The impact of single-step and sequential embryo culture systems on obstetric and perinatal outcomes in singleton pregnancies: the Massachusetts Outcomes Study of Assisted Reproductive Technology.

Out of a total 13, 690 pregnancies at MGH and BWH from 2004–2017 linked to SART CORS with laboratory data, 9032 records were in-state deliveries linked to PELL. Of these deliveries, 7371 were conceived with autologous fresh embryo transfer and 5403 were singletons. After restricting to blastocyst transfers, there were 1532 deliveries included in our analysis; 1058 pregnancies were conceived following culture in single step media and 474 were conceived following culture in sequential media. Demographic characteristics are shown in Table 1 . The culture media groups were similar in age distribution (single step culture: 45.4% under age 35 vs. 54.6% 35 years and over; sequential culture: 47.3% under age 35 and 52.7% 35 years and older, p = 0.493) and in race distribution (single step culture: 75.6% non-Hispanic white and 24.4% other; sequential culture: 81.4% non-Hispanic white and 18.6% other, p = 0.013). Most women in both groups had a college education and had private health insurance. There were more cycles in our cohort performed at MGH than BWH (59% vs. 41% overall; p < 0.0001). Chronic hypertension was present in 2.1% of the single step media group, which was similar to the sequential media group (p = 0.969). The single step culture group was characterized by more deliveries from women with diminished ovarian reserve (18.5% vs. 12.9%, p = 0.006) and polycystic ovarian syndrome (PCOS) or ovulation disorders (22.8% vs. 15.6%, p = 0.001). Most deliveries arising from the single step culture group resulted from embryos cultured in 5% oxygen, while most deliveries in the sequential culture group resulted from embryos cultured in 20% oxygen (94.4% vs. 74.1%, respectively, p < 0.0001, Table 2 ). Similarly, protein supplementation content differed between the groups, with 49.5% of single step culture deliveries with embryos receiving SSS compared to 86.3% of sequential culture deliveries (p < 0.0001). A higher proportion of single step culture deliveries were conceived with ICSI (57% of single step culture vs. 48.1% of sequential culture, p = 0.001). Single embryo transfers, which became more common over time, were performed in 69.1% of single step culture pregnancies, compared to 40% of sequential culture pregnancies (p < 0.0001). The proportion of deliveries complicated by cesarean section, placental abnormalities, pregnancy-induced hypertension, preterm birth, low birthweight, and SGA were not significantly different in either the raw data or adjusted model ( Table 3 ). There was a higher proportion of gestational diabetes in the single step culture group (7.8% vs. 3.8%, p = 0.004) prior to adjusting for confounders but no difference in odds of gestational diabetes in our adjusted model (aOR 1.35, 95% CI 0.64 – 2.85, Table 3 ). Interestingly, there was a significantly increased odds of delivering an LGA infant in the single step culture group that persisted in our adjusted model (7.1% vs. 5.9% in the sequential media group; aOR 2.1, 95% CI 1.04 – 4.22, Table 3 ). Similar findings were observed when additionally controlling for PCOS diagnosis and when each fertility laboratory was examined separately ( Supplemental Tables 1 and 2 , respectively).

This historical cohort study included singleton live births arising from autologous fresh transfers at either Massachusetts General Hospital Fertility Center (MGH) or Brigham and Women’s Hospital Center for Infertility and Reproductive Surgery (BWH). All transfers occurred on day 5 with the exception of one day 6 transfer. Deliveries took place between July 1,2004 and December 31,2017 and were linked to the MOSART database using maternal, paternal, and infant names and dates of birth. IVF cycle identification numbers from MGH and BWH were then added to the MOSART data set and used at each site to abstract laboratory data. Frozen embryo transfers, out of state deliveries, and cycles with incomplete embryology data were excluded. The MOSART database links ART data from the Society for Assisted Reproductive Technology Clinic Outcome Reporting System (SART CORS) database to birth certificate and hospital discharge data recorded in the Massachusetts Pregnancy to Early Life Longitudinal (PELL) system.( 12 ) The linkage is based on maternal, paternal, and infant identifying information such as name and date of birth.( 12 ) SART CORS is a clinical cycle-based ART database containing demographic, cycle specific, treatment outcome and pregnancy outcome data. The SART CORS database contains cycle-based data entered by US ART clinics and reported to the Centers for Disease Control and Prevention in compliance with the Fertility Clinic Success Rate and Certification Act of 1992 (Public Law 102–493). The database contains ART cycle specific demographics, infertility diagnoses, ART treatment, pregnancy, and outcome data and is maintained by Redshift Technologies, Inc. under contract to SART. Data are obtained from approximately 90% of ART clinics in the United States and all Massachusetts clinics are included in the database. SART CORS data are validated annually with random clinics having on-site visits for chart review based on an algorithm for clinic selection. During these visits, data reported by the clinic are compared with information recorded in patients’ charts. In 2017, of data fields selected for validation, only cycle start date and several infertility diagnoses were found to have discrepancy rates >5%.( 13 ). The PELL system links vital records from birth certificates to hospital discharges, outpatient visits, and emergency room visits in Massachusetts since 1998. PELL was established through a collaboration among Boston University School of Public Health, the Massachusetts Department of Public Health (MDPH), and the Centers for Disease Control and Prevention (CDC). The MOSART database was established as a linkage of SART CORS and PELL data in Massachusetts under a Memorandum of Understanding between SART, the MDPH, and the project principal investigators. It was approved by the SART Research Committee.( 12 ) This study was approved by the Institutional Review Boards of Mass General Brigham, the Massachusetts Department of Public Health, and Dartmouth-Hitchcock Health. Embryos were cultured individually to the blastocyst stage prior to fresh transfer. Single step medium in this study includes Continuous Single Culture (CSC+), (FujiFilm, Irvine, CA) and Global (GLB)/Global Total (GLB-T) (CooperSurgical Inc, Trumbull, CT); and sequential media includes Quinn’s Advantage Cleavage and Blastocyst Media (CooperSurgical Inc, Trumbull, CT) and G-Series media (G1/G2) (Vitrolife, Sweden). Both MGH and BWH off-gas all plasticware before use and prepare media 18–24 hours before use to ensure proper equilibration before introducing gametes or embryos. Benchtop and box incubators were used at both fertility centers. The types of oil used varied with year and fertility center and included Light Mineral Oil (Irvine Scientific, Santa Ana, CA, USA), OVOIL (Vitrolife, Göteborg, Sweden), and LiteOil (Coopersurgical, Trumbull, CT, USA). Maternal demographic variables including age, race, education level, parity, delivery payer source, and year of delivery were obtained from birth certificate data in the PELL system. Chronic hypertension as a preexisting comorbidity was determined from birth certificates or hospital discharge data (ICD-9 codes 401, 402, 403, 404, or 405). Infertility diagnoses were obtained from the BWH and MGH laboratory cycle data. Variables with an n of less than 11 are required to be suppressed by the Massachusetts Department of public health. Live births were linked via SART ID to embryology laboratory data at the respective institutions. Embryology variables included oxygen tension (5% or 20%), protein content (human serum albumin [HSA], or serum substitute supplement [SSS]), use of intracytoplasmic sperm injection (ICSI), use of assisted hatching, and number of embryos transferred. The primary outcomes of interest were obstetric—method of delivery, placental abnormalities, pregnancy hypertension, and gestational diabetes—and perinatal—prematurity, low birthweight, small-for-gestational age (SGA), and large-for-gestational age (LGA). Method of delivery was determined from the birth certificate data. Data on placental abnormalities (defined clinically as placental abruption (ICD9: 641.2, 762.1/ICD10: O45), placenta previa (ICD9: 641.0, 641.1/ICD10: O44), vasa previa (ICD9: 663.5/ICD10: O69.4), retained placentas without hemorrhage (ICD9: 667.0 / ICD10: O73.0), and/or placenta accreta (ICD10: O43.21)) were obtained from hospital discharge delivery as well birth certificate records in PELL. Pregnancy-induced hypertension and gestational diabetes were identified in PELL from either the birth certificate or hospital discharge delivery records (pregnancy-induced hypertension, ICD9: 642/ICD10: O13, O14, O15, O16; gestational diabetes, ICD9: 648.83 / ICD10: O24.4). Length of gestation was calculated using the SART CORS outcome date minus the date of fertilization plus 14 days. Preterm births were defined as a length of gestational less than 37 weeks. Low birthweight was defined as weight under 2500g. Based on birthweight, Z scores were calculated using generated gender-, race/ethnicity-, and gestation-specific birthweight means and SD using all live births from 1998 through 2017 in Massachusetts as a reference, as previously described.( 14 , 15 ) Infants with birth weights less than the 10th percentile for gestation and gender were classified as small-for-gestational age (SGA), and those greater than the 90 th percentile for gestation and gender were classified as large-for-gestational age (LGA). We compared frequencies of maternal and birth outcomes of infants born to mothers for whom single step or sequential media were used to culture embryos to the blastocyst stage, using Chi-square statistics (alpha = 0.05). Multivariate logistic regressions were used to assess the association between single step and sequential media type among fresh singleton deliveries and adverse health outcomes, adjusting for maternal age at delivery (<35, 35+ years), race/ethnicity (non-Hispanic white, all other), maximum education level (high school, some college, college or graduate/professional school), insurance status at birth (public, private), parity (1, 2+), oxygen tension (5% or 20%), protein supplementation (HSA or SSS), fertilization method, and number of embryos transferred (1, 2+). Crude and adjusted odds ratios (aOR) and 95% confidence intervals (CI) are reported. All analyses were performed using the SAS software, version 9.4 (SAS Institute, Cary, NC).

Embryo culture systems play an important role in optimizing preimplantation embryo development. Ultimately, it is most important to understand how laboratory conditions during in vitro fertilization affect maternal and fetal health. Our study demonstrates that single step culture is associated with an increased risk of LGA in singleton pregnancies conceived with fresh embryo transfer, suggesting that single step culture is one laboratory factor that may have a lasting impact on fetal growth and directly influence clinical outcomes.

As the number of children conceived with ART increases, continued scrutiny of laboratory variables that may influence obstetric and perinatal outcomes is of paramount importance to the field of reproduction. In addition, the number of cycles utilizing day 5–7 blastocyst transfer continues to rise, resulting in more embryos cultured for a longer duration, and so optimizing embryo culture systems to improve obstetric and perinatal outcomes is crucial. Prior studies have shown that single step culture media systems may offer some advantage over sequential media culture systems in terms of early outcomes such as blastocyst development and usable blastocyst rate, but they appear similar in terms of pregnancy rates.( 6 , 7 , 10 , 11 , 16 – 18 ) However, information regarding obstetric and perinatal outcomes after single step versus sequential culture media systems is lacking. If single step versus sequential culture can affect early embryo development, then there may be long term consequences of these different culture systems. Contrary to our original hypothesis, our findings indicate that single step media culture is associated with increased odds of LGA (aOR 2.1), even after controlling for key demographic and laboratory confounding variables. LGA is associated with increased maternal and neonatal morbidity,( 19 ) and so our observation may be a sign of other potential neonatal risks related to embryo culture that were not assessed with our outcome variables. Although GDM is a risk factor for delivering an LGA infant, the incidence of GDM did not differ between our study groups in our adjusted models, which took into account maternal age and ethnicity, two risk factors for GDM.( 20 ) Furthermore, since women with PCOS are at higher risk of GDM and LGA, particularly in the setting of elevated BMI, an additional analysis controlling for PCOS diagnosis was performed, showing no changes in our primary outcomes. The mechanism underlying the link between single step culture and LGA in our cohort is not yet clear. Culture media provides nourishment as the inner cell mass and trophectoderm differentiate. Animal data illustrate that signaling between the embryo and its environment from the earliest stages has a lasting effect on cellular metabolism and other important functions.( 21 ) Kleijkers et al. illustrated that different culture media lead to different levels of embryo gene expression in pathways related to apoptosis, protein degradation, metabolism, and cell-cycle regulation.( 22 ) In fact, individual commercial embryo culture media are remarkably variable in terms of overall composition as well as type of protein supplementation.( 23 – 26 ) A few prior studies have examined the association between particular commercial medias and fetal growth. For example, Dumoulin et al. demonstrated that amongst embryos randomly allocated to Vitrolife or Cook sequential culture media and transferred on day 2–3, differences in birthweight were apparent,( 27 ) with increased fetal growth noted in the Vitrolife group by 20 weeks gestation.( 28 ) Vitrolife G1-PLUS has also been implicated in an increased rate of LGA infants from blastocyst transfers.( 29 , 30 ) One hypothesis described by Sunde et al. proposes that if early embryos in particular lack the nutrients they require from culture media, they may be reprogrammed to have increased growth once placed in the nutrient-rich maternal environment.( 31 ) Another possibility is that different culture medias influence epigenetic regulation during early embryo and placental development.( 32 ) Although we controlled for both protein supplementation and oxygen concentration, two key laboratory variables that have been shown to play an important role in blastocyst development,( 23 , 24 , 33 , 34 ) there may be other ingredients in the commercial medias used in this study that affect embryo development and fetal growth. Of note, Vitrolife G-Series medium was used in the sequential media group and so cannot be linked to a greater risk of LGA in our cohort. Prior studies have also proposed that single step culture reduces embryo stress due to decreased handling and fewer environmental fluctuations.( 17 , 35 ) It is therefore possible that a combination of different nutrient composition and utilization, as well as reduced stress, in the single step culture systems lead to an altered fetal growth trajectory. The study is limited by its retrospective design. The study population is skewed towards educated, non-Hispanic white women with private insurance, and so our findings may not be as generalizable to other populations. The primary outcomes assessed were those that were available in the MOSART database, and so other indicators of obstetric and perinatal morbidity were not available. Body mass index (BMI) data was not available for the patients, which would have been helpful to assess as a confounding variable since elevated BMI is a risk factor for LGA. Only those deliveries with complete embryology data were eligible for inclusion, although this was unlikely to have introduced selection bias. As with any study assessing outcomes from a wide date range, controlling for the most relevant laboratory variables for our desired outcomes may not have fully captured every laboratory change over time affecting obstetric and perinatal outcomes. Oxygen tension was the most significant laboratory covariate that changed over time, and so further studies examining the impact of oxygen tension on perinatal outcomes in a larger cohort are needed. We controlled for fertilization method and number of embryos transferred, as the use of ICSI and increased numbers of embryos transferred have both been associated with adverse outcomes.( 36 , 37 ) Because embryo transfers and deliveries were performed at different academic centers, there may be other laboratory variables that could affect our primary outcomes, such as type of oil used, which deserve further investigation in future studies. However, we also assessed the rates of our primary outcomes by fertility center and found no changes in the direction of adjusted odds ratios, though this analysis was limited by lower power. In terms of our analysis, we were unable to correct for multiple deliveries per patient (52 women, 3.5% of the total cohort, had multiple deliveries during the study period) due to a lack of convergence of the models. Finally, we did not control for infertility diagnosis for parsimony. There are several strengths of this study. It is the first to analyze singleton deliveries from a well-validated database that links SART CORS ART data to maternal and fetal outcomes via birth certificate and hospital discharge data in the PELL system.( 12 ) Detailed embryology data was available for each delivery, enabling us to control for key confounding variables, including protein supplementation and oxygen concentration. We were also able to control for the number of embryos transferred as well as the use of ICSI, which have both been suggested previously as risk factors for adverse obstetric outcomes.( 36 , 38 ) We excluded frozen cycles, as cryopreservation introduces another variable which has been associated with distinct placental and obstetric outcomes compared to fresh transfers.( 39 , 40 ) However, frozen embryo transfers now account for almost 75% of transfers in the United States( 41 ) and have also been linked to an increased risk of LGA( 42 ); embryo culture conditions that may contribute to this outcome are important to explore in future studies. Ideally, our findings should be investigated further with a randomized, prospective study design at a single center where all other laboratory variables can be controlled.

The evolution of IVF culture media, a fundamental aspect of embryo development in vitro, has played an important role in improving assisted reproductive technology (ART) pregnancy rates over time. After early observations of the physiologic variation in substances such as glucose, lactate, and pyruvate between the fallopian tube and the uterus,( 1 ) two distinct culture system models emerged: sequential, in which the embryo is placed in media that reflects the in vivo environment appropriate for its developmental stage, and single step, in which the media contains the necessary components for the embryo to “choose” components as needed.( 2 , 3 ) Despite the early improvement in pregnancy rates, disparate obstetric and perinatal outcomes between singletons conceived with and without ART persist,( 4 ) due to either components of fertility treatment such as embryo culture, underlying infertility diagnosis, or perhaps a combination of both. A study performed in 2002 comparing “stage-specific” sequential media to the commonly used human tubal fluid (HTF) medium formulation at that time found that the improved sequential media was associated with better embryo morphology and usable embryo rate.( 5 ) However, more recent studies comparing single step versus sequential culture systems with newer commercial media have identified no significant differences in pregnancy or miscarriage rates.( 6 – 9 ). One study suggests that single step media (Global) is associated with improved cleavage stage embryo quality and implantation rates compared to sequential media (ECM/Multiblast).( 10 ) Another prospective study that compared sibling oocytes randomly cultured in either single step (Global) or sequential media (ISM1/BlastAssist) found a higher proportion of high quality blastocysts in the single step medium group.( 11 ) There is inadequate evidence at this point to recommend one of these culture systems.( 6 ) Importantly, no studies to our knowledge have yet assessed obstetric and perinatal outcomes in singleton pregnancies conceived with blastocyst transfer cultured in single step versus sequential media. The Massachusetts Outcome Study of Assisted Reproductive Technology (MOSART) database links records from birth certificates, hospital discharges, and program data from child health and developmental programs in Massachusetts, affording a unique opportunity to evaluate laboratory conditions on obstetric and perinatal outcomes in an infertility insurance-mandated state. In this study, we compared obstetric and perinatal outcomes of singleton deliveries conceived with embryos cultured in either single step or sequential media. We hypothesized that there would be no differences in outcomes based on culture system.

Supplemental Table 1. Obstetric and perinatal outcomes by type of culture media system among fresh cycles with blastocyst transfer that link to MOSART singleton deliveries, additionally adjusting for PCOS diagnosis Supplemental Table 2. Obstetric and perinatal outcomes by type of culture media system among fresh cycles with blastocyst transfer that link to MOSART singleton deliveries, by fertility clinic site