Prevalence of a Good Perinatal Outcome With Cryopreserved Compared With Fresh Donor Oocytes.

This retrospective cohort study was declared exempt by the Duke Institutional Review Board as the data were de-identified. All 2012-2015 donor oocyte IVF cycles with intent for fresh embryo transfer, no use of a gestational carrier, and no preimplantation genetic testing (PGT) were queried from the Society for Assisted Reproductive Technology Clinic Outcomes Reporting System (SART CORS). The SART CORS database contains comprehensive data from >90% of all clinics performing ART cycles in the United States. The data were collected through voluntary submission, verified by SART, and then 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). SART maintains HIPAA-compliant business associates agreements with reporting clinics. In 2004, following a contract change with CDC, SART gained access to the SART CORS data system for the purposes of conducting research. The data in the SART CORS are validated annually with select clinics having on-site visits for chart review based on an algorithm for clinic selection. 20 During each visit, data reported by the clinic were verified with information recorded in patients’ charts. 20 In 2012, records for 2,045 cycles at 35 clinics were randomly selected for full validation, along with 238 egg or embryo banking cycles. The full validation included review of 1,318 cycles for which a pregnancy was reported. Among the non-donor cycles, 331 were multiple-fetus pregnancies. Ten out of 11 data fields selected for validation were found to have discrepancy rates of ≤5%. The exception was the diagnosis field, which, depending on the diagnosis, had a discrepancy rate between 2.1% and 9.2%. 20 A cycle was defined by the initiation of medications to prepare the recipient’s uterus for embryo transfer. Cycle cancellation was defined as initiation of endometrial preparation for the recipient without subsequent embryo transfer. The stage at embryo transfer was defined as cleavage or blastocyst stage based on morphologic assessment and the day of transfer. Elective single embryo transfer (eSET) was defined as single embryo transfer with at least one excess embryo cryopreserved. A good quality embryo was defined by SART CORS as an embryo free of or with only minor imperfection. 21 The primary outcome was a “good perinatal outcome,” defined as a singleton live birth at ≥ 37 weeks with birth weight ≥ 2,500 g and ≤ 4,000 g. Secondary outcomes included live birth rate, defined as the proportion of cycles with a live birth entered in SART CORS, multiple birth rate, calculated as the proportion of live births with two or more infants, and prematurity rate, defined as the proportion of live births occurring at less than 37 weeks gestational age. Additionally, high-order multiple birth was defined as the birth of three or more infants. Low birth weight was defined as birth weight less than 2,500 grams. Clinical pregnancy rate was defined as the proportion of cycles with a gestational sac on first-trimester ultrasound. Implantation rate was defined by SART CORS as the greater of the number of fetal hearts on ultrasound or the number of live births plus still births, divided by the total number of embryos transferred. All analyses were conducted in R version 3.5.3 (Vienna, Austria). 22 After cycles with missing parity and oocyte data were excluded, cycles with missing body mass index (BMI) were imputed using the mean observed BMI. Generalized estimating equation (GEE) models with an unstructured variance-covariance matrix were used to account for the correlation induced by repeated cycles within a patient. A multivariable GEE model with a binomial distribution and logit link was then implemented to test the effect of oocyte type (cryopreserved or fresh) on good perinatal outcome with and without covariate adjustments. The covariates included in the GEE model were the recipient’s age, BMI, smoking status, parity, race, infertility diagnosis, number of available oocytes, ICSI, use of assisted hatching, clinic region, and year. Race was categorized as “unknown” for cycles with “patient not asked,” “patient refused,” or “unknown” selected for that field. Additionally, the stage of embryo development and the number of embryos transferred were combined into one covariate to avoid estimation errors caused by correlation of these variables. For cycles in which no oocytes were retrieved or thawed, the number of available oocytes was set to zero. Similarly, for cycles that did not proceed to transfer, the number of embryos transferred was set to zero. Cycle characteristics and secondary outcomes were analyzed without covariate adjustments using GEE models with binomial distribution (logit link) for binary dependent variables and with Gaussian distribution (identity link) for continuous dependent variables. Next, a sensitivity analysis of independent cycles was performed by using a multivariable logistic regression model with data from the first cycle per woman and the same set of covariate adjustments. Finally, we performed a second sensitivity analysis based on complete cases in which at least one embryo was transferred. Odds ratios (ORs) and 95% confidence intervals (CI) were obtained for all variables. A P- value of less than .05 was considered statistically significant.

A total of 37,402 cycles were queried. After exclusion of 190 cycles (0.5%) with missing data for good perinatal outcome and 287 cycles (0.8%) with missing data for parity or the number of available oocytes, the analysis cohort consisted of 36,925 unique cycles from 31,124 women ( Figure 1 ). Among these women, 26,580 underwent one IVF cycle while 4,544 underwent two or more IVF cycles ( Table 1 ). Of the 36,925 cycles included in the analysis, 8,381 (22.7%) were performed with cryopreserved oocytes and 28,544 (77.3%) with fresh oocytes ( Table 2 ). The mean recipient age was 41 years, and the mean donor age 26 years, in both groups. The vast majority of donors were anonymous. Data on recipient BMI were missing for 21.6% of cycles; among those with BMI data, the mean BMI was clinically similar between groups (25.9 kg/m 2 compared with 25.6 kg/m 2 ). Diminished ovarian reserve was the most common infertility diagnosis in both groups. As expected, cryopreserved oocytes were used in an increasing proportion of cycles during the 4-year study period. The median numbers of available oocytes and fertilized oocytes were significantly lower in cryopreserved oocyte cycles than in fresh oocyte cycles ( Table 3 ). Intracytoplasmic sperm injection (ICSI) and assisted hatching were used more commonly in cryopreserved cycles, while blastocyst-stage embryo transfer was performed in a larger proportion of fresh oocyte cycles. Patients who utilized cryopreserved oocytes were significantly more likely to have one embryo transferred (42.3% compared with 32.9%, OR 1.46, 95% CI 1.39-1.54, P <.001). At least one good-quality embryo was transferred in a significantly smaller proportion of embryo transfers from cryopreserved oocytes than from fresh oocytes (68.1% compared with 80.6%, OR 0.51, 95% CI 0.48-0.55, P <.001), and the median number of excess embryos cryopreserved was also significantly smaller (0 compared with 3, β = −3, 95% CI −3.1 to −3, P <.001). The proportion of cycles resulting in a good perinatal outcome was 22.0% in the cryopreserved oocyte group and 24.1% in the fresh oocyte group (OR 0.90, 95% CI 0.85-0.96, P= .001, Table 4 ). The multivariable GEE model demonstrated significantly lower odds of a good perinatal outcome with cryopreserved oocytes (adjusted OR 0.88, 95% CI: 0.81-0.95, P =.002, Table 4 ). Implantation, clinical pregnancy, and live birth rates were all significantly lower with cryopreserved oocytes than fresh oocytes ( Table 5 ). In contrast, rates of multiple birth (22.3% compared with 31.2%, OR 0.63, 95% CI 0.58-0.69, P <.001), prematurity (27.6% compared with 30.6%, OR 0.86, 95% CI 0.79-0.94, P =.001), and low birth weight (25.3% compared with 29.8%, OR 0.80, 95% CI 0.73-0.87, P <.001) were also significantly lower in the cryopreserved oocyte group ( Table 5 ). Sensitivity analysis using the first cycle per woman was performed to determine if cycle order might have biased the observed association between oocyte type (cryopreserved versus fresh) and the primary outcome ( Table 4 ). The odds of a good perinatal outcome remained significantly lower with cryopreserved oocytes (adjusted OR=0.89, 95% CI 0.81-0.97, P= .01). Given that a higher starting number of oocytes may influence the likelihood of a cycle proceeding to embryo transfer, a sensitivity analysis was performed of 31,982 cycles from 28,500 women in which at least one embryo was transferred ( Table 4 ). The odds of a good perinatal outcome remained significantly lower with cryopreserved oocytes (adjusted OR=0.88, 95% CI 0.81-0.95, P =.002).

In this national study of 36,925 donor oocyte IVF cycles from 2012-2015, the use of cryopreserved compared with fresh donor oocytes in IVF cycles was associated with marginally lower odds of a good perinatal outcome after adjusting for confounders including the number of available oocytes. The findings were similar in a sensitivity analysis of cycles that proceeded to embryo transfer. Two initial studies on this topic used composite clinic data, preventing any adjustment for covariates. 15 , 16 Furthermore, there were no data on the number of available oocytes. In a fresh cycle, the recipient typically obtains the total number of retrieved oocytes. Cryopreserved oocytes, however, are commonly purchased in batches of 6-8 eggs. Logically, a smaller starting number of oocytes will yield fewer fertilized oocytes and fewer cleavage-stage embryos. In turn, clinics may be more likely to perform embryo transfer at cleavage stage rather than extend culture to blastocyst stage, to avoid the risk of having no embryos to transfer. Ultimately, a lower live birth rate could be solely due to the initial difference in oocyte number. Another recent study from the CDC examined 2,223 cryopreserved and 9,691 fresh donor oocyte IVF cycles from 2013. 10 Consistent with our findings, the live birth rate was lower with cryopreserved than fresh donor oocytes. When the analysis was restricted to cycles with at least one embryo transferred, however, there was no significant diffference in live birth rates. The authors attribute this to a difference in cancellation rates. Because the cancellation rate was lower in the cryopreserved oocyte group, however, removal of the cancelled cyles would exacerbate any difference in live birth rates rather than negate it. A more likely explanation is that the authors adjusted for the number of embryos transferred in the per-transfer analysis, but not the per-cycle analysis. They acknowledged this limitation and explained that it was considered a missing value for cancelled cycles. The importance of this limitation was understated, as they also did not control for the number of available oocytes. In the present study, the number of transferred embryos was set to zero for cancelled cycles and included as a covariate. Addtionally, the CDC authors did not have data on smoking status, BMI, or embryo quality, all of which were adjusted for in our analyses. The present study has many strengths. Unlike previous studies examining only one year of data, our analysis included data from three years. Because the SART CORS database includes data from the vast majority of clinics performing ART in the United States, external validity is another strength of this study. The analysis adjusted for several covariates, including ICSI, assisted hatching, and a combined covariate including the stage of embryo development and number of embryos transferred. Additionally, we demonstrated that differences in outcomes persist even after adjusting for the starting number of oocytes. These findings support the notion that the process of cryopreserving and thawing oocytes may have a modest detrimental impact on the potential of those ooccytes to result in a live birth. Although our analysis does not allow us to draw conclusions about the underlying cause, we demonstrated a significantly lower proportion of high-quality embryos among the cryopreserved oocyte group. A reduction in embryo quality could be due to meiotic spindle damage, as has been shown previously. 23 Another strength of the present study is its focus on a good perinatal outcome, rather than live birth. Whereas previous studies did not report rates of multiple birth or prematurity, a good perinatal outcome has been advocated as the most relevant parameter for couples seeking fertility care. 24 Exclusive reliance on live birth rate underestimates the impact of multiple birth on perinatal morbidity and mortality 18 , 25 . In 2015, ART accounted for only 1.7% of all births in the United States, but 5.3% of preterm and 5.4% of very preterm births. 26 The majority of preterm births from ART result directly from the transfer of multiple embryos. 26 Elective single embryo transfer effectively reduces the multiple birth rate with minimal impact on the rate of live birth in young patients. 27 For that reason, national guidelines state that only one embryo should be transferred in donor oocyte IVF cycles. 28 Our study revealed that these guidelines were not followed for the majority of cycles, leading to high rates of multiple birth and prematurity in both groups. Multiple birth and prematurity rates were even higher with fresh oocytes than with cryopreserved oocytes, due to a higher proportion of cycles with two or more embryos transferred (67.1% of fresh oocyte cycles compared with 57.8% of cryopreserved oocyte cycles). Elective single embryo transfer rates remained comparable between groups, reflecting the reduced number of excess embryos for cryopreservation among women who used cryopreserved oocytes. Ultimately, less than 25% of cycles led to the birth of a term, normal-weight infant. These statistics highlight the importance of single embryo transfer among all patients undergoing donor oocyte IVF. The present study also has limitations. For example, unknown race was reported for 45% of cycles, consistent with existing SART literature. 29 We opted to include “unknown” as a category for race, since it is a choice for clinics to select when inputting data and is listed as such in the dataset generated by SART. From a statistical standpoint, we felt that it was better to proceed in this fashion rather than exclude 45% of patients from the analysis; however, we acknowledge this limitation. A third limitation is lack of data on oocyte maturity. The number of available fresh oocytes reflects the total number retrieved, but a proportion of retrieved oocytes are immature. Therefore, we were unable to compare fertilization rates between groups. Another limitation was the inability to evaluate cumulative live birth rates with all fresh and frozen embryos derived from the initial cohort of oocytes. Because patients using fresh oocytes were more likely to have excess cryopreserved embryos, cumulative live birth rates may be even higher for that cohort. In conclusion, this retrospective national study demonstrated that the use of cryopreserved compared with fresh donor oocytes in IVF cycles is associated with marginally lower odds of a good perinatal outcome. The clinical significance of this finding is relatively modest in light of the many known benefits of cryopreserved donor oocytes. Patients should be encouraged to consider the advantages and disadvantages of both approaches when planning for donor oocyte IVF.

In vitro fertilization (IVF) with donor oocytes is standard treatment for diminished ovarian reserve or advanced reproductive age. Between 2000 and 2010, the annual number of donor oocyte cycles in the United States almost doubled. 1 Given the increased demand for donor oocytes, cryopreserved oocytes are an attractive option. Benefits include decreased cost, less wait time, and elimination of hormonal synchronization for the donor and recipient. Embryo development is not compromised by oocyte cryopreservation, but data on live birth and perinatal outcomes are sparse. 2 - 8 Initial studies were limited by small sample size, no control group, or inability to adjust for confounders due to the use of aggregate, rather than patient-level, data. 9 - 16 A national study demonstrated a reduction in live birth rate with cryopreserved oocytes, although this reduction was eliminated after restricting the analysis to cycles that proceeded to embryo transfer. 17 That study was limited by the inclusion of only one year, lack of information on multiple birth or prematurity, and inability to adjust for the number of available oocytes. Their findings could be attributed to the fact that cryopreserved oocytes are obtained from commercial banks in small batches of only six to eight eggs, compared to fifteen or more fresh oocytes from a donor retrieval. We aimed to compare outcomes between fresh and cryopreserved oocytes while adjusting for covariates, including the number of available oocytes. Our primary outcome was chosen to highlight the importance of a healthy birth, or a “good perinatal outcome” defined as a singleton full-term neonate with a normal birth weight. 18 , 19