The Risk of Birth Defects With Conception by ART

Barbara Luke; Morton B. Brown; Ethan Wantman; Nina E. Forestieri; Marilyn L. Browne; Sarah C. Fisher; Mahsa M. Yazdy; Mary K. Ethen; Mark A. Canfield; Stephanie Watkins; Hazel B. Nichols; Leslie V. Farland; Sergio Oehninger; Kevin J. Doody; Michael L. Eisenberg; Valerie L. Baker


Hum Reprod. 2021;36(1):116-129. 

In This Article


We found that singleton infants conceived with ART (using autologous oocytes, fresh embryos, and without ICSI) were 18% more likely than naturally conceived infants to have a major nonchromosomal birth defect; with ICSI in the absence of male factor diagnosis, the risk increased to 30%; with male factor, the risk increased to 42%. The 18% increased risk of a major nonchromosomal birth defect in singleton infants conceived with ART without ICSI (~36% of ART births), the 30% increased risk with ICSI without male factor (~33% of ART births), and the 42% increased risk with ICSI and male factor (~31% of ART births) translates into an estimated excess of 386 major birth defects among the 68 908 singleton children born by ART in 2017.

Our prevalence rates of birth defects are in accord with both US and European rates (State Birth Defects Surveillance Program Directory, 2016; EUROCAT prevalence rates, 2020), as well as our prior research in Massachusetts (Luke et al., 2017a,b). These findings are consistent with the pooled estimate of 1.32 (95% CI 1.24, 1.42) in 45 studies of ART reported by Hansen et al. (2013). Our findings of higher birth defects rates among twins compared to singletons are also in accord with prior studies (Hansen et al., 2013).

Comparison to other published studies is challenging due to differences in case ascertainment, birth defects definitions, reporting of ART treatment parameters, time periods and periods of follow-up, and failure to differentiate births by plurality. In the CDC population-based 2000–2010 study in Florida, Massachusetts and Michigan, Boulet et al. (2016) reported on a limited number of birth defects because of differences in case ascertainment across States (passive surveillance in Florida and Michigan, and active surveillance in Massachusetts), focusing only on defects usually diagnosed at birth. Although their reported singleton major nonchromosomal birth defects rates (per 10 000 live births) by oocyte source-embryo state combinations were much lower (ranging from 56 to 79) compared to our study (189 to 264), both studies found no significant difference by the combination of these two factors. Our study also differed because our ART-conceived group, which was compared to natural and OI/IUI conceived groups, included only infants conceived using autologous oocytes and fresh embryos (to be more physiologically comparable to naturally conceived infants), whereas in the CDC study the ART group included both autologous and donor oocytes as well as fresh and thawed embryos. Also, in contrast, our analyses did not show an increased risk of birth defects with the use of assisted hatching (data not shown), whereas the CDC study showed a 55% increased risk; other studies have also reported no increased risk (Jwa et al., 2015).

Blastogenesis Risk

Halliday et al. (2010), in their singleton study of 20 838 non-ART controls and 6946 ART children, reported increased blastogenesis defects risks with both ICSI (AOR 2.33, 95% CI 1.12, 4.87) and fresh embryos (AOR 3.65, 95% CI 2.02, 6.59). Our results indicate lower blastogenesis risks compared to those reported by Halliday, including increased risks with ICSI in the absence and presence of male factor (AOR 1.49, 95% CI 1.08, 2.05, and AOR 1.56, 95% CI 1.17, 2.08, respectively among singletons, and AOR 1.50, 95% CI 1.01, 2.23 with ICSI in the absence of male factor among twins). Other factors we found to be significantly associated with an increased risk of blastogenesis defects in singletons were older maternal age (≥44 years, AOR 1.80, 95% CI 1.12, 2.88), diabetes (pre-pregnancy or gestational, AOR 1.46, 95% CI 1.25, 1.69), and male infant sex (AOR 1.17, 95% CI 1.08, 1.27 in singletons and AOR 1.30, 95% CI 1.01, 1.67 in twins). These differences may have been due to our exclusion of cardiac defects in defining blastogenesis defects, and limiting to live births only, as well as changes in culture media over the study periods (1991–2004 in the Halliday study, and 2004–2016 in our study); the Halliday study also included pregnancies terminated for a birth defect at any gestation. As blastogenesis defects may have an environmental etiology, including aspects of ART treatment, these associations should be investigated further.

ICSI and Birth Defects Risk

As a commonly used procedure in ART, the use of ICSI has increased in the USA from 11% in 1995 to 67% in 2017 (Toner et al., 2016; Centers for Disease Control and Prevention, 2019). This trend is also evident internationally, with 66% of cycles using ICSI in 2010, ranging from 56% of cycles in Asia to 96% of cycles in the Middle East (Dyer et al., 2016). The use of ICSI offers hope of genetic parenthood for men with profound oligospermia (low sperm count) and, by means of testicular biopsy and epididymal aspiration, even for men with azoospermia (absence of sperm). However, it is increasingly being used even in the absence of male factor infertility. There are several theoretical concerns, though, regarding ICSI and the potential risks for the offspring (de Kretser, 1995; Palermo, 2008; Woldringh et al., 2010): the risks of using sperm that potentially carry genetic abnormalities; the risks of using sperm with structural defects; the potential for mechanical and biochemical damage by introducing foreign material into the oocyte; and the risks associated with circumventing natural selection by injecting a single spermatozoon. The analyses of the outcomes of children born after ICSI have shown mixed results, including a 3-fold-increased risk of congenital heart defects (Tararbit et al., 2013), a twofold risk of major birth defects and a 50% increased risk of minor birth defects (Hansen et al., 2002; Katalinic et al., 2004; Yan et al., 2011; Davies et al., 2012a; Farhi et al., 2013), while other studies have shown no difference (Lie et al., 2005). Our results, which were limited to children conceived using fresh embryos, indicated a 30% increased risk of a major nonchromosomal birth defect with the use of ICSI in the absence of male factor diagnosis, increasing to 42% in the presence of male factor diagnosis, compared to naturally conceived singletons. These findings support the judicious use of ICSI only when medically necessary in ART-treated patients.

Fresh Versus Thawed Embryos and Birth Defect Risk

The use of frozen embryo transfer has increased by more than 80% since 2006 owing to better cryo-preservation techniques, improved live birth rates, lower risk of ectopic pregnancies, and more physiologically normal hormonal and endometrial environments (Toner et al., 2016; Centers for Disease Control and Prevention, 2019). Results indicate that singletons born after frozen embryo transfer have comparable or lower risks for low birthweight, SGA birthweight and preterm birth compared to singletons born after fresh ART, but worse outcomes compared to singletons born after natural conception, including an excess of LGA birthweights, pregnancy-induced hypertension and placenta accreta (Wada et al., 1994; Källén et al., 2005; Belva et al., 2008, 2016; Shih et al., 2008; Pinborg et al., 2010; Luke et al., 2019, 2020; Hwang et al., 2019). Belva et al. (2008) reported rates of major malformations to be highest in children born from cryopreserved embryos with ICSI (6.4%) compared to children born from cryopreserved embryos without ICSI (3.1%) and fresh embryos with ICSI (3.4%). Other studies have reported malformation rates in frozen cycles ranging from 1.0% (Wada et al., 1994) to 8.7% (Källén et al., 2005). Pinborg et al. (2010), in their study of Danish singleton births in 1995–2002, reported higher nonsignificant differences in major birth defects rates among infants conceived using fresh embryos (5.9%, 5.8% with ICSI) and thawed embryos (5.4%; 4.5% with ICSI) compared to infants of fertile controls (4.7%). Recent analyses of infants born in 2004–2013 in Massachusetts confirm small but nonsignificant differences in birth defect risks from fresh versus thawed embryos (1.8% vs 1.7%, respectively for nonchromosomal defects) (Hwang et al., 2019). A Belgian study of births in 2008–2013 reported similar results for singletons (fresh, 2.8%, thawed, 2.6%) and twins (fresh, 2.7%, thawed, 2.4%) (Belva et al., 2016). Our findings are in line with these prior reports, with frequencies of major birth defects among singletons of 2.4% and 2.5% for autologous-fresh and autologous-thawed, and 2.3% and 2.7% for donor-fresh and donor-thawed, respectively; and among twins of 3.4% and 3.8% for autologous-fresh and autologous-thawed, and 3.0% and 3.5% for donor-fresh and donor-thawed, respectively. Among singleton ART births comparing to births from autologous oocytes and fresh embryos, there were no significant differences in the risks of birth defects, except for chromosomal defects, which were decreased with the use of donor oocytes (fresh, AOR 0.12, 95% CI 0.04, 0.43; thawed, AOR 0.09, 95% CI 0.01, 0.67).

Sibling Studies

The choice of an appropriate comparison group in infertility research poses a special challenge. Although most studies compare women treated for infertility to fertile women, this approach has several potential disadvantages, including differences in age, socioeconomic status, education and reproductive history. Comparisons within families, as repeat pregnancies to the same woman, have the advantage of eliminating the fixed effects of the parents (mainly the genetic contribution), with adjustments possible for her change in age, parity, and, if appropriate, method of conception. In our prior studies of siblings in Massachusetts, declining fertility status, with or without ART, was associated with increasing risks for adverse outcomes, greatest for women whose fertility status declined the most between the two pregnancies (Luke et al., 2016a). In addition, we previously demonstrated that among singleton siblings both conceived with ART, frozen embryo state was associated with an increased risk of LGA birthweight (AOR 1.74, 95% CI 1.45, 2.08), with a birthweight difference of 222 g (SE 11) (Luke et al., 2017c). Henningsen et al. (2011) reported similar results in singleton siblings with fresh versus frozen embryo status, with a birthweight difference of 286 g. Shih et al. (2008) in their large Australian study reported a difference of 244 g in ART-conceived siblings conceived using fresh versus frozen embryos. Only one study reported on the risk of birth defects in siblings. In an additional sibling analysis to their Australian study (Davies et al., 2012a), Davies et al. (2012b) reported an increased risk of birth defects among ART-conceived siblings compared to naturally conceived siblings (crude odds ratio, 1.50, 95% CI, 1.08, 2.09). Among singleton ART siblings, our analysis showed increased risks of any defect (AOR 1.15, 95% CI 1.08, 1.23) and musculoskeletal defects (AOR 1.32, 95% CI 1.04, 1.67), and among twin ART siblings, increased risks of any defect (AOR 1.26, 95% CI 1.01, 1.57).

Our findings of an increased risk of birth defects among ART siblings who were conceived without ART suggests that subfertility may be a contributing factor. It remains difficult to separate the relative contribution of the biology of the subfertile couple versus aspects of the ART treatment to this increased risk. This information regarding birth defects should be included when counseling patients about the risks and benefits of ART. In addition, the larger context of risk versus benefit of ART versus other treatment options, such as expectant management and controlled ovarian stimulation with IUI, should be considered. ART treatment will typically lead to a shorter time to conception, mitigating the effect of advancing maternal age. ART also enables a more controllable situation with respect to the risk of multiple gestation compared with ovarian stimulation with IUI, with twins and triplets associated with many serious adverse consequences for both the mother and the children. Furthermore, some couples have fertility factors that are not treatable other than by ART. The potential for increased risk of birth defects associated with ART needs to be balanced against the potential risks associated with other options.

Challenge of Monitoring of Births and Birth Defects From ART in the USA

Unlike other countries that track their citizen's health from cradle to grave, the USA does not have a uniform system to monitor health. The US Certificate of Live Birth is the only consistent mechanism to assess population-based data on births for all States and territories. Revised periodically, the 2003 version of the birth certificate includes checkbox questions regarding the use of infertility treatment. Although the birth certificate has been suggested as a mechanism to identify children conceived with infertility treatment (Lynch et al., 2011), several validation studies of the accuracy of ART indicated on the birth certificate have reported low sensitivity (ranging from 27 to 28% overall, higher with multiples, 43%), with only 36–50% accurately reported (Zhang et al., 2010; Cohen et al., 2014; Thoma et al., 2014; Luke et al., 2016b). Birth defects have also been indicated on the birth certificate, but validation studies of the 1989 and 2003 versions compared to birth defects registry data have shown low sensitivities (23% and 19.1%, respectively) (Boulet et al., 2011; Salemi et al., 2017). Birth defects have also been included in the outcome data of ART cycles in the SART CORS, but again, validation studies showed low sensitivity (38.6% for any birth defect, ranging from 18.4 to 50% for specific birth defect categories) (Stern et al., 2016), making research findings based on these data questionable (Xiong et al., 2017; Kirby and Boulet, 2017). The birth defect variables have since been removed from research data provided by SART. When the birth certificate is used as the sole data source of both infertility treatment and birth defects, the research is doubly flawed (Shechter-Maor et al., 2018).


This study has a number of strengths, including a large sample size, population-based design, and contemporary time period. The four study States include racially and ethnically diverse populations, with high linkage rates to the SART CORS, vital records, and birth defects registries, and their birth defects registries utilize the similar case definitions and data collected. We were able to stratify our analysis by plurality, non-ART and ART conception, and additionally within the ART group by oocyte source, embryo state, and the use of ICSI, as well as including naturally conceived ART siblings. The infertility data and birth defects data were independently collected, minimizing the risk of ascertainment bias.


This study is subject to several limitations. In the SART CORS database, it was not possible to differentiate method of embryo freezing (slow-freezing vs vitrification); data on ICSI was only available in the fresh embryo ART group, not the thawed embryo group; and data were unavailable on duration of infertility, which has been reported to be related to birth defect risk (Ghazi et al., 1991; Zhu et al., 2006). Data on preimplantation testing were not available, other than the infertility diagnosis of PGD; these births were excluded from the analysis. For the OI/IUI group, it was not possible to differentiate type of non-ART treatment utilized (e.g. IUI, ovulation stimulation). We were not able to reconstruct sibling sets in twin births. Data on birth defects were not available on miscarriages, terminations or stillbirths, only on live births; this limitation is also noted in other population-based studies (Källén et al., 2005, 2010), which for legal reasons could not be included in the linkages or analyses, making it impossible to study conditions that are more likely to be terminated after prenatal detection. In addition, data were unavailable on imprinting disorders. Because of the lack of a national registry for non-ART infertility procedures, the OI/IUI group is likely underrepresented, with some treated women included in the naturally conceived groups. This underrepresentation would tend to bias toward finding less of a difference between the OI/IUI and naturally conceived groups. Although we limited the ART siblings to those naturally conceived, as indicated in the Materials and Methods, there is a possibility that this group also included children conceived with OI or IUI, or were conceived with ART treatment performed outside the USA or ART treatments not nationally reported.