure a follicle had been recruited and the endometrial stripe measured ∼8 mm. Then, patients were monitored with daily bloodwork until the LH surge (LH ≥ 20 IU/l) was detected. Two days following the LH surge, the patient was instructed to start luteal supplementation with vaginal progesterone (Crinone, Merck Kenilworth, NJ, USA; or Endometrin, Ferring Pharmaceuticals, Parsipanny, NJ, USA) (Bartels et al., 2019). The embryo transfer was scheduled for 6 days following the LH surge. For FET cycles in which previously biopsied blastocysts were transferred, only euploid blastocysts were transferred.

Outcome measures

Demographic and baseline cycle information was collected from our fertility center’s electronic medical record (EMR) including age, BMI, smoking status, parity, history of pre-existing diabetes (DM) or chronic hypertension (HTN), primary infertility diagnosis, number of embryos transferred, use of PGT, indication for PGT (aneuploidy testing, detection of mono-genic gene condition or detection of structural rearrangement) and type of FET protocol (natural or programmed). Outcome data, including mode of delivery and occurrence of obstetric or perinatal complications, were collected from information that is stored in our EMR and reported to SART. Our procedure for collecting these data consists of phone calls from nurses to each patient following delivery, during which they solicit delivery details and information regarding any obstetric or perinatal complications.

The primary outcomes were rate of HDP and FGR. HDP included the spectrum of HDP including pre-eclampsia (with and without severe features), gestational hypertension, the syndrome of hemolysis, elevated liver enzymes and low platelet count (HELLP) or eclampsia. HDP was further classified into severe cases (eclampsia, HELLP syndrome, pre-eclampsia with severe features, pre-eclampsia with delivery prior to 32 weeks gestational age) and non-severe cases (gestational hypertension, pre-eclampsia without severe features, pre-eclampsia with delivery after 32 weeks gestational age). FGR was defined as birthweight percentile <10th percentile for gestational age at delivery calculated as a function of z score. Secondary outcomes included delivery details and rate of other obstetric complications (mode of delivery, placenta accreta, placenta previa, postpartum hemorrhage, placental abruption, chorioamnionitis, premature preterm rupture of membranes (PPROM) and gestational diabetes (GDM)) as well as perinatal outcomes (mean birthweight, proportion of low birthweight (≤2500 g) and very low birthweight (≤1500 g) infants, gestational age at delivery, neonatal intensive care unit (NICU) admission and overall birth defects (including major and minor)). Major birth defects were classified as major and minor as described in previous studies (Bonduelle et al., 2002). A major congenital anomaly was defined as a condition that reduces the viability or compromises the quality of life and requires medical treatment or a condition that causes functional impairment or requires surgical correction. The remaining malformations were considered minor. A minor malformation was distinguished from normal variation by the fact that it occurred in ≤4% of the infants from the same ethnic group.

Statistical analysis

Statistical analysis was done using IBM SPSS© Statistics version 26.0 (IBM, Chicago, IL, USA). Student’s t-test was used to analyze parametric continuous variables. Chi square or Fisher’s exact test was used for categorical variables and, where appropriate, post hoc analysis involving pairwise comparisons were performed if there are ≥3 independent groups, using z-test of two proportions with a Bonferroni correction. Continuous variables were presented as mean ± SD and categorical variables presented as percentage and count. Binary logistic regression was used to determine odds of developing obstetric and perinatal complications. Potential covariates that may be associated with abnormal placentation and other obstetric complications adjusted for in the model included age and BMI as continuous variables and smoking status, parity, history of DM or HTN, polycystic ovary syndrome (PCOS)/anovulation diagnosis versus other diagnoses, number of embryos transferred and type of FET protocol as categorical variables. Covariates associated with perinatal complications adjusted for in the model included age and BMI as continuous variables and smoking status, parity, history of DM and HTN as categorical variables. All variables were entered using a forced entry method, and all the predictor variables were tested in one block to assess their predictive ability while controlling for other predictors in the model. Results are presented as adjusted odds ratio (aOR) with 95% CI. A two-sided P-value of 0.05 was considered statistically significant.

Results

Baseline characteristics

Patients who transferred biopsied blastocysts were significantly older, more often nulliparous, had a fewer mean number of embryos transferred and more often utilized a natural FET protocol compared to patients in the unbiopsied group, respectively (P < 0.01) (Table I). In addition, the patients in the biopsy group more often had a diagnosis of recurrent pregnancy loss, whereas the patients in the unbiopsied group were more often anovulatory. The most common indication for PGT was aneuploidy testing (PGT-A: 88.0%), followed by testing for mono-genic gene conditions (PGT-M: 10.0%) and then structural rearrangements (PGT-SR: 2.1%). PGT-A was done for the indications of advanced maternal age and recurrent pregnancy loss, as well as entirely elective. Of all patients who did PGT-A, 66.5% (141/212) were ≥35 years old and 17.5% (37/212) had the diagnosis of recurrent pregnancy loss. There were no significant differences between groups in terms of smoking status, BMI, history of pre-existing DM or chronic hypertension.

Table I

Baseline characteristics for FET groups with and without trophectoderm biopsy.

Baseline characteristics	Biopsied	Unbiopsied	P-value*
Age (years, mean ± SD)	35.9 ± 4.1	33.8 ± 3.6	<0.001
Age group			<0.001
<35 years (%, n)	38.2 (92/241)	63.3 (326/515)
≥35 years (%, n)	61.8 (487/241)	36.7 (191/515)
BMI (kg/m2, mean ± SD)	26.4 ± 5.5	26.8 ± 6.2	0.371
BMI group			0.211
≤30 kg/m2 (%, n)	78.0 (188/241)	73.8 (380/515)
>30 kg/m2 (%, n)	22.0 (53/241)	26.2 (135/515)
Smoker (%, n)	3.3 (8/240)	3.1 (16/519)	0.884
Parity			<0.001
Parous (%, n)	8.7 (21/241)	31.7 (163/515)
Nulliparous (%, n)	91.3 (220/241)	68.3 (352/515)
History of pre-existing diabetes			1.000
Yes (%, n)	0.0 (0/241)	0.4 (2/515)
No (%, n)	100.0 (241/241)	99.6 (513/515)
History of chronic HTN			0.670
Yes (%, n)	0.4 (1/241)	1.0 (5/515)
No (%, n)	99.6 (240/241)	99.0 (510/516)
Diagnosis			<0.001
Unexplained (%, n)	25.7 (62/241)	18.6 (96/515)
PCOS/anovulation (%, n)	12.4 (30/241)	23.5 (121/515)
Male Factor (%, n)	22.0 (53/241)	29.7 (153/515)
DOR (%, n)	4.6 (11/241)	3.7 (19/515)
Endometriosis (%, n)	2.5 (6/241)	8.3 (43/515)
Fibroids (%, n)	4.1 (10/241)	2.1 (11/515)
RPL (%, n)	15.4 (37/241)	3.1 (16/515)
Tubal factor (%, n)	4.6 (11/241)	7.8 (40/515)
Other (%, n)	8.7 (21/241)	3.1 (16/515)
Cycle characteristics
Mean no. embryo transferred (n, mean ± SD)	1.1 ± 0.3	1.4 ± 0.5	<0.001
FET protocol			0.009
Programmed FET (%, n)	41.9 (101/241)	52.0 (268/515)
Natural FET (%, n)	58.1 (140/241)	48.0 (247/515)

FET, frozen–thawed embryo transfer; HTN, hypertension; PCOS, polycystic ovary syndrome; DOR, diminished ovarian reserve; RPL, recurrent pregnancy loss.

*Student’s t-test was used for continuous variables and chi square test was used for categorical values.

Obstetric outcomes

Trophectoderm biopsy was associated with an increased risk of HDP (aOR 1.943; 95% CI, 1.072–3.521; P = 0.029) when compared with unbiopsied blastocyst transfer after adjusting for potential covariates including age, BMI, smoking status, parity, history of pre-existing DM or HTN, diagnosis of PCOS/anovulation versus other diagnoses, number of embryos transferred and method endometrial preparation for FET cycle (Tables II and III). In addition, the adjusted odds of developing a severe case of HDP was 1.770 (95% CI, 0.383–8.183; P = 0.465) and of a non-severe case of HDP was 1.896 (95% CI, 1.014–3.547; P = 0.045) in the biopsy compared to the unbiopsied groups (Table II).

Table II

Obstetric outcomes for frozen–thawed embryo transfer (FET) groups with and without trophectoderm biopsy.

Obstetric complications	Biopsied	Unbiopsied	Adjusted OR*(95% CI)	Adjusted P-value
HDP (%, n)	12.4 (30/241)	9.3 (48/515)	1.943 (1.072–3.521)	0.029
Severe cases of HDP** (%, n)	13.3% (4/30)	10.4% (5/48)	1.770 (0.383–8.183)	0.465
Non-severe cases of HDP** (%, n)	86.7% (26/30)	89.6% (43/48)	1.896 (1.014–3.547)	0.045
Placental abruption (%, n)	1.7 (4/241)	0.6 (3/515)	1.684 (0.268–10.572)	0.578
Placenta accreta (%, n)	0.4 (1/241)	0.4 (2/515)	2.673 (0.194–36.806)	0.462
Placenta previa (%, n)	1.7 (4/241)	3.1 (16/515)	0.401 (0.116–1.392)	0.150
Postpartum hemorrhage (%, n)	1.7 (4/241)	1.0 (5/515)	1.398 (0.278–7.021)	0.684
Cesarean delivery (%, n)	43.5 (104/239)	50.8 (259/510)	0.626 (0.432–0.905)	0.013
Gestational diabetes (%, n)	10.0 (24/241)	8.9 (46/515)	1.049 (0.562–1.956)	0.881
Preterm premature rupture of membranes (%, n)	1.7 (4/241)	1.6 (8/515)	0.717 (0.183–2.811)	0.633

HDP, hypertensive disorders of pregnancy; OR, odds ratio.

*Binary logistic regression was used to calculate the adjusted OR and controlled for the following covariates: age, BMI, smoking status, parity, history of diabetes or hypertension, polycystic ovary syndrome/anovulation versus other diagnoses, number of embryos transferred diagnosis and type of FET protocol.

**Severe cases of HDP included diagnoses of the hemolysis, elevated liver enzymes and low platelet count syndrome, pre-eclampsia with severe features, eclampsia, pre-eclampsia with delivery ≤32 weeks gestational age; non-severe cases of HDP included gestational hypertension, pre-eclampsia without severe features, pre-eclampsia with delivery at >32 weeks.

Table III

Adjusted OR of predictive variables and participants with HDP.

Variables	HDP
	aOR*(95% CI)	P-value
PGT biopsy
Unbiopsied	1
Biopsied	1.943 (1.072, 3.521)	0.029
Age (years)	0.929 (0.870, 0.993)	0.957
BMI (kg/m2)	1.099 (1.057, 1.142)	<0.001
Parity
Parous	1
Nulliparous	0.552 (0.258, 1.179)	0.125
History of smoking
Non-smoker	1
Smoker	0.448 (0.058, 3.481)	0.442
Diagnosis
Other diagnoses	1
PCOS/anovulation	0.983 (0.531, 1.820)	0.957
FET protocol
Natural	1
Programmed	2.740 (1.526, 4.920)	0.001
Number of embryos transferred	1.252 (0.709, 2.210)	0.438

HDP, hypertensive disorders of pregnancy; aOR, adjusted odds ratio; PGT, preimplantation genetic testing; PCOS, polycystic ovary syndrome; FET, frozen–thawed embryo transfer.

*Binary logistic regression was used to calculate the adjusted OR and controlled for the following covariates: age, BMI, smoking status, parity, history of diabetes (DM) or hypertension (HTN), PCOS/anovulation versus other diagnoses, number of embryos transferred diagnosis and type of FET protocol. However, history of DM and HTN was not considered in the final model because there were only a few cases of DM and HTN.

Additionally, women who underwent trophectoderm biopsy had a significantly lower probability of cesarean delivery. There were no significant differences in adjusted odds of placental previa, placenta accreta, placental abruption, postpartum hemorrhage, PPROM or GDM between groups (Table II). There was also no difference between groups in the risk of chorioamnionitis between biopsied and unbiopsied groups (0.0% (0/241) versus 0.6% (3/515); P = 0.555), respectively.

When subgroup analysis was performed, there was no significant difference in the adjusted odds of HDP between the biopsy and unbiopsied groups in nulliparous patients undergoing first FET cycles that resulted in singleton live birth (1.683; 95% CI, 0.899–3.151; P = 0.103), however, this result was likely limited by the small sample size (biopsy group, n = 140; unbiopsied group, n = 104).

Perinatal outcomes

After adjusting for potential covariates, including BMI and age, smoking status, history of DM or HTN and parity, there was no significant difference in FGR between the biopsied and unbiopsied groups (aOR: 1.397; 95% CI, 0.815–2.395; P = 0.224) (Table IV). There were also no significant differences in mean birthweight (3286.8 ± 597.5 g versus 3347.8 ± 666.2 g; P = 0.228) and gestational age at birth (38.5 ± 2.2 versus 38.6 ± 2.6 weeks; P = 0.663) between biopsied and unbiopsied groups, respectively. There was also no significant difference between groups in regards to proportion of low (≤2500 g) and very low birthweight (≤1500 g) infants born. Finally, there was no significant difference between biopsied and unbiopsied groups in the rate of major birth defects (1.7 (4/241) versus 1.9% (10/515); P = 0.789) or minor birth defects (0.4% (1/241) versus 0.0% (0/515); P = 0.319), respectively.

Table IV

Perinatal outcomes for frozen–thawed embryo transfer groups with and without trophectoderm biopsy.

Perinatal complications	Biopsied	Unbiopsied	Adjusted OR*(95% CI)	Adjusted P-value
Gestational age <37 weeks (%, n)	15.4 (37/241)	12.4 (64/515)	0.789 (0.490-1.272)	0.331
Weight groups
Low birthweight (≤2500 g, %, n)	10.5 (25/239)	7.1 (36/508)	0.603 (0.336–1.084)	0.091
Very low birthweight (≤1500, %, n)	0.8 (2/239)	2.6 (13/508)	2.948 (0.613–14.177)	0.177
Fetal growth restriction (<10th percentile)	11.7 (28/239)	8.6 (44/510)	1.397 (0.815–2.395)	0.224
NICU admission (%, n)	5.4 (13/241)	6.6 (34/515)	0.810 (0.396–1.659)	0.565
Overall birth defects (%, n)	2.1 (5/241)	1.9 (10/515)	1.025 (0.312–3.372)	0.968

OR, odds ratio; NICU, neonatal intensive care unit.

*Binary logistic regression was used to calculate the adjusted OR and controlled for the following covariates: age, BMI, smoking status, parity, history of diabetes and hypertension.

Discussion

This study investigated whether trophectoderm biopsy adversely impacts obstetric and perinatal outcomes in FET cycles that involved transfer of only previously vitrified–thawed blastocysts, which resulted in a singleton live birth. We demonstrated that in this large US cohort, FET cycles with transfer of trophectoderm biopsied blastocyst were associated with an increased risk of HDP compared to cycles with transfer of unbiopsied blastocysts. However, trophectoderm biopsy did not increase the probability of other placentation disorders, such as FGR, placenta previa or placenta accreta. Furthermore, there was no difference in any of the other perinatal outcomes we assessed between groups. To our knowledge, this is the largest study to date to provide a detailed report on the aforementioned outcomes with respect to trophectoderm biopsy in only FET cycles.

Most notably our data demonstrated a significant increase in the risk of HDP in pregnancies from biopsied blastocysts in FET cycles. It should be noted, however, that when HDP was analyzed by severity, there was no significant difference between groups in the adjusted odds of severe cases of HDP, likely because of very few cases of severe HDP in each group. Nevertheless, the adjusted odds of non-severe HDP were significantly higher in the biopsy group, suggesting that while trophectoderm biopsy may increase the risk of HDP, it may be a mild effect.

There have only been three prior studies that have investigated differences in obstetric outcomes with trophectoderm biopsy. Forman et al. (2014) compared singe euploid blastocyst transfer to double unbiopsied blastocyst transfer and stated that they collected information regarding pregnancy complications but did not report specifically on rate of HDP. Our finding was in agreement with Jing et al. (2016) who demonstrated a higher rate of HDP in trophectoderm biopsied embryos. However, in this earlier study, they compared trophectoderm biopsy in FET cycles to blastomere biopsy in fresh cycles (Jing et al., 2016). Thus, it is challenging to determine whether the protocol (i.e. fresh versus frozen) or day of embryo biopsy is responsible for the result, especially as it is well established that the risk of HDP is significantly higher in FET compared to fresh cycles (Roque et al., 2013; Chen et al., 2016; Coates et al., 2017). Finally, a more recent study by Zhang et al. (2019) also concluded there was an increased risk of pre-eclampsia with trophectoderm biopsy (Zhang et al., 2019). Although these investigators analyzed recent data and limited the analysis to blastocyst transfer, one key difference between their study and ours is that their cohort included pregnancies from both fresh and FET cycles. When they limited their analysis to FET cycles only (which comprised >70% of their total included cycles), they found no significant difference in the rate of pre-eclampsia. Their result was likely limited by small sample size.

Placenta previa and accreta are also conditions associated with abnormal placentation. We did not find a significant difference between groups in these outcomes. Unlike our results, Bay et al. (2016) did report an increased risk of placenta previa with PGT, but only in comparison to spontaneous conception. When compared to IVF pregnancies without genetic testing, the groups were no longer significantly different (Bay et al., 2016). As such, it is possible that these conditions represent more severe defects of placentation, which are not significantly impacted by trophectoderm biopsy.

Finally, we found that the rate of cesarean delivery was significantly lower in the biopsy group. While the odds of cesarean were adjusted for parity, we did not have information regarding previous mode of delivery or indication for mode of delivery. Thus, it is difficult to determine specifically why the odds of having a cesarean delivery were significantly lower in the biopsy group.

In terms of perinatal outcomes, we found no significant difference in FGR, mean birthweight or proportion of low and very low birthweight infants between groups. Since the placenta is an important determinant of normal fetal growth, one can argue that if trophectoderm biopsy disrupts normal placentation, a significant difference should be observed in mean birthweight and the proportion of growth-restricted infants. It is possible that clinically significant differences in fetal weight may only become apparent with more severe cases of placental disorders, which is consistent with our finding that the significant differences were noted only in non-severe HDP and not severe HDP. Nevertheless, our findings are in agreement with older publications, which although reported on data from blastomere biopsy, similarly found no significant difference (Desmyttere et al., 2012; Eldar-Geva et al., 2014; Bay et al., 2016). In fact, Forman et al. (2014) actually found a significantly higher rate of low birthweight, preterm delivery and NICU admission after transfer of unbiopsied blastocysts compared with single trophectoderm biopsied blastocyst transfer. However, they compared single euploid transfer to double unbiopsied embryo transfer. As such, there were a significant number of twin pregnancies in the unbiopsied group, which likely accounts for their results. Furthermore, all transfers were done in fresh cycles (Forman et al., 2014).

Our study did not detect significant differences in the rate of congenital anomalies between groups. The rate of overall anomalies in the biopsy compared to the unbiopsied group was 2.1% versus 1.9%, respectively. This is fairly comparable to the average rate of congenital anomalies reported in the general population of 2–3% (CDC, 2007). Similarly, we also found no significant difference in the odds of chorioamnionitis between groups. The incidence of chorioamnionitis was lower than what has been reported in the general population, which is estimated to complicate ∼2–5% of term deliveries (American College of Obstetricians and Gynecologists, 2017). However, our result for this outcome was limited by sample size.

The major strengths of our study are the large sample size, the analysis of contemporary data and inclusion of only FET cycles. Previous studies, as noted earlier, are difficult to interpret in the context of current IVF practice because they report on outdated techniques such as cleavage stage blastomere biopsy rather than the current standard of care, trophectoderm biopsy (Desmyttere et al., 2012; Eldar-Geva et al., 2014; Bay et al., 2016; Jing et al., 2016). The generalizability of our results are further increased because we appropriately include only IVF pregnancies and we limited our analysis to only FET cycles, which is the cycle protocol predominantly used for transfer of biopsied embryos as genetic testing results are usually not available for several days to weeks, during which time embryos must be frozen and transferred in a subsequent cycle. Finally, we limited analysis to only singleton live births, thus eliminating the known adverse impact of multiple gestation on the obstetric and perinatal outcomes in question.

Likewise, it is known that multiple factors can modify risk of HDP in patients such as age, BMI and parity (American College of Obstetricians and Gynecologists, 2019). Moreover, several studies have shown higher rates of HDP in women with PCOS undergoing IVF (Vryonidou et al., 2005; Chen et al., 2007, 2016; Sazonova et al., 2012; Opdahl et al., 2015). In fact, earlier studies have shown that women with PCOS have reduced vascular compliance and endothelial dysfunction, which may explain the higher risk of hypertensive disease (Paradisi et al., 2001; Kelly et al., 2002).

In terms of ART, the method of endometrial preparation for FET has been also shown to significantly influence risk of HDP with higher risk observed in programmed cycles. This is consistent with the results of our multiple logistic regression as displayed in Table III, which shows FET protocol type may be more predictive of HDP than the use of trophectoderm biopsy, possibly owing to the lack of corpus luteum, which may secrete important protective vasoactive substances (Ernstad et al., 2019; Saito et al., 2019; Zhang et al., 2019). For example, corpora lutea are known to secrete potent vasodilatory substances, such as relaxin, which promote vascular compliance (Marshall et al., 2016). Consequently, we controlled for these aforementioned factors a priori, including age, BMI, parity and diagnosis, in our binary logistic regression model and found the adjusted odds of HDP remained significantly higher with trophectoderm biopsy.

Our study is limited because all data are derived from a single fertility center, although that also provides us detailed knowledge of the lab protocols performed. There may be variables, such as history of HDP in prior pregnancy or history of autoimmune disorders or renal dysfunction, that are known to increase the risk of HDP that are not recorded in our EMR. Furthermore, it is possible that our groups were heterogeneous. To control for this, we performed an additional analysis limiting to only first FET cycles in nulliparous patients and found no significant difference in the adjusted odds of HDP between groups, although this was limited by the small sample size. Furthermore, data on obstetric and perinatal complications are based on patient report, which is subject to recall bias. Ideally, we would have confirmed the delivery details with a patient’s delivery record in the EMR, but accessing these records is generally logistically difficult because patients have obstetric care in several different hospitals. Furthermore, the retrospective design of this study makes it subject to confounding and selection bias. Nevertheless, binary logistic regression was performed to control for potentially confounding covariates, but it is possible that there may be other confounders we did not account for or could not be completely controlled for with regression analysis. Finally, we did not perform a Bonferroni correction for multiple comparisons. Bonferroni adjustment may increase type II error and obscure truly important findings (Perneger, 1998; Armstrong, 2014). It has further been suggested that Bonferroni adjustment is more appropriate when there is no pre-established hypothesis or when there are multiple subsamples as in stratified group analyses performed (Perneger, 1998), which was not the case in our study.

Conclusion

Trophectoderm biopsy may increase the risk of HDP in FET cycles, although this may be mainly limited to non-severe cases of HDP. However, before definitive conclusions are drawn, further studies with larger sample size or prospective randomized trials should be performed to confirm these findings.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Acknowledgments

We would like to thank the physicians and nursing staff at CARS for their support.

Authors’ roles

R.M.: participated in study conception and design and execution, data analysis, manuscript drafting and critical discussion and final approval of manuscript. C.B., P.G., A.B., A.D., J.N., D.G. and C.B.: participated in study conception and design, critical discussion and final approval of manuscript. L.E.: participated in study conception and design and execution, data analysis, manuscript drafting and critical discussion and final approval of manuscript. All authors approved the final version and agree to be accountable for all aspects of the work.

Funding

No external funding was either sought of obtained for this study.

Conflict of interest

The authors declare no conflict of interest related to the present study.