Discussion
The main finding of this study is that in vitro culture of human embryos in medium from Cook resulted in singletons with a lower weight during the first 2 years of life compared with singletons born after embryo culture in medium from Vitrolife. This indicates that the effect of culture medium on fetal development and birthweight (Dumoulin et al., 2010; Nelissen et al., 2012, 2013) persists during the first 2 years of life.
Since the publication of our previous study (Dumoulin et al., 2010; Nelissen et al., 2012), there have been several other studies published on the effect of culture medium on birthweight. Several of these observational studies, in which other culture media were compared, found no differences in birthweight between the culture medium groups (Eaton et al., 2012; Vergouw et al., 2012; Carrasco et al., 2013; Lin et al., 2013). However, in these studies the culture media have been used in consecutive time periods. Carrasco et al. performed a small randomized study (n = 98), comparing newer versions of the Cook and Vitrolife media (Carrasco et al., 2013). Although, the average birthweight was lower in the Cook group compared with the Vitrolife group, they found no significant difference in birthweight between the groups. Recently, Eskild et al. published a large study including the birthweights of 2435 singletons after the use of 3 different culture media and compared those with the birthweights of offspring from spontaneous conceptions (n = 698 359) during the corresponding time periods (Eskild et al., 2013). They found a significant effect of culture medium on birthweight of the offspring, also after adjustment for changes in birthweight after spontaneous pregnancies. To our knowledge, there are no prior studies on the effect of culture medium on post-natal growth in humans.
This study allows a valid comparison of the effect of two commercially available embryo culture media on post-natal growth in IVF children, since exactly the same ovarian stimulation, fertilization, culture and embryo transfer procedures were applied in each group. Furthermore, several factors that might influence post-natal growth were included in the analysis, for example parental weight and height, which are known to be contributors to post-natal child growth (Blair et al., 2004; Griffiths et al., 2007; Hindmarsh et al., 2008; Mesman et al., 2009). However, it was not possible to include all factors, e.g. childhood diseases or nutrition, as this information was not available.
Although the components in the media from this study are known, the exact concentrations of the different components were not disclosed by the suppliers. Therefore, it would be speculative to relate the differences found here to specific components in the culture media.
Since the first use of IVF, there has been a major interest in the health of IVF children. Several studies have investigated the growth of children born after IVF. The vast majority of these studies did not find differences in post-natal weight and height between IVF children and naturally conceived children (Ceelen et al., 2009; Basatemur et al., 2010; Beydoun et al., 2010; Lee et al., 2010; Woldringh et al., 2011). However, Ceelen et al. found that growth velocity was higher in IVF children compared with naturally conceived children and that rapid weight gain during early childhood was associated with increased cardiovascular risks in IVF children (Ceelen et al., 2009). Furthermore, it was found that healthy IVF children and adolescents displayed systemic and pulmonary vascular dysfunction, which appeared to be related to the IVF procedure itself (Scherrer et al., 2012).
It is too early to know if the differences that we have found in this study have any clinical significance. However, the findings are in line with the growing body of evidence that suggests that the fetus is sensitive to the uterine environment, and that adaptations of the fetus to its environment have lasting consequences for its development, growth and health. This was first proposed by Barker, who found an association between low birthweight and increased rates of coronary heart disease and the related disorders of stroke, hypertension and type 2 diabetes during adult life (Barker, 1997). In the Dutch famine study it has been observed that adults who had been exposed to famine in early gestation had higher rates of coronary heart disease, a more atherogenic lipid profile, disturbed blood coagulation, increased stress responsiveness and were more often obese than those exposed in mid- or late gestation and non-exposed individuals (Roseboom et al., 2006). Furthermore, it has been suggested that, even during the periconceptional period, gametes and preimplantation embryos adapt to their environment with long-term health consequences (Steegers-Theunissen et al., 2013). A maternal low-protein diet given to pregnant mice or rats exclusively during the preimplantation period resulted in offspring with increased weight, sustained hypertension, and abnormal anxiety-related behaviour (Kwong et al., 2000; Watkins et al., 2008). Maternal under-nutrition around the periconceptional and preimplantation period affects fetal and placental growth in sheep (MacLaughlin et al., 2005). Fernandez-Gonzalez et al. showed that female mice aged 31–70 weeks, that had been cultured during the preimplantation period in medium with serum, were significantly heavier compared with female mice cultured as an embryo in medium without serum (Fernandez-Gonzalez et al., 2004). Furthermore, mice sacrificed at 20 months of age showed large organs when cultured in medium with serum during the preimplantation period, as well as long-term neurodevelopmental and behavioural effects (Ecker et al., 2004; Fernandez-Gonzalez et al., 2004).
Several animal studies have shown that culture medium affects gene expression (van Montfoort et al., 2012). Rinaudo and Schultz (2004) compared global patterns of gene expression in mouse blastocysts cultured in either Whitten's medium or KSOM/AA with that of blastocysts that developed in vivo. Culture in Whitten's medium affected the expression of 114 genes, while only 29 genes were differently expressed after culture in KSOM/AA compared with in vivo counterparts. A side-by-side comparison of five commercial culture media showed that all five media had a varying but compromised ability to maintain genomic imprinting in comparison with in vivo-derived mouse embryos (Market-Velker et al., 2010). Khosla et al. showed that the presence of serum in culture medium reduces fetal weight and influences the regulation of multiple growth-related imprinted genes in mice (Khosla et al., 2001). In humans, the effect of culture medium on gene expression is currently unknown.
The culture medium is the direct environment of the preimplantation embryo for several days during an IVF treatment. It seems that the embryo makes adaptations, likely by epigenetic modifications, during this short period of time, which will have long-lasting effects. In view of these results and the growing number of children conceived with ART, it is important that the effect of embryo culture medium is studied in the human more extensively, to select the optimal culture medium and to minimize short-term risks and perhaps even susceptibility to disease in later life.