11.1 Introduction

Since the birth of Louise Brown, the first ‘test tube baby’, in 1978, assisted reproductive technology (ART) has become one of the standard treatments for couples with subfertility problems. Today, it is estimated that 8 million children have been born via ART worldwide and up to 6% of newborns in Europe are conceived via this technique (Adamson et al., 2019). Surprisingly, relatively little is known about the short- and especially long-term effects of ART manipulations on the health risks for the children.

In the beginning of IVF, multiple embryos were transferred to increase the chance of live birth. This practice resulted in a high rate of multiple pregnancies, with the inherent increased risk for preterm birth, low birthweight, and other health problems. The transition towards single embryo transfer (SET) decreased the multiple birth rate considerably and concomitantly the perinatal risks for ART children decreased as well. Nevertheless, a smaller, but still increased risk remained in singleton ART children. Berntsen et al. (2019) summarized the outcomes of the large cohort studies and meta-analyses. The results of the studies are consistent, showing an increased risk for preterm birth (adjusted risk (AR) 1.41–2.04), very preterm birth (AR 1.68–3.07), low birthweight (AR 1.6–1.7), very low birthweight (AR 1.8–3.0), small for gestational age (AR around 1.5), and perinatal mortality (AR 1.7–2.0) when compared to spontaneous conceptions. Alongside, an increased risk for obstetric complications, such as gestational diabetes, pregnancy related hypertension, placenta praevia, and abruption and preterm rupture of membranes is reported (Berntsen et al., 2019). The adverse perinatal outcomes are to a large part attributable to subfertility-related factors in the parents, as the risk for these outcomes is higher in spontaneously conceived women with a time to pregnancy of more than 1 year as compared to less than 1 year (Pinborg et al., 2013). However, among siblings, the child conceived via ART has a poorer perinatal outcome than the child conceived spontaneously. This implies that ART related manipulations also play an important role.

Studies investigating long-term health effects are largely reassuring as no increase or no consistently reported increase in malignancies, adverse neurodevelopmental health outcomes, or growth alterations have been reported (Berntsen et al., 2019). However, the cardiovascular health of ART offspring deserves attention. In a meta-analysis, including 872 IVF-ICSI offspring and 3034 spontaneously conceived offspring, the systolic and diastolic blood pressure of ART offspring was 1.88 (95% CI 0.27–3.49) and 1.51 (95% CI 0.33–2.70) mmHg higher, respectively. When stratifying the analysis according to decade in which the children were born, the higher blood pressure was only present in the older cohorts born between 1990 and 1999 and not in the more recent cohorts born between 2000 and 2009 (Guo et al., 2017). On the other hand, in five studies in which a large proportion of children were born in the 20th century, suboptimal cardiac diastolic function and cardiovascular morphology, such as intima media, thickness were reported (Guo et al., 2017). As ART offspring are growing older, future will reveal if these cardiovascular findings are evident in ART-conceived adults.

These reported adverse perinatal outcomes together with the cardiovascular alterations are reasons for concern as it is known from other non-IVF related cohorts, starting with the landmark studies from Professors Barker and Osmond, that prematurity and low birthweight are associated with an increased risk for long-term health effects, such as cardiovascular diseases (Roseboom, 2018). This association is captured in the developmental origins of health and disease (DOHaD) hypothesis. Low birthweight is seen as an indicator of altered fetal growth and other adaptations (programming) that might affect long-term health and disease. Environmental exposures during pregnancy, but possibly also during early embryonic development, may lead to adaptations in the physiology and metabolism of organs and cells. These adaptations are in all probability beneficial for the short term, but may increase the vulnerability for diseases in later life. An illustrative example comes from people exposed in utero to the Dutch famine in World War II. When exposed early in pregnancy, they had a similar birthweight as nonexposed, but a doubled risk for cardiovascular disease in later life (Roseboom, 2018). In animals, nutritional changes during the few days around conception only, already lead to long-term health effects in the offspring, such as impaired glucose tolerance and cardiovascular dysfunction (Roseboom, 2018). Altogether, this makes it plausible that IVF children could be at risk and urges for research into the health effects of IVF on the children, the specific parts of IVF that are causative, and the mechanisms by which these programming processes work.

Obtaining reliable results from human follow-up studies is, however, not so obvious, given the (methodological) issues associated with these kind of studies. These include the retrospective design of most studies and the small sample sizes (while large numbers are needed to detect the relative small differences that are to be expected). Furthermore, phenotypic and/or health parameters such as birthweight, child growth, cognitive development, and cardiometabolic parameters are subject to many confounding factors, such as socioeconomic status and lifestyle, as well as genetic factors for which correction is not always possible. Inherently associated with human ART follow-up studies are the absence of ‘old’ ART cohorts as the technique is relatively young and the parental subfertility that always interferes when studying effects of ART procedures. Finally, a risk for publication bias exists as studies that do find an association between ART and health outcome parameters will be published more readily.

Given these challenges, animal studies can prove to be valuable with respect to unraveling potential health effects of ART-related procedures. One of the most important advantages is that patient-related confounding factors, such as subfertility, genetics, and lifestyle (Pinborg et al., 2013), are of no importance in animal studies. However, it must be realized that differences between species and in ART protocols make it difficult to deduce definitive answers. For example, in domestic animals (cattle, sheep, horses) embryos are obtained from in vitro matured oocytes while most human ART oocytes mature in vivo. Furthermore, embryos in animal studies are mainly transferred to young fertile recipients in a cycle without superovulation (Duranthon & Chavatte-Palmer, 2018). Also, relatively few studies have been published, concerning the long-term effects of ART procedures on the offspring in animal models, which is perhaps surprising as the shorter lifespan would simplify these studies. The limited success and use of ART, resulting in relatively low number of ART offspring produced in many animal models, are likely to be the main causes. We will therefore focus on two animal species in which these problems are least prominent, cattle and mice, to discuss the effects of embryo culture on health of the offspring, and relate to the limited data available from human studies. In the end, we will briefly touch upon possible programming mechanisms that are affected by embryo culture.

11.2 Embryo Culture and IVF Offspring Outcome – Animal Studies

In cattle, ART is widely used to accelerate the genetic improvement in breeding programs and almost a million embryos are produced yearly (Duranthon & Chavatte-Palmer, 2018; Ealy et al., 2019). The independent effect of in vitro culture is difficult to derive as most studies in cattle use in vitro produced (IVP) embryos that usually originate from in vitro matured oocytes that are fertilized and cultured in vitro to the blastocyst stage (Duranthon & Chavatte-Palmer, 2018). In vitro production affects the success rates in terms of live born calves after transfer (Wondim et al., 2014; Duranthon & Chavatte-Palmer, 2018; Salilew-Vrooman & Bartolomei, 2017, Wrenzycki, 2018; Ealy et al., 2019). Pregnancy rates are 10–40% lower compared to superovulation only and pregnancy losses are higher, ranging from 60% to 85%, compared to 10–65% for in vivo produced embryos after artificial insemination.

Morphokinetic differences exist between IVP and in vivo-generated conceptuses during the peri-attachment period (day 12–20 of gestation), as well as differences in molecules secreted by the conceptus and the maternal endometrium during their interaction at implantation. In early gestation, IVP fetuses are reduced in size compared to fetuses after in vivo conception, while in late gestation, IVP fetuses are generally larger. Adverse perinatal outcomes after IVP, commonly known as the large offspring syndrome, have been reported with birthweights of up to twice the expected weight and a high incidence of congenital abnormalities, such as limb deformities, an enlarged tongue (macroglossia), and heart abnormalities. Neonatal health is also compromised after IVP with increases in the incidences of stillbirth and neonatal death (Duranthon & Chavatte-Palmer, 2018; Ealy et al., 2019).

Few studies describe long-term health effects of IVP up to the adult age, as animals are often sold to clients and/or are culled at a relatively young age. In two studies, IVP-generated calves surviving the neonatal period had normal mortality rates and economic outcome parameters, such as reproductive function and milk production. In another study in which X-sorted semen was used, an increased risk of mortality and reduced milk yield in a subgroup of IVP-generated calves was reported (summarized in Duranthon & Chavatte-Palmer, 2018; Ealy et al., 2019).

Although the reported effects of IVP may be attributed to ART aspects other than the in vitro culture of embryos, suboptimal conditions such as the use of coculture with feeder cells or the presence of serum in culture medium, have shown to induce abnormal perinatal large offspring syndrome phenotype (Ealy et al., 2019).

Rodents such as the rat, rabbit, and especially the mouse have been widely used as models for the evaluation of the effects of ART on fetal development and long-term health of the offspring. Many studies have shown that fetal growth and birthweight are affected when embryos are cultured in vitro during the preimplantation embryo stages, as compared to in vivo developed controls (see Zandstra et al., 2015 for overview). Studies found that in vitro culture could affect postnatal growth and organ size, as well as adult health including glucose metabolism, obesity, cardiovascular health, and behavior, often in a sex- and strain-specific manner (see extensive reviews: Calle et al., 2012; Duranthon & Chavatte-Palmer, 2018; Feuer & Rinaudo, 2017; Gardner & Kelley, 2017; Vrooman & Bartolomei, 2017). However, whether or not these adverse health effects also result in a shortened life span remains to be determined. In one study, no effect of embryo culture was found on longevity of offspring in mice, while in another study, a significantly shorter lifespan was found, but only in combination with being fed a high fat western diet (Feuer & Rinaudo, 2017).

It remains to be determined which of the different ART culture aspects contributes the most to these health effects. It must be realized that during in vitro culture, embryos are exposed to many stressors that are absent or different in the in vivo environment in which they normally develop, such as fluctuations in temperature, oxygen, pH, light exposure, the absence of growth factors and cytokines, the type of protein source used, and the relatively large volume of surrounding medium. However, it is clear that the environment in which embryos are cultured during their preimplantation development can have a profound effect on perinatal and postnatal health. Two aspects of the in vitro environment of embryos are most widely studied: culture media and the oxygen concentration in the gas phase.

Several investigators have studied the effect of specific culture media components on long-term health outcomes. The group of Gutiérrez-Adán studied the effect of supplementing the culture medium with either fetal calf serum (FCS) or bovine serum albumin in a mouse model and found that offspring of the FCS group had a significantly higher body weight and increased liver and heart weight at 20 weeks of age, with females being more affected than males (discussed in Duranthon & Chavatte-Palmer, 2018). The group of Banrezes and Ozil studied the effect of transiently (only 15 hours during the zygote stage) modifying the redox potential by providing only a single exogenous carbohydrate (either pyruvate or lactate) to mouse embryos, which resulted in induced NAD(P)H oxidation and a low mitochondrial activity in the embryos. Significant differences up to the age of 20 weeks were found in the growth profiles of the offspring from the different experimental groups (Duranthon & Chavatte-Palmer, 2018). The group of Fleming studied the effect of supplementing the culture medium with a low concentration of insulin and so-called branched-chain amino acids (BCAA) valine, isoleucine, and leucine. The concentration of insulin and BCAA was chosen to mimic the effect on the insulin level in serum and BCAA level in uterine luminal fluid found in their in vivo model in which females were fed a low protein diet. They found a significant increase in birthweight and weight gain during early postnatal life up to 10 weeks of age. Furthermore, relative hypertension in male offspring was found as well as reduced heart/body weight in female offspring (Velazquez et al., 2018). In addition, accumulated waste products in the culture medium, such as ammonium, have been found to correlate to incidences of neural tube birth defects (Gardner & Kelley, 2017).

The oxygen concentration in the gas phase used during in vitro culture is a second important factor that affects embryo development and perinatal outcome. The concentration of oxygen in the reproductive tract of all investigated species is typically between 2% and 8%, compared to atmospheric oxygen concentration of approximately 20%. Culture of embryos under atmospheric oxygen concentrations has been common practice in the past (and still is in many laboratories today) in both animal and human IVF. As summarized in several studies (Salilew-Wondim et al., 2014; Wale & Gardner, 2016; Gardner & Kelley, 2017; Mani & Mainigi, 2018), culture of preimplantation embryos under the more physiological concentration of 5% oxygen resulted in more optimal development with higher blastocyst rates, higher mean number of cells per blastocyst, and higher fetal development rates in many species including sheep, cow, goat, pig, and mice, as well as humans.

Despite evidence, indications that atmospheric oxygen levels lead to long-term effects are scarce. However, it makes an embryo more susceptible to a second stressor such as suboptimal culture medium, single embryo culture, and/or exposure to ammonium (Feuer & Rinaudo, 2017, Gardner & Kelley, 2017). This was demonstrated in the studies by Feuer and Rinaudo (2017). They cultured mouse embryos in different culture media and oxygen concentrations and found that offspring from embryos that were cultured in Whitten’s medium under 20% oxygen had higher body weights and exhibited glucose intolerance at 19 weeks of age compared with control mice (in vivo-derived blastocysts transferred to foster mothers) while offspring from embryos that were cultured in KSOM medium with amino acids under 5% oxygen had a normal postnatal growth and glucose homeostasis (Feuer & Rinaudo, 2017).