Could we observe long term cognitive impacts of artificial reproductive technologies?

Introduction 

There has been a substantial increase in the conception rates for women aged 40 and over in the UK, for the fourth consecutive year, according to the Office of National Statistics in 2019 (Conceptions in England and Wales - ONS, 2020). As age is the single most important determinant in a couple’s fertility and pregnancy outcomes, assisted or otherwise (Balasch, 2010), the increasing maternal and paternal ages in the UK population make the investigations into the outcomes of Assistive Reproductive Technology imperative for the future of humankind. The risks associated with the offspring of infertile couples are not exclusive to the factor of maternal/paternal age, it extends to various factors namely the stress caused by infertility (Rooney & Domar, 2018). Maternal stress during pregnancy causes an increase in stress hormones such as cortisol (Cao-Lei et al., 2016, 2020; Kovas & Fatos, 2021) which can cause epigenetic changes in offspring. Epigenetics regulate gene expression through DNA methylation and histone modification, and are essential for development and cellular differentiation, including brain cells, meaning that these factors affect neurological development (Kundakovic & Jaric, 2017). Epigenetic alternations resulting from in-utero stress exposure have been found to have short- and long-lasting effects, as exposures to stress and toxins, such as BPA (Perera et al., 2012) and cigarette smoke (Toledo-Rodriguez et al., 2010), can disrupt normal brain development and contribute to neurodevelopmental disorders, and psychiatric disorders (Kundakovic & Jaric, 2017). Embryonic preimplantation is thought to be especially vulnerable to endogenous and exogenous environmental factors, such as stress and nutrition because the epigenome in the cells of an embryo undergoes drastic changes. These factors have been shown to affect disease susceptibility in offspring (Li et al., 2019).

It is known in the literature that the use of Assistive Reproductive Technology (ART) is increasing, with the number of ART babies reaching 0.5 million as a global estimate in 2011 (Adamson et al., 2018). It is also well documented that ART children are at an increased risk of congenital malformations (Pinborg, Henningsen, et al., 2013) and perinatal complications such as preterm birth (PTB), low birth weight (LBW) and being considered small for gestational age (SGA) according to Helmerhorst et al., (2004), McDonald et al., (2009) and Pinborg, Wennerholm, et al., (2013).  Specifically, studies indicate that the relative risk for ART singleton very preterm birth (<32 weeks) is 3.27, and 2.04 for preterm (<37 weeks). Helmerhorst et al., (2004) found that the relative risk for preterm birth of assistive conception was double that of the naturally conceived in the 12 matched studies assessed (namely, Dhont et al., 1999, 1999; Helmerhorst et al., 2004b; Isaksson et al., 2002; Koivurova et al., 2002; Koudstaal et al., 2000; Nuojua-Huttunen et al., 1999; Reubinoff et al., 1997; Tallo et al., 1995; Tan et al., 1992; Tanbo et al., 1995; Verlaenen et al., 1995) and 2 non-matched studies (Addor et al., 1998; Olivennes et al., 1993). Although the risk is higher for ART singletons compared to naturally conceived singletons, this disparity is not the case for ART twins. Interestingly, ART twins present with a 40% lower risk of perinatal mortality when compared with naturally conceived twins (Helmerhorst et al., 2004). In essence, when investigating the effects of ART on outcomes, the factor of twinning is important to consider. The common outcomes associated with low birth weight, are thus applicable to twins and ART singletons because they have higher proportional rates of LBW. The development of ART children may differ from naturally conceived children, in that they are more likely to suffer these effects.

Risks Associated with The ART Process Itself

Various factors have been attributed to these differences in risk between ART and non-ART offspring. For instance, it has been speculated that the stages involved in IVF treatment can facilitate short and long-term health risks (McDonald et al., 2009) by affecting development during the embryo transfers, which are environmental and epigenetic factors that affect growth (Steel & Sutcliffe, 2009) during the germinal and embryonic stages of life.  In the sense that hormonal stimulation and oocyte retrieval are extreme environmental manipulations which can affect outcomes of offspring in terms of DNA expression. Pinborg and Wennerholm, et al., (2013) suggested investigating milder ovarian stimulation to provide improved endometrial conditions and embryo quality avoiding perinatal risks by reducing epigenetic modification of DNA caused by the ART process.  Although Olivennes et al., (1993) found no differences between outcomes of IVF and natural pregnancies, the process of IVF itself was not to blame, more notably, adverse outcomes were more common when considering ovarian stimulation in both IVF and Natural groups. This alone emphasises the importance of the specificity of ART treatment in research in the area. The process itself may be a causal factor for differences in obstetric risk and developmental outcomes. One can infer this when looking at outcomes for ART and non-ART siblings from the same mother, as the ART singletons tend to have poorer developmental outcomes compared to their naturally conceived siblings (Pinborg, Henningsen, et al., 2013). In addition, parental subfertility itself can be considered a risk factor for ART offspring (Pinborg, Henningsen, et al., 2013). It is the case that environmental risk factors such as maternal gestational stress, gestational diabetes and smoking during pregnancy are influenced by maternal genetics and the genetically influenced behaviour of the mother. These factors are of particular consideration to the ART samples, where prenatal stress relating to genetics and genetically influenced behaviour may have affected fertility in these samples (Thapar et al., 2007). 

Parental Hypothesis

The attachment of parents to ART children has also been mentioned as a way in which ART children may differ from non-ART. Prenatal attachment in pregnancy has indeed been found to be the same or higher for ART couples (Ranjbar et al., 2020). However, it is a leap to suggest that overall longitudinal attachment outcomes are vastly different between these groups, or that any differences equate to differences in outcomes despite the well-established relationships between attachment and developmental outcomes. These relationships include outcomes such as cognitive performance in school (Moss & St-Laurent, 2001), and quality of adult relationships (Bartholomew, 1993;  Murphy et al., 2014). It is the case that, IVF parents have been found to show little difference in parenting style, besides from a small, but significant increase in controlling behaviours (Y.-M. Wang et al., 2014). Conversely, ART parents have been classified as “inept” and “at-risk” more often than non-ART parents when investigated in a study examining their interactive styles with their offspring (Agostini et al., 2020). This was less so when their offspring were conceived within the first cycle. This could be related again to the ART processes themselves, being that the stress of repeated attempts of conception may impact their propensities for certain interactions with their offspring.

Obstetric risk and Cognitive Development

 

Considering the considerable maternal stress of multiple cycles of treatment and other various factors, it should be stated that brain development is subject to a complicated interplay of genetic and environmental factors (Giannopoulou et al., 2018). This could include parenting, maternal stress, obstetric complications and the epigenetic markers associated with IVF treatments as discussed.

Foetal hypoxia is a common hypoxia-related obstetric complication, which has long-term effects that can persist into adulthood. There is a particular concern for the subsequent dopaminergic system disturbances which can impact neurology and cognition, in part due to the relation between these disturbances and psychopathologies (Giannopoulou et al., 2018). Hypoxia during the neonatal stage is an environmental risk factor for schizophrenia, ADHD, and ASD, all of which relate to cognitive ability. Although reduced cognitive ability is affected by various factors beyond obstetric risk and oxygen deprivation. Other Intrauterine environmental factors also relate to neurocognitive outcomes, such as poor maternal nutrition is also related to schizophrenia (Neugebauer, 2005), and gestational stress is associated with anxiety and depression in offspring (Glover & O’Connor, 2002). Weight at birth is also a major consideration for cognitive outcomes.

There is much concern over the links between the negative effects of low birthweight (LBW) on chronic diseases such as cardiovascular disease and diabetes (Thapar et al., 2007), and the relationships with intellectual development. These links have been well documented and may even extend to low birthweights in the normal ranges (Luciano et al., 2004; Martyn et al., 1996; Richards et al., 2001; Shenkin et al., 2001). For instance, Intrauterine growth restriction (IUGR) is associated with poor outcomes, such as that mathematics and memory ability as found by Swamy et al., (2018) when comparing the heavier and lighter monochorionic twins. Previous data has shown that twins are more at risk for neurodevelopmental disabilities in part because they account for a significant proportion of preterm and low BW infants. To the point where twinning may itself be considered a neurodevelopmental disadvantage (Luu & Vohr, 2009).

Looking at cognitive ability in relation to BW specifically,  Richards et al.'s (2001) study found an established positive correlation between birth weight (>2500g) and cognitive performance, with significance at ages 8, 11, 15 and 26, although they were unrelated at age 43.  In addition, a study by (Jefferis et al., 2002) investigated verbal cognitive ability and birthweight with consideration of socioeconomic status and found that each contributed independently, which is concordant with Shenkin et al., (2001) who also found BW explained a 3.8% variance in IQ at age 11, but with no relationship between BW and social class. However, models on verbal IQ predicting birth weight have been found to be a better prediction than birth weight as the predictor. Given that birthweight precedes verbal IQ at 16, this study suggested verbal IQ could be considered a proxy for parental education, in that parents who provide better learning environments, leading to better verbal IQ ability in their children, were also likely to provide better prenatal environments (Boomsma et al., 2001). This expresses the complexity surrounding such associations between physiological differences in offspring and developmental outcomes, as it is difficult to assess parental differences when investigating associations like these. Given the associations between cognitive ability and birthweight across the lifespan, and the increased risk for preterm birth and low, and very low birthweight in ART children, it is important to establish if this association is also applicable to this group. 

Conclusion

In conclusion, it is possible that the bodily risks associated with the processes of ART could affect cognitive functioning. It is pertinent, given the mixed literature on cognitive performance, that a longitudinal cohort study be used to examine any group difference in IVF and naturally conceived children. This would show differences, if they are statistically significant, throughout the life course. Work on this is currently being conducted by Dr. Yulia Kovas and her research team at Goldsmiths, University of London. I will attach a link to the research when it is published for anyone interested in reading their study.