The impact of using donor sperm in assisted reproductive technology cycles on perinatal outcomes.

A total of 134,710 fresh autologous ART cycles were evaluated from the SART database from 2012 to 2013, of which 7,043 used donor sperm and 127,667 used partner sperm. After exclusion criteria, 2,186 donor sperm and 43,875 partner sperm cycles that resulted in singleton clinical pregnancies were further analyzed. To eliminate the possibility of induced immune tolerance with prior donor sperm exposure, we further restricted the final analysis to the first ART cycles, which included 2,123 donor sperm and 42,799 partner sperm ART cycles. The demographic and treatment characteristics of these cycles are described in Table 1 . There were significant differences between the donor and partner sperm groups in most categories, thus necessitating the linear or logistical regression method that we chose to adjust for these variables in the analyses. After adjusting for all potential confounding variables (including maternal age, race, body mass index [BMI], smoking status, gravidity, history of preterm birth, highest follicle-stimulating hormone [FSH] count, blastocyst transfer percentage, total embryo transferred, and etiology of infertility), no significant differences in miscarriage rates, preterm births, very preterm births, LBW, and VLBW were observed ( Table 2 ). Birthweight was significantly lower in the partner sperm group than in the donor sperm group (3,292 ± 601 and 3,233 ± 592 g in the donor and partner sperm groups, respectively, adjusted P =.003); however, this small absolute difference (adjusted effect estimate 42 g, 95% CI 14.7–70.9) does not carry clinical significance ( Table 2 ).

After receiving approvals from the local institutional review board and SART Research Committee, a population-based historical cohort study of ART cycles from the SART CORS database was conducted. All fresh autologous ART cycles from 2012 to 2013 in the SART CORS database were evaluated for perinatal outcomes. Only those cycles that resulted in clinical pregnancies were included in the initial analysis. Exclusion criteria included visualization of more than one fetal heart on ultrasound, donor oocyte and donor embryo cycles, frozen/thawed embryo transfer cycles, gestational carriers, ectopic and heterotopic pregnancies, and cycles resulting in no pregnancies or in biochemical pregnancies. To eliminate the possibility of induced maternal immune tolerance with prior donor sperm exposure, we further restricted the final analysis to the first ART cycle. The primary outcomes evaluated include gestational age, birthweight, and miscarriage rate. Miscarriage was classified as presence of gestational sac on ultrasound without a live birth. Live birth was defined as birth after 24-week gestational age (GA). Preterm birth was defined as live birth <37 weeks, and very preterm birth was defined as live birth <34 weeks. Low birthweight (LBW) was defined as birthweight <2,500 g, and very low birthweight (VLBW) was birthweight <1,500 g. Baseline patient characteristics are summarized as mean (with standard deviation [SD]) or median (with interquartile range [IQR]) for continuous variables and frequency (%) for categorical or nominal variables. Baseline continuous variables were compared between donor sperm and partner sperm groups using Student’s t -test or Wilcoxon’s rank sum test as appropriate. Categorical or nominal variables between donor sperm and partner sperm groups were contrasted using the χ 2 test. Linear or logistic regression was used depending on the distributional form of outcome variables considered in comparing the donor sperm and partner sperm groups. Analysis was adjusted for a priori identified potential confounding variables including maternal age, gravidity, body mass index, prior full-term births, prior preterm births and prior miscarriages, etiology of infertility, assisted hatching, percentage of blastocyst transfer cycles, and total embryos transferred. In the models used, the shape of the association between maternal age and study outcomes while adjusting for other potential confounding factors is determined by using the natural cubic spline with 5 knots based on its empirical distribution. Missing data were handled by using multiple imputation. SAS version 9.4 (2002 – 2012 by SAS Institute Inc., Cary, NC) was used for statistical analysis. Results are presented as point estimates and corresponding 95% confidence interval (CI). A P value ≤.05 was considered statistically significant.

After controlling for various potential confounding variables, we found no significant differences in adverse pregnancy outcomes including miscarriage, preterm births, or low birthweight in the first fresh autologous ART cycles conceived with donor sperm vs. partner sperm. These findings were contrary to our hypothesis, which was that adverse pregnancy outcomes related to placental insufficiency are increased in pregnancies conceived through donor sperm as compared to those using partner sperm because of higher incidences of maternal immune intolerance. Our hypothesis was based on the lack of prior exposure to paternal antigens in donor sperm ART cycles, leading to potential inadequate maternal immune adaptation and poor trophoblast and placenta formation. Our findings could indicate that the exposure to paternal antigens during ART cycles may be in such a low dosage that the impact on the maternal immune maladaptation may not be as significant as we had expected. Without the exposure to donor semen or seminal fluid during an ART cycle, the maternal immune response to foreign paternal antigens expressed by the embryo appears to be minimal. The strengths of our study include using a large database extracted from a national registry, and carefully controlling for various potential confounding variables. By using fresh autologous cycles and strict exclusion criteria, we started the analysis with the first ART cycles that were conceived with fresh embryo transfers and resulted in singleton clinical pregnancies. By restricting the analyses to the first ART cycles only, we minimized the possibility of induced immune tolerance with prior donor sperm exposure. By including fresh autologous ART cycles only, we eliminated the other possible foreign antigen exposures as occurs in donor oocyte or donor embryo or gestational carrier cycles. Variables that may intrinsically alter pregnancy outcomes, such as obstetric history, infertility diagnosis, and the number and quality of embryos transferred, were controlled for in our linear or logistic regression model, so that our results could be specifically linked to the effect of the sperm source. There are several limitations associated with our study. Although ART outcome data submitted to SART CORS have been validated, all births and perinatal outcome data are self-reported, and thus recall bias is inevitable. Most perinatal outcomes are not reported to SART CORS; therefore, we could not include some important variables that may reflect the direct result of maternal immune maladaptation and abnormal placentation, such as preeclampsia or gestational hypertension ( 17 – 19 ). Most, but not all, clinics in the United States report to SART, so the dataset does not encompass all U.S. practices. In addition, donor sperm cycles are not limited to anonymous donors, and no information on prior ART cycles is available in SART CORS; therefore, prior exposure to paternal antigen of known or the same donor cannot be excluded. The frequency of exposure to partner sperm is also unknown. Severe male factor infertility could potentially affect the perinatal outcomes. However, the SART database does not provide information on sperm quality, and therefore we were unable to exclude cycles with severe oligospermia. We did not request sperm retrieval methods for the dataset and therefore were unable to analyze the cycles using surgically retrieved partner sperm as a subgroup. Additionally, data on the sperm source are not available in the SART database for frozen embryo transfer cycles; therefore, we limited the scope of this study to fresh autologous cycles only. In conclusion, by analyzing a large database reported to SART CORS and adjusting for various potential confounding variables, we observed that the use of donor sperm in fresh autologous ART cycles was not associated with increased miscarriage, preterm births, or low birthweights, as compared to cycles using partner sperm. This information can be used in patient counseling to reassure patients who use donor sperm in ART cycles.