How much does the uterus matter? Perinatal outcomes are improved when donor oocyte embryos are transferred to gestational carriers compared to intended parent recipients.

In 2014, there were a total of 21,391 cycles using donor oocytes. Of these, 1,067 canceled cycles and 80 cycles with a simultaneous autologous embryo transfer were excluded. As a result, the analysis included 20,244 donor oocyte cycles, of which 1,927 were gestational carrier cycles and 18,317 were intended parent recipient cycles. The oocyte donor age ranged from 18 to 50, with a mean age of 26.2 ( Table 1 ). The largest percentage of intended parent recipients were 41–45 years old (44.2%), with a mean age of 41.6. Most gestational carriers were under 35 years old (78.3%) with a mean age of 31.6. Most of the intended parents were non-Hispanic white, nonsmokers, and had performed 2.3 prior ART cycles. Gestational carrier cycles had significantly more frozen-thaw embryo transfers compared to intended parent recipient cycles (61% vs 41.9%, P <.0001). Intended parents using a gestational carrier had a diagnosis of uterine infertility and diminished ovarian reserve in 17.6% and 46% of cycles, respectively. Of note, 52% of cycles reported “other” for the infertility diagnosis. Intended parent recipients had a diagnosis of diminished ovarian reserve in 78.8% of cases. Among cycles using a gestational carrier, 31% used PGS/PGD and 74.4% had an eSET. In cases of intended parent recipients, only 12% utilized PGS/PGD and 78.5% transferred a single embryo ( Table 1 ). Among cycles using a gestational carrier, there were 1,009 deliveries, with 716 singletons and 293 sets of multiples. Among intended parent recipients, there were 7,374 deliveries, of which 5,634 were singletons and 1,740 were multiples ( Table 2 ). Overall, pregnancy outcomes were significantly better for cycles using a gestational carrier compared to intended parent recipients ( Table 2 ). Patients using a gestational carrier had significantly higher odds of clinical pregnancy (adjusted OR [aOR]1.33, 95% CI 1.17–1.51) and live birth (aOR 1.37, 95% CI 1.21–1.55), compared to intended parent recipients. The miscarriage rate was higher for intended parent recipients but not statistically significant. There were significantly more multiples among the gestational carrier cycles compared to the intended parent recipients (aOR 1.32, 95% CI 1.07–1.62). Live birth rates were consistently lower for intended parent recipients compared to gestational carriers and begin to sharply decline after age 45 ( Fig. 1 ). Among singleton deliveries, the incidence of prematurity (<37 weeks) was significantly lower in gestational carriers compared to intended parent recipients (aOR 0.78, 95% CI 0.61–0.99). Prematurity from 32 to 37 weeks was also reduced in gestational carrier cycles (aOR 0.77, 95% CI 0.59–0.99). Very preterm delivery (<32 weeks) had a similar incidence of 3.1% in both groups, which may be due to a small sample size of 22 cases in the gestational carriers and 176 in the intended parent recipients. The overall trends of low birthweight and preterm delivery are illustrated in Figure 2 . Intended parent recipients have a consistently higher percentage of low birthweight infants compared to gestational carriers, and that percentage starts to rise after age 43 ( Fig. 2A ). Similarly, preterm delivery rate starts to gradually rise after age 44 among intended parents ( Fig. 2B ). The odds of overall low birthweight among singletons was significantly reduced in gestational carrier cycles (aOR 0.62, 95% CI 0.44–0.89), relative to intended parent recipients ( Table 2 ). Low birthweight at term (1.0% vs. 1.9%) and among preterm neonates (5.4% vs. 10.2%) were also reduced in gestational carrier cycles. Neonatal outcomes were similar in fresh and frozen embryo transfer cycles ( Table 2 ). A sub-analysis was performed to compare pregnancy outcomes of 3,858 intended parent recipients with at least one prior full-term birth and at least one prior ART cycle to 1,768 gestational carriers ( Supplemental Table 1 ). The mean age of parous intended parent recipients and gestational carriers were 42.2 and 31.6, respectively. Clinical pregnancy and live birth were significantly higher in gestational carriers than in parous intended parent recipients with a history of prior ART ( Supplemental Table 1 ), whereas the miscarriage rate was lower in gestational carriers than parous intended parent recipients (aOR 0.74, 95% CI 0.55–0.99). No significant associations were observed for birth outcomes. In the sub analysis of parous intended parent recipients without a history of prior ART, the pregnancy and birth outcomes were like gestational carriers ( Supplemental Table 2 ).

The study was approved by the University Hospitals Cleveland Medical Center Institutional Review Board. Data on all gestational carrier cycles and intended parent recipient cycles in the United States utilizing donor oocytes were collected from the 2014 Society for Assisted Reproductive Technologies (SART) database. SART collects anonymous data from approximately 95% of clinics in the United States ( 18 ). Variables collected included intended parent’s demographics, smoking status, cycle type, infertility diagnosis, number of embryos transferred, elective single embryo transfer (eSET), assisted hatching, use of preimplantation genetic screening/diagnosis (PGS/PGD), intracytoplasmic sperm injection (ICSI), and treatment outcomes including pregnancy outcome, birth weight, and gestational age at delivery. We considered clinical pregnancy (CP), live birth (LB), miscarriage and plurality as pregnancy outcomes. CP was defined by the presence of a gestational sac. LB was defined as a live born neonate greater than 20 weeks gestation. We also considered preterm delivery (PTD), late PTD, very early PTD, and low birthweight (LBW) as birth outcomes. Gestational age at delivery was calculated using the date of embryo transfer. We defined PTD in 3 ways: live birth delivery at <37 weeks as PTD; ≥32 weeks and <37 weeks as late PTD; and <32 weeks gestation as very early PTD. LBW was defined as a neonate weighing less than 2500 grams. Demographic and reproductive characteristics were compared between gestational carrier cycles and intended parent recipient cycles using Fisher’s exact tests and chi-square tests for categorical variables and Student’s t-test for continuous variables, as appropriate. Infertility diagnosis and ART factors were also compared by gestational carrier cycles versus intended parent recipient cycles. Two main analyses were performed: the association between cycle type (gestational carrier vs. intended parent recipient) and pregnancy outcome (CP, LB, miscarriage, and plurality); andthe association between cycle type and birth outcomes (including PTD, late PTD, very early PTD, and LBW) among singleton live births. Each PTD classification was compared to full term deliveries and LBW was compared to normal birth weight. We also investigated associations between cycle type and LBW stratified by gestational age (i.e., among term or preterm neonates). Cycles with no embryo transferred or donor embryos (generated from infertile couples) were excluded from the analysis. Generalized linear regression models with robust error variance to account for multiple cycles for the same woman were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) to investigate these associations between cycle type (gestational carrier vs. intended parent recipient) and pregnancy and birth outcomes. Models were adjusted for carrier/recipient age, fresh or frozen embryo transfer, use of preimplantation genetic testing (PGD or PGS), assisted hatching, ICSI, number of embryos transferred. Models for birth outcomes were further adjusted for reduction in fetal hearts. To isolate the effects of infertility, we performed two sub-group analyses: (1) intended parent recipients with at least one prior live term birth and a history of at least one prior ART cycle compared to gestational carriers and (2) parous intended parent recipients without a history of ART compared to gestational carriers. The SART data do not provide any information on parity in gestational carriers; however, we assumed gestational carriers had at least one prior uncomplicated term delivery, according to the suggested ASRM criteria for selecting potential gestational carrier candidates ( 19 ). We also compared percentages of LB, LBW, and PTD from gestational carrier and intended parent recipient cycles by age group categories (i.e., <35, 36–38, 39–41, 42–44, 45–47, 48–50, and 51–53 years); the analysis was restricted to women of age ≤53 due to small number of cycles from women of age greater than 54 and no statistical comparisons were made. All data were analyzed using SAS version 9.4 (SAS Institute, Cary, North Carolina).

This is the first study to directly compare the risks of low birthweight and preterm delivery among pregnancies conceived using donor oocytes in gestational carriers and intended parent recipients. We found that gestational carriers had higher odds of live birth after transfer of donor oocyte embryos compared to women using their own uterus. The odds of having a preterm delivery or low birthweight neonate were lower in gestational carriers compared to intended parent recipients. Finally, in our sub-group analysis, we found that gestational carrier cycles demonstrated better reproductive outcomes compared to parous intended parent recipients with a history of prior ART, though no such differences were found for intended parents without a history of ART suggesting that a history of infertility may be a contributing factor in reproductive outcomes. In addition, we found similar neonatal outcomes among parous intended parent recipients with or without prior ART compared to gestational carriers. We have also demonstrated that increasing age of the intended parent is associated with lower rates of live birth and poorer neonatal outcomes ( Figs. 1 and 2 ). The mechanisms behind the associations between uterine age and adverse perinatal outcomes are still largely unknown. A recent study in mice looking at the effect of uterine age on embryo development suggests that abnormal placentation and decidualization in an older uterine environment is the cause of these negative outcomes ( 19 ). A mouse model was used to differentiate the effect of age on the oocyte and the uterus and found that uterine aging itself was associated with intrauterine growth restriction (IUGR) and congenital anomalies. Using RNA-seq on placentas at mid-gestation, they found key regulators of decidual differentiation ( Bmp2 and Igf1 ) and uterine immune cells were abnormally regulated with older age. In addition, the uterine stroma of older mice was refractory to implantation due to lower expression of progesterone and estrogen receptors in the endometrium and reduced proliferation. Even in the setting of normal embryos, neonatal low birthweight may be due to abnormal placentation resulting from abnormal decidual response to pregnancy ( 20 ). Several studies have shown that parity is an independent predictor of uterine blood flow during a subsequent pregnancy, suggesting some permanent modifications on uterine vasculature ( 21 – 23 ). The validated SART live birth calculator also found that a prior history of a live birth improves subsequent ART success ( 24 ). However, even after controlling for parity in our subgroup analysis, clinical pregnancy and live birth rates were significantly higher in gestational carriers compared to parous intended parent recipients with a history of prior infertility treatment. It is interesting to note that the neonatal outcomes were similar in parous intended parent recipients regardless of a history of previous ART. This may have to do with a lower risk of preeclampsia and other disorders of placentation in parous compared to nulliparous women ( 25 ). Strengths of our study include the large number of ART pregnancies from the most recent complete SART database in 2014. Unlike previous years, the new 2014 SART data is organized by retrieval initiated and includes use of eSET, PGD, and subsequent frozen embryo transfers, which is a more accurate representation of total cycle success. Our study included over 1,000 gestational carriers under 35 years old and 3,845 intended parent recipients over 45 years old. By restricting the analysis to donor oocyte embryos, we were able to have similar uterine preparations whether having a fresh or frozen transfer and oocyte quality across groups; this allowed us to isolate the effect of infertility on pregnancy outcomes, which is a key strength of our study. Limitations of our study include the lack of data available on gestational carriers other than age. For example, body mass index, smoking status, and prior obstetrics history would be important confounding factors that are not reported to SART. Among the intended parents’ recipients using their own uterus, 2.6% were smokers, 5.2% had endometriosis, 3.4% had PCOS and 6.5% had tubal disease. We could not control for these confounders in our models and it is possible some may be contributing to the adverse outcomes observed in the intended parent recipients. Smoking has been linked to an increased risk in preterm delivery and low birth weight ( 26 ). In addition, SART does not collect information on pregnancy complications such as preeclampsia, or other maternal complications including cardiac or cerebral venous events, maternal death, or embolic events. There is no data collected on protocols used for preparation of the endometrium for embryo transfer, whether it was a natural or medicated cycle, or the type of luteal support utilized. We had a small sample size in the very advanced maternal age category with 491 intended parent recipients and 2 gestational carriers over 50 years old. Additionally, the increased use of PGD/PGS in gestational carrier cycles may influence pregnancy outcomes but should not have affected the outcomes of preterm delivery or low birth-weight. A major barrier to utilizing a gestational carrier is cost. Future studies analyzing the cost-effectiveness of gestational carrier cycles in older intended parent recipients would be helpful, particularly if they considered the cost of extended hospital stays due to complications from maternal, perinatal, and neonatal morbidities in older mothers. Postnatal lifetime expenses should also be considered as IUGR and low birthweight have been linked to a higher risk of chronic adult diseases ( 27 ). The trend of older women carrying pregnancies is likely to grow in the future, as more women freeze their eggs and may serve as their own “egg donor” later in their reproductive years. Future studies should expand on this study by using several years of data to increase the sample size and focus on pregnancy outcomes of women in their fifth and sixth decades receiving donor and autologous oocytes. More animal and human research studies are needed to explore the molecular mechanisms underlying the effect of infertility on adverse perinatal outcomes. Finally, counseling patients on the potential effects of age on pregnancy outcomes, and emphasizing the importance of eSET in this context, is vital for ensuring the safest outcomes for older fertility patients.