Comparison of perinatal outcomes between spontaneous vs. commissioned cycles in gestational carriers for single and same-sex male intended parents

Abstract

Purpose

To determine whether gestational carrier (GC) in vitro fertilization (IVF) cycles (commissioned cycles) for same-sex or single male intended parents have an increased incidence of adverse perinatal outcomes compared with spontaneous cycles in the same GCs.

Design

GC singleton pregnancies were identified from a database of 895 commissioned cycles from a large fertility center. Of these, 78 commissioned cycles met inclusion and exclusion criteria and were compared with 71 spontaneous cycles by the same GCs. The primary outcome was the composite score for adverse perinatal outcomes. Secondary outcomes included mode of delivery, birthweight, and gestational age. Chi-square test of association and Mann-Whitney U tests were used to compare categorical and continuous variables between the cohorts, respectively. Logistic and linear regressions controlling for GC age were constructed to determine the influence of GC cycle type on adverse perinatal outcomes.

Results

Commissioned cycles were significantly associated with adverse perinatal outcomes (25.6% vs. 9.9%; p = 0.02) and lower average gestational age (38.7 ± 1.5 vs. 39.4 ± 0.9; p < 0.001) compared with spontaneous cycles. Commissioned cycle increased the likelihood of adverse perinatal outcomes (OR 3.3; p = 0.03) and was a significant independent predictor of a lower average gestational age (β = 0.897; p < 0.001). There were no significant differences in the incidence of vaginal deliveries or cesarean sections between commissioned and spontaneous cycles.

Conclusions

Commissioned cycles confer a greater incidence of composite perinatal complications and were independently associated with a lower average gestational age when compared with spontaneous pregnancies carried by the same GC despite a confirmed healthy uterine environment, sperm samples, and donor oocytes.

Introduction

Assisted reproductive technologies (ARTs) and gestational carriers (GCs) are established treatment modalities for the management of infertility and for same-sex male relationships. The use of a GC has risen in the USA from approximately 1.0 to 2.5% of total in vitro fertilization (IVF) cycles between 1999 and 2013 [1]. Studies have shown that the use of a GC during an IVF cycle can lead to many benefits including higher live birth rates (LBRs) in cases of uterine factor infertility [2]. GC cycles using non-donor or donor oocytes were reported to have better implantation, clinical pregnancy, and LBR than non-gestational carrier cycles, although rates of multiple and preterm delivery increased as well [1], potentially from the higher incidence of double embryo transfer [3–5]. Additionally, LBR and perinatal outcomes between GC cycles and autologous IVF cycles were compared favorably including risk for preterm birth, early preterm birth, very low birthweight, and congenital anomalies [6, 7].

Despite the known benefits of a GC, more recent data have suggested an increased incidence of adverse perinatal and pregnancy outcomes associated with the use of ART even within GC cycles [8]. Additionally, GC cycles occasionally involve the use of donor oocytes. The use of a donor oocyte in and of itself can potentially increase the risk of adverse outcomes such as preterm birth through an elevated immune response, thereby adding an additional layer of ART-related mechanisms that may influence perinatal outcomes [9]. Such adverse effects are important to consider for patients whose only option for genetically similar offspring is to reproduce using a GC and donor oocytes, as is the case for same-sex male couples and single male intended parents [10, 11]. These patients often have to deal with questions surrounding choice of spermatozoa between partners for insemination; use of anonymous vs. known donor oocytes; and the financial, emotional, and legal burdens of gestational carrier use and parenthood [12, 13]. Additionally, gestational carriers themselves must endure the demands of pregnancy and accept the potential risk of complications. Therefore, appropriate counseling of the risks and benefits of third-party gestation is essential for patient-centered care.

However, criticisms of previous studies examining perinatal outcomes from ART included limitations such as using autologous eggs from an infertile intended parent (IP) that may be of decreased quality, increased age of the GC at the time of the commissioned cycle compared with the age at the time of the same GC’s spontaneous pregnancy, and changes in BMI of the GC between the commissioned and spontaneous cycles [14]. This study aims to overcome some of these limitations by uniquely exploring the use of GCs and egg donors for same-sex and single male intended parents to further isolate the role of the ART in resulting perinatal and pregnancy outcomes while also adjusting for age. We hypothesize that if ART influences embryo quality and the uterine immune response potentially leading to adverse perinatal outcomes, then such outcomes can possibly be observed in GCs undergoing IVF despite a healthy uterine environment, donor oocyte, and fertile male sperm.

Materials and methods

Design, setting, and patient population

A retrospective cohort analysis of perinatal and pregnancy outcomes in patients undergoing a GC cycle was performed. For simplicity, GC-IVF and commissioned cycles are terms that are used interchangeably to refer to the same type of cycle. IRB approval was obtained from a large fertility practice before starting the study. All GCs in this study have had at least one successful spontaneous pregnancy before their commissioned cycle and were identified through an agency with no direct relation to the intended parents. Such GCs who then underwent a commissioned frozen embryo transfer (FET) cycle for other intended parents were identified from a database of the fertility practice between 2012 and 2016 with uniformity in the labs, procedures, and techniques performed. In some instances, GCs underwent more than one FET cycle, often for the same intended parents. All successful FET cycles that a GC underwent and resulted in a live birth were included in this study. Inclusion criteria consisted of gestational carriers for same-sex male or single male intended parents. Exclusion criteria consisted of any cycles with missing records/data from either the IVF or spontaneous pregnancies, pregnancies that did not result in a live birth, and those that resulted in twin gestations due to the inherently higher incidence of complications resulting from multiple gestation pregnancies. Selection of study participants was based specifically on the use of a GC for same-sex male or single male intended parents and not on any specific demographic characteristic of the GC. History of miscarriage or preterm delivery did not preclude GCs from being included in this study. Fertile male sperm in this study was defined as sperm used from the male partner/donor with a normal semen analysis as defined by the World Health Organization (WHO) [15]. Data were collected regarding a GC’s most recent pregnancies prior to her commissioned cycle. In the instances where a GC had multiple spontaneous pregnancies that abided by the recommended pregnancy interval of 18 months to 5 years, each of those pregnancies was included [16, 17]. However, any spontaneous pregnancies greater than 5 years out from the start of the GC’s commissioned cycle were not included. Therefore, this parameter allowed for some GC to have up to 2 spontaneous pregnancies included in analysis. All pregnancies, whether spontaneous or commissioned, were at least 18 months apart from one another. Only agency donor oocytes were used in this study’s population with age of donor ranging from 21 to 28 years old per our fertility practices policies. Out of 895 commissioned frozen embryo transfer cycles, 105 met the above criteria, 78 of which resulted in a singleton pregnancy and 27 of which resulted in a twin gestation (Fig. 1).

Fig. 1.

Flowchart representation of gestational carrier study population selection

Endpoints

Demographic data was obtained through review of clinical charts and scanned documents. The primary endpoint was the composite incidence of perinatal complications which includes preterm delivery, postpartum hemorrhage, preeclampsia, gestational hypertension, gestational diabetes, IUGR, oligohydramnios, abnormal placentation, placental abruption, and NICU admission between the commissioned and spontaneous groups. Total frequency of each individual complication was also recorded. The secondary endpoints were mode of delivery, birthweight, and gestational age. Gestational age at birth was calculated using the date of embryo transfer. Standard of care practices as outlined and determined by the American College of Obstetricians and Gynecologists (ACOG) for diagnosis of perinatal and pregnancy outcomes was assumed to be followed by each GC’s general obstetrician for both spontaneous and commissioned pregnancies. Additionally, data on whether 1 or 2 embryos were transferred and whether preimplantation genetic testing for aneuploidy (PGT-A) was implemented or not was also collected to allow for stratification analysis based on these parameters.

Endometrial preparation

Preparation of the endometrium for embryo transfer was initiated by 2 mg of oral estrace once daily for cycle days 1–2, followed by two doses of 2 mg oral estrace BID for cycle days 3–4 and four doses of 2 mg oral estrace BID for cycle days 5–12. Starting day 13, 1 mL of 50 mg/mL progesterone in oil was given IM in addition to 4 doses of 2 mg oral estrace BID. Embryo transfer occurred on cycle day 18. This regimen was continued until day of pregnancy test, or 10 days after day of embryo transfer and, if positive, continued through early pregnancy.

Statistical analysis

Statistical analysis was performed using SPSS statistical software (IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY, IBM Corp.). GC commissioned and spontaneous cycles were treated as independent variables. Categorical data are presented as frequencies with percentages and were compared with the chi-square test of association. Fisher’s exact test was used when comparing individual pregnancy complications where the cell size was < 5. Continuous variables are presented as means with standard deviations or ranges and were compared with the non-parametric Mann-Whitney U test after determining non-normality of distribution for our data via the Shapiro-Wilk test. Age matching was performed between spontaneous pregnancies (controls) and commissioned pregnancies (cases). Both age-unadjusted and age-adjusted analyses of composite perinatal and pregnancy outcomes between the commissioned and spontaneous cycles were performed. Logistic and linear regression models were constructed to determine the influence of cycle type (commissioned vs. spontaneous) on the stated primary and secondary endpoints. A p value of less than 0.05 was considered to indicate statistical significance. In instances of multiple comparisons between individual complications, embryo transfer, and use of PGT-A, a Bonferroni correction with an adjusted p value of 0.0045 was utilized.

Ethical approval

The present study was approved by the New England Institutional Review Board as a retrospective research study and has a yearly renewal of IRB approval, the most recent of which was January 15, 2019.

Results

Demographic information of each GC, their spontaneous pregnancies, and their commissioned cycles are reported in Table 1. We identified 66 GCs that resulted in 78 commissioned cycles and 71 spontaneous pregnancies. Five GCs had two spontaneous pregnancies that met the parameters described in the “Materials and methods” section which accounts for 66 GCs having 71 spontaneous pregnancies that were analyzed. Similarly, 12 of the GCs underwent two successful FET cycles resulting in live births, accounting for 66 GCs having 78 FET cycles analyzed. Data for ethnicity and education level was missing for 8 and 11 GCs respectively; otherwise, every GC had complete demographic data. GCs averaged 2.8 term deliveries, 0.2 preterm deliveries, and 0.5 miscarriages. Of note, most GCs were multiparous before their IVF cycle. 81.8% of GCs were married and 60.6% have a college or university education. There was no loss to follow-up or exclusion of outliers.

Table 1.

Demographics of gestational carriers (n = 66 women with complete demographics)

Variable	Data
Cycle type (n = 149)
Commissioned	78
Spontaneous	71
Age (years)
Commissioned cycle range	23–45
Spontaneous cycle range	20–39
Commissioned cycle average	32.8 ± 4.3
Spontaneous cycle average	28.1 ± 4.5
BMI at commissioned cycle (kg/m2)	26.7 ± 4.3
Gravidity/parity before commissioned cycle
Gravidity	3.7 ± 1.8
Term	2.8 ± 1.0
Preterm	0.2 ± 0.5
SAB/miscarriage	0.5 ± 1.1
Live	3.0 ± 1.1
Relationship status
Married	54 (81.8)
Single	12 (18.2)
Ethnicity
White/Caucasian	53 (80.3)
Black/African American	2 (3.0)
Hispanic	2 (3.0)
Asian	1 (1.5)
Unknown	8 (12.1)
Educational level
High school	12 (18.2)
College/university	40 (60.6)
Graduate	3 (4.5)
Unknown	11 (16.7)

Data presented as mean ± SD or n (%)

Commissioned versus spontaneous pregnancy analysis

Age-unadjusted and age-adjusted perinatal and pregnancy complications and outcomes are listed in Table 2. Commissioned pregnancies were significantly associated with an increased composite incidence of perinatal complications compared with spontaneous pregnancies before and after adjusting for age (p = 0.01; p = 0.02). There was a higher frequency of preterm delivery (10.2% vs. 2.8%) and hypertensive disorders (9% vs. 1.1%) between the commissioned and spontaneous groups, respectively. Out of the 7 patients that had complications in the spontaneous pregnancy cohort, only 2 of those patients subsequently had a complication in their commissioned cycle. There was no association between type of cycle and mode of delivery whether data was adjusted for age or not (p = 0.53; p = 0.73). Commissioned pregnancies had a significantly lower average gestational age (38.7 ± 1.5 weeks vs. 39.4 ± 0.9 weeks; age-adjusted p = < 0.001) but a non-significant lower average birthweight (3455 ± 508 g vs. 3563 ± 472 g; age-adjusted p = 0.48).

Table 2.

Perinatal complications and pregnancy outcomes

	Commissioned (n = 78)	Spontaneous (n = 71)	Age-unadjusted p value	Age-adjusted p value
Composite score of perinatal complications
Total cycles	20 (25.6)	7 (9.9)	0.01	0.02
Individual incidence of perinatal complications*
Preterm delivery (< 37 weeks)	8 (10.2)	2 (2.8)	0.10
Preeclampsia/GHTN	7 (9)	1 (1.1)	0.07
GDM	1 (1.3)	1 (1.1)	1.0
Postpartum hemorrhage	2 (2.6)	1 (1.1)	1.0
IUGR	2 (2.6)	1 (1.1)	1.0
Oligohydramnios	2 (2.6)	0 (0)	0.50
Placental abruption	1 (1.3)	0 (0)	1.0
NICU	1 (1.3)	1 (1.1)	1.0
Abnormal placentation	1 (1.3)	0 (0)	1.0
Mode of delivery			0.53
Cesarean section	9 (11.5)	6 (8.5)
Vaginal	69 (88.5)	65 (91.5)
Pregnancy outcomes
Gestational age (weeks)	38.7 ± 1.5	39.4 ± 0.9	0.003	< 0.001
Birthweight (grams)	3455 ± 508	3563 ± 472	0.40	0.48

Data presented as mean ± SD or n (%). p < 0.05 considered significant

GHTN gestational hypertension, GDM gestational diabetes mellitus, IUGR intrauterine growth restriction, NICU neonatal intensive care unit

*Analyses compared via Fisher’s exact test and performed with Bonferroni corrections for multiple comparisons against an adjusted p value of 0.0045

Commissioned pregnancy sub-analysis

Commissioned pregnancies were subsequently stratified by (1) number of embryos transferred (1 or 2) and (2) the use of PGT-A to determine if differences existed in the incidence of adverse perinatal complications as shown in Table 3. There was no statistically significant difference between number of frozen embryos transferred or the use of PGT-A and the incidence of perinatal complications (p = 0.38; p = 0.63).

Table 3.

Perinatal complications by number of transferred embryos per cycle or use of PGT-A

	Cycles	Complications	p value*
Number of transferred embryos			0.38
DET	43 (55.1)	9 (20.9)
eSET	30 (38.4)	9 (30)
Use of PGT-A			0.63
PGT-A	37 (47.4)	9 (24.3)
No PGT-A	26 (33.3)	5 (19.2)

Data presented as n (%). p < 0.05 considered significant. All cycles resulted in singleton pregnancies

DET double embryo transfer, eSET elective single embryo transfer, PGT-A preimplantation genetic testing for aneuploidy

*Analyses compared via chi-square test of associations and performed with Bonferroni corrections for multiple comparisons against an adjusted p value of 0.0045

Bivariate logistic and multivariate linear regression analyses

A bivariate logistic regression analysis was performed to determine the independent effect of type of cycle on adverse perinatal outcomes while controlling for GC age. This analysis demonstrated that GC-IVF pregnancies were associated with an increased likelihood of an adverse perinatal outcome compared with GC-spontaneous pregnancies (odds ratio [OR] 3.3, 95% confidence interval [95% CI] 1.1–9.3; p = 0.03) (Table 4). When controlling for gestational carrier age in a linear regression analysis, type of cycle was found to be a significant independent predictor of gestational age with GC-IVF pregnancies associated with lower average gestational ages (β = 0.897, 95% CI 0.431–1.363, p < 0.001) (Table 5). No such significance was found between type of cycle and birthweight.

Table 4.

Logistic regression predicting likelihood of perinatal complications based on type of cycle

	OR	p value	95% CI for OR
Type of cycle age	3.3	0.03	1.1–9.3
Age	1.0	0.89	0.9–1.1

p of < 0.05 considered significant

OR odds ratio, CI confidence interval

Table 5.

Multivariate linear regression association of gestational age and birthweight with type of cycle

	β	p value	95% CI for β
Gestational age	0.897	< 0.001	0.4–1.4
Birthweight	− 81.8	0.39	− 268.7–105.3
Age	0.04	0.07	− 0.004–1.4

p of < 0.05 considered significant

β unstandardized beta, CI confidence interval

Discussion

The results of the present study showed an age-unadjusted and age-adjusted significant increase in the composite incidence of various adverse perinatal outcomes in commissioned versus spontaneous cycles. Absolute risk of perinatal complications was 25.6% and 9.9% in the commissioned and spontaneous groups, respectively. Logistic regression analysis also indicated a significantly increased likelihood of adverse perinatal outcomes with commissioned pregnancies being 3.3 times more likely to result in an adverse perinatal outcome than spontaneous pregnancies even after controlling for GC age. As indicated in the results, only 2 out of the 7 patients who had complications in the spontaneous cohort went on to have a complication in their commissioned cycle. Therefore, the additional 18 patients who had complications in the commissioned cohort had previously uncomplicated spontaneous pregnancies. There was also a significant decrease in average gestational age at time of birth in commissioned cycles. Similarly, when controlling for age, a commissioned pregnancy was found to be a significant independent predictor of a lower average gestational age via linear regression analysis. There was also a higher frequency of hypertensive disorders and preterm delivery. Medical conditions such as hypertensive disorders are associated with advanced maternal age and so it is important to control for this potential confounder, which was done in this analysis. Additionally, the lower average gestational age in commissioned cycles is clinically significant as early term births, defined as births between 37 0/7 and 38 6/7 weeks, are associated with worse perinatal outcomes and higher neonatal morbidity [18]. No such clinical significance can be assigned to the difference in average birthweight. Patient parity can also influence pregnancy outcomes especially as it pertains to nulliparous or grand-multiparous patients, defined as parity greater than or equal to 5. Data on complications due to grand-multiparity has been controversial but studies agree that there is a probable increase in the risk of placenta previa/abruption, postpartum hemorrhage, macrosomia, and umbilical cord prolapse [19, 20]. Only four GCs in this study qualified as grand-multiparous leading to three incidences of complications including preterm delivery and two cases of preeclampsia. Most GCs had parity between 1 and 4 and therefore, the outcomes of this study were not likely significantly influenced by parity. However, 3 out 20 (15%) of complications is still a considerable proportion of total complications and although overall significance is not affected when excluding grand-multiparous GCs (23.0% vs. 9.9% in commissioned vs. spontaneous cycles, p = 0.03), grand-multiparity may independently impact complication risk. This remains a clinical question to be addressed in larger prospective studies.

This study also examined mode of delivery between commissioned and spontaneous cycles and the incidence of perinatal outcomes in relation to number of embryos transferred or the use of PGT-A. There was no significant difference with mode of delivery between commissioned and spontaneous cycles. GCs are therefore not at an increased risk for a cesarean section by this analysis. When we stratified the data to compare the incidence of composite adverse perinatal outcomes in single versus double embryo transfers and whether PGT-A was implemented or not, there was no statistically significant difference. This finding is important due to the potential detrimental effect on perinatal outcomes when transferring 2 embryos since elective single embryo transfer is known to decrease the risk of preterm delivery and NICU admission [21]. However, our sample sizes were small and likely underpowered for this stratified analysis. Therefore, no definitive conclusions can be made, and further study is needed. Additionally, there are differences in endometrial preparation between spontaneous cycles where no hormone replacement/supplementation was provided, and FET cycles in this study that utilized the hormone replacement (HR) protocol described previously. There is a lack of large randomized control trials to compare outcomes in FET cycles vs. natural cycle FET or between various endometrial preparation protocols within FET cycles. Despite the scarcity of high-level data, studies have shown that there was an increase in the rate of miscarriage rate but no difference in implantation, clinical pregnancy or live birth rate, or pregnancy outcomes in hormone replacement FET cycles vs. natural cycle FET [22–24]. Therefore, the endometrial preparation in this study’s population was unlikely to significantly affect pregnancy outcomes.

Limitations of this study include the inability to control for potential confounding variables such as BMI, smoking status, and GC medical disorders and the inability to exclude pregnancies resulting after vanishing twin syndrome. Furthermore, our study’s cohort design limits the ability to control for all known and potential unknown confounding variables. The major limitation of this study was the small sample size. Evaluation of only 66 GCs resulted in 78 commissioned and 71 spontaneous pregnancies. However, a post hoc power analysis of proportions using an alpha error probability of 0.05 and the difference in the incidence of composite perinatal complications with our sample sizes indicated that the current study is powered at 71% for this outcome measure. Although this is slightly lower than 80% power, this serves as the best evidence that we currently have for these rare populations compared with current literature. Therefore, the study sample size is relatively large given the analysis of this study’s niche population of GC cycles that utilize donor oocytes. Nevertheless, the authors of this study agree with the opinions set forth by Spandorfer in that a large, multi-center prospective study is warranted and necessary [14]. Additionally, unknown perinatal factors influencing perinatal outcomes in a GC’s first commissioned cycle would be more likely to also influence the same GC’s second commissioned cycle. To minimize this possibility, a prospective study would only include the GC’s first commissioned cycle for analyses while also anticipating needing a larger sample size of patients as this would limit the number of data points. Our study was unable to overcome the potential vanishing twin effect due to both eSET and DET cycles being included in the analysis. Future studies that examine perinatal morbidity in only eSET cycles can help to remove this potential confounding variable. Moreover, another future study of interest could factor in the age of the sperm provider and the effect that may have on embryo quality and the intrauterine environment since fathers that conceive through surrogacy may on average be older than fathers that conceive through unassisted reproduction.

Additionally, there was difficulty in calculating the BMI from GCs’ spontaneous pregnancies due to missing data/charts, including a patient’s height and weight, from preconception or initial prenatal visits. Our study participant’s average BMI was mildly overweight at 26.7 kg/m² with a range of 19.7 to 38.1 kg/m². However, only two GCs had a BMI above 35 kg/m² and fourteen between 30 and 34.9 kg/m². Data surrounding the influence of BMI on pregnancy outcomes for patient’s undergoing fertility treatments varies. Wittemer et al. [25] found a decreased response to ovarian stimulation in obese patients but no difference in clinical pregnancy or miscarriage rate between underweight, overweight, and obese patients. Coyne et al. examined the effects of increasing BMI specifically in the GC population and found no significant differences in implantation, clinical pregnancy, or live birth rates as well as no difference in GDM, cesarean section rate, or preterm or NICU admissions. Therefore, although obesity can influence pregnancy outcomes, we do not believe that our data was significantly influenced by differences in BMI between spontaneous and commissioned pregnancies. GCs are also stringently vetted and should meet the standards proposed by ASRM including assessment of medical conditions, psychosocial stability, support, infectious disease, and virology screening as well as having a general healthy lifestyle which should preclude those who are current smokers from participating [26]. Donor oocytes undergo strict screening standards as well to ensure that healthy gametes are being used [27]. Our study therefore assumes that the GCs were mostly healthy and fertile participants with good psychosocial habits and support. Most of the GCs in this study were college educated with a support system in the form of a spouse or partner. Moreover, demographic factors are unlikely to dramatically change between spontaneous and commissioned cycles in the same GC.

Our results are supported by previous studies that also found associations between ART and significant increases in adverse perinatal and pregnancy outcomes including low birthweight, preterm birth, cesarean delivery, and abnormal placentation, among others [28–30]. These risks and others including hypertensive disorders seem to not be mitigated by the use of a donor oocyte [31–33]. The possible rationale behind these adverse perinatal outcomes are changes associated with epigenetic reprogramming and DNA methylation induced by ART, specifically laboratory procedures such as the use of culture medium and incubator systems [34, 35]. Woo et al. were able to provide a promising model for evaluating and determining IVF vs. maternal effect on perinatal outcomes. Similar to the findings in our study, those authors found a significant increase in adverse perinatal outcomes including preterm birth, hypertension, and abnormal placentation compared with spontaneously conceived pregnancies despite a confirmed healthy uterine environment [8].

However, critiques of the Woo et al. study included the lack of data on number of embryos transferred, differing lab techniques used between 1995 and 2010, and data being presented as age-unadjusted [14]. Additionally, there remains the question of whether observed adverse perinatal outcomes from the study are the result of unknown effects of the infertility condition of the intended parent donors despite the use of a known healthy GC uterine environment. Kapiteijn et al. [36] examined this possibility and showed that subfertility alone could not explain the incidence of poor perinatal outcomes since there were increased odds of low birthweight, very low birthweight, and preterm birth in a controlled ovarian stimulation/IVF group. Nevertheless, Woo et al. attempted to answer this question by comparing commissioned pregnancies between patients using autologous eggs versus donor eggs but had insufficient numbers for statistical significance. Our study can help add to this growing body of knowledge by addressing some of the above limitations and by evaluating embryos derived from donor oocytes and subsequently carried by a gestational carrier as Woo et al. had proposed.

The potential influence of age on perinatal outcomes can be related to the effects of age on the uterine microenvironment and was the motivation behind controlling for GC age in this study. Although there is a clear link between maternal age and a decline in fertility due to age-related effects on ovarian function and diminished ovarian reserve, the link to an aging uterine environment is less clear. It is plausible to think that GCs carrying an oocyte, whether from the intended parent or a donor, at increasing maternal ages can contribute to adverse perinatal outcomes observed in commissioned cycles. Segal et al. [37] showed that an increasing age in intended parents is associated with lower LBR and a higher rate of poorer neonatal outcomes such as low birthweight and preterm delivery even when using a donor oocyte. These findings suggest that the uterine microenvironment, independent of a healthy donor oocyte, plays an important role in pregnancy outcomes and can be influenced by increasing maternal age. Recent studies describe increases in proinflammatory elements and uterine and immune system senescence that lead to an age-dependent, low-grade, chronic, and systemic inflammatory state, as negatively affecting pregnancy via a dysfunctional immune tolerance [38]. Since our study was able show a significant increase in adverse perinatal and pregnancy outcomes despite the use of younger GCs and adjusting for age, ART may potentially play a more prominent role in influencing the uterine microenvironment leading to adverse pregnancy outcomes despite young and healthy uteri.

The possible effects of the donor oocyte, as a form of ART, on outcomes warrant discussion as well. The benefits of a donor oocyte are well known. A study by Yeh et al. [39] showed that donor oocyte recipients had significantly higher rates of implantation, clinical pregnancy, and LBR versus autologous fresh cycles, which was potentially attributed to the supraphysiological hormonal environment in autologous fresh cycles. However, the use of a donor oocyte may not be completely benign and can influence the immune response, thereby driving the increase in perinatal complications. Placentas examined from donor oocyte cycles were significantly more likely to demonstrate pathologic findings such as chronic villitis and deciduitis, increased perivillous fibrin, ischemic changes/infarction, increased proinflammatory immune expression, and intervillous thrombi [40, 41]. These significant histological and immunohistochemical changes in the placentas of donor oocyte cycles can potentially lead to adverse pregnancy outcomes. Furthermore, there was a higher rate of placental disorders of pregnancy such as gestational hypertension and preeclampsia in donor oocyte versus autologous IVF pregnancies [31]. Therefore, the increase in adverse outcomes in our study population, where a donor oocyte must be used, may also be related to the donor oocyte itself rather than just the previously discussed effects of ART on the uterine microenvironment and warrants further study comparing autologous FET and donor oocyte FET.

This study demonstrates the unique ability to further isolate the role of ART in perinatal and pregnancy outcomes with the goal of providing better patient education and counseling. Grover et al. highlighted the importance of counseling patients about the emotional and medical demands of having a child using ART, especially in same-sex and single male intended parents. Supportive psycho-educational counseling helps these intended parents to make informed decisions with regard to fertility treatment and creating a family [12]. Furthermore, gestational carriers assume potential physical risk in support of the reproductive interests of the intended parents as well as the risk that difficult ethical decisions might arise and therefore, GCs may be in even greater need for the ability to make informed decisions [42, 43]. These informed decisions can include the need for the GC to deliver at a tertiary medical center or to potentially have increased antepartum testing and high-risk obstetrical management. Additionally, the initiation of low-dose aspirin starting at 12 weeks gestation, as is done in patients with a history of preeclampsia, may be a viable preventative measure due to the potential independent risk factor of being a GC or using a donor oocyte on developing hypertensive disorders of pregnancy. Therefore, the importance of this data lies in educating and empowering both gestational carriers and same-sex or single male intended parents, and in general all intended parents seeking to utilize a GC, about the risks and benefits of ART/GC cycles so that all parties can make appropriate decisions on the journey to starting a family. Future studies can hopefully utilize the model proposed by Woo et al. in combination with the model used in this study, namely studying embryos derived from donor oocytes and subsequently carried by a GC, to further delineate the role of ART in perinatal and pregnancy outcomes while addressing previously listed limitations and controlling for potential confounders. A larger, multi-center prospective trial utilizing GCs, donor oocytes, and fertile male sperm can help to accomplish these goals and potentially analyze the specific physiology behind perinatal outcomes observed in ART.

Conclusion

This study uniquely analyzes the incidence of adverse perinatal and pregnancy outcomes while controlling for a healthy uterine environment, donor oocyte, and fertile male sperm. We accomplished this by examining embryos derived from donor oocytes and subsequently carried by GCs for either same-sex male or single male intended parents. Our data suggests that the use of ART in commissioned cycles may have an adverse effect on the quality of the embryos despite a confirmed healthy uterine environment, sperm sample, and donor oocytes. Therefore, although absolute risk of perinatal complications is low at 25.6%, GC/ART use may still be associated with higher adverse perinatal risks. Our observational study describes a trend but is not definitive and the authors of this study again encourage a larger, multi-center trial that can aid in further confirmation of these trends. Overall, this data can help drive counseling and recommendations for both gestational carriers and intended parents to ensure appropriate education and management are offered.

Compliance with ethical standards

The present study was approved by the New England Institutional Review Board as a retrospective research study and has a yearly renewal of IRB approval, the most recent of which was January 15, 2019.

Conflict of interest

The authors declare that they have no conflict of interest.