Assisted reproduction technologies (ART)

ART are defined as all treatments or procedures for initiating pregnancy that include the in vitro handling of both oocytes and sperm or embryos, predominantly in vitro fertilization (IVF), intracytoplasmic sperm injection (ICSI), gamete and embryo cryopreservation, preimplantation genetic diagnosis and preimplantation genetic screening.Reference Zegers-Hochschild, Adamson and de Mouzon ¹ Briefly, the routine IVF procedure includes three steps: ovarian hyperstimulation, IVF and embryo culture, and embryo transfer. First, high doses of gonadotropins are administrated to induce development of multiple follicles. Then, the oocytes are retrieved from the ovaries using a transvaginal ultrasound-guided fine needle and inseminated with the prepared sperm in vitro to achieve fertilization. Embryos are cultured in a prepared medium in the incubator for 3 days to reach the eight-cell stage or 5 days to develop to the blastocyst stage. Finally, the embryos are transferred into the uterus or frozen for future transfer. Although IVF is beneficial for most of infertile couples with female infertility, unexplained infertility and some cases of male infertility, ICSI in which a single spermatozoon is injected into the oocyte cytoplasm is required to treat severe male infertility.

Since 1978, more than 5 million children have been born by ART treatment, mostly by IVF and ICSI.Reference Ferraretti, Goossens and Kupka ² Concerns about the potential health implications of ART remain.Reference Schieve, Rasmussen and Buck ^{3 ,} Reference Sutcliffe and Ludwig ⁴ Increasing evidence shows ART treatment is associated with adverse perinatal outcomes, which are related to subfertility of patients, multiple pregnancies and ART technologies.Reference Sutcliffe and Ludwig ^{4 –} Reference Basatemur and Sutcliffe ⁹ As it is well accepted that multiple pregnancies, after multiple embryos transfer, are associated with low birth weight and preterm delivery,Reference Hart and Norman ¹⁰ which can also impact long-term health risks, this review will limit its focus to the health outcomes of ART singletons born from IVF and ICSI v. singletons from natural conception.

Obstetric and perinatal outcome in ART singleton pregnancies

As summarized in Table 1, singleton pregnancies after ART are associated with adverse obstetric and perinatal outcome as compared with spontaneous conception.Reference Sutcliffe and Ludwig ^{4 ,} Reference Pinborg, Wennerholm and Romundstad ^{11 ,} Reference Henningsen, Pinborg and Lidegaard ¹² These outcomes include an increased risk of low birth weight, preterm birth, small for gestational age, stillbirth, perinatal mortality, admission to a neonatal intensive care unit, antepartum haemorrhage, hypertensive disorders of pregnancy, preterm rupture of membranes, gestational diabetes, induction of labour and caesarean section.Reference Wisborg, Ingerslev and Henriksen ^{8 ,} Reference Pinborg, Wennerholm and Romundstad ^{11 –} Reference Buckett, Chian and Holzer ¹⁷ It should be noted that vanishing twin pregnancies, which contribute to about 10% of IVF singletons pregnancies, increase perinatal risk in IVF singletons.Reference Pinborg, Lidegaard and Freiesleben ^{18 ,} Reference Pinborg, Lidegaard and la Cour Freiesleben ¹⁹ However, whether the procedure of IVF itself, or the underlying parental characteristics or genetics are the main contributors to this increase in obstetric and perinatal risk is not clear. Some studies have shown that IVF singletons have an increased risk of adverse perinatal outcome v. their non-IVF siblings.Reference Pinborg, Wennerholm and Romundstad ¹¹ However, other studies have shown that perinatal outcomes from spontaneous conception are also poorer in subfertile women v. those with normal fertility,Reference Basso and Baird ^{20 ,} Reference Zhu, Obel and Hammer Bech ²¹ and that perinatal outcomes are comparable after IVF or natural conception in subfertile women.Reference Raatikainen, Kuivasaari-Pirinen and Hippelainen ²² A large recent study using siblingship analysis suggested that maternal characteristics such as subfertility and maternal age but not IVF treatment are associated with lower birth weight in IVF children.Reference Seggers, Pontesilli and Ravelli ²³ This discrepancy in the literature requires further study in larger cohorts that control for as many confounders as possible, and also further pre-clinical study. When investigating risk, the type of ART procedure is not always reported, and may contribute to adverse perinatal outcomes. A recent study shows that frozen embryo transfer increases pregnancy rates, improves obstetric and perinatal outcomes, and reduces the risk of ovarian hyperstimulation syndrome in patients with polycystic ovary syndrome.Reference Chen, Shi and Sun ²⁴ A meta-analysis of 11 studies supports this, reporting that singletons born after the transfer of frozen thawed embryos had better obstetric and perinatal outcome as compared with those after the transfer of fresh IVF embryos.Reference Maheshwari, Pandey and Shetty ²⁵ The relative risks (RR) and 95% confidence intervals (CI) of antepartum haemorrhage (RR=0.67, 95% CI, 0.55–0.81), preterm birth (RR=0.84, 95% CI, 0.78–0.90), small for gestational age (RR=0.45, 95% CI, 0.30–0.66), low birth weight (RR=0.69, 95% CI, 0.62–0.76) and perinatal mortality (RR=0.68, 95% CI, 0.48–0.96) were lower in women who received frozen embryos.Reference Maheshwari, Pandey and Shetty ²⁵ The characteristics for each study was shown in the review.Reference Maheshwari, Pandey and Shetty ²⁵ This data suggests that suboptimal endometrial development, induced by hormone stimulation, may be a contributor to poorer perinatal outcome after IVF.Reference Chung, Coutifaris and Chalian ^{7 ,} Reference Henningsen, Pinborg and Lidegaard ¹² A recent retrospective cohort study suggests natural cycle IVF may decrease the risk of low birth weight v. conventional stimulated IVF,Reference Mak, Kondapalli and Celia ²⁶ but this is not reported universally.Reference LaMarca and Polyzos ²⁷ The length of embryo culture is emerging as another potential confounder when considering perinatal outcomes, but consideration may also be required as to whether sequential or single-step culture media is employed.Reference Wirleitner, Vanderzwalmen and Stecher ²⁸ Increasing evidence suggests the type of culture medium may also impact birth weight in IVF singletons.Reference Vergouw, Kostelijk and Doejaaren ^{29 ,} Reference Dumoulin, Land and Van Montfoort ³⁰ Although some studies show that blastocyst transfer is associated with a higher cumulative live birth and pregnancy rates,Reference Glujovsky, Farquhar and Quinteiro Retamar ³¹ a recent meta-analysis of six studies suggest that blastocyst transfer may increase the risk of preterm birth in IVF singleton pregnancies.Reference Dar, Lazer and Shah ³² Taken together, it seems that parental characteristics, and ART procedures themselves contribute to the adverse perinatal outcomes of singleton pregnancies after ART. Careful further studies are warranted to determine whether cleavage embryo transfer, sequential media with blastocyst transfer, minimal stimulation protocols or natural IVF improve obstetric and perinatal outcomes, and the long-term health outcomes of this on ART children.

Table 1 Obstetric and perinatal outcome in in vitro fertilization (IVF) singleton pregnancies

ICSI, intracytoplasmic sperm injection; CI, confidence intervals; ART, assisted reproductive technologies; aOR, adjusted odds ratio; OR, odds ratio.

Birth defects in ART singleton pregnancies

Major malformations were defined as those causing functional impairment or requiring surgical correction,Reference Bonduelle, Liebaers and Deketelaere ^{33 ,} Reference Bonduelle, Wilikens and Buysse ³⁴ whereas the others were considered minor malformations. The prevalence of major birth defects such as chromosomal and musculoskeletal defects diagnosed by 1 year of age is two-fold higher in infants conceived by IVF or ICSI than in naturally conceived infants born between 1993 and 1997 in Western Australia.Reference Hansen, Kurinczuk and Bower ³⁵ Importantly, this study controlled for parental factors such as maternal age and parity, the gender of the infant and correlation between siblings. In Israel, the percentage of major malformations in infants conceived by ART in 1986–1994 and 1995–2002 was also double that of the general population during the same periods.Reference Merlob, Sapir and Sulkes ³⁶ Similar reports have been observed in Spain,Reference Gutarra-Vilchez, Santamarina-Rubio and Salvador ³⁷ France,Reference Tararbit, Houyel and Bonnet ³⁸ CanadaReference El-Chaar, Yang and Gao ³⁹ and the United StatesReference Reefhuis, Honein and Schieve ^{40 ,} Reference Boulet, Kirby and Reefhuis ⁴¹ (summarized in Table 2).

Table 2 Birth defects in in vitro fertilization (IVF) singleton pregnancies

ICSI, intracytoplasmic sperm injection; CI, confidence intervals; ART, assisted reproductive technologies; aOR, adjusted odds ratio; OR, odds ratio; aRR, adjusted relative risks; IUI, intrauterine insemination.

Large meta-analyses have been conducted and show that children born after ART have a 30–40% increased risk of birth defects compared with spontaneous conceptions.Reference Rimm, Katayama and Diaz ^{42 –} Reference Qin, Sheng and Wang ⁴⁷ However, it is not entirely clear if the contributing factor is the ART procedure, or the underlying infertility itself. One Italian study of >7000 infants born after ART or ovulation induction suggested the increased prevalence of birth defects associated with non-spontaneous conception was largely due to confounding factors such as maternal age,Reference Parazzini, Cipriani and Bulfoni ⁴⁸ which is associated with poorer oocyte quality, mitochondrial dysfunction, aneuploidy and epigenetic alteration.Reference Ge, Schatten and Zhang ⁴⁹ A large Danish longitudinal study found singletons born of infertile couples who conceived naturally (time to pregnancy >12 months) or after infertility treatment including different types of ART or surgeries had a higher prevalence of congenital malformations compared with singletons born of fertile couples (time to pregnancy ⩽12 months).Reference Zhu, Basso and Obel ⁵⁰ Notably, the prevalence of congenital malformations increased with increasing time to pregnancy, suggesting that infertility per se was an independent risk factor. On the other hand, amongst infertile couples, infertility treatment was associated with an increased prevalence of genital organ malformations in singletons compared with natural conception.Reference Zhu, Basso and Obel ⁵⁰ Further, Davies et al.Reference Davies, Moore and Willson ¹⁵ compared risks of birth defects among pregnancies in women who received ART treatment, spontaneous pregnancies in women who had a previous birth with assisted conception and spontaneous pregnancies in women with or without a record of infertility. An increased risk of birth defects was significantly associated with infertility per se, independently of assisted conception. An increased risk of birth defects was also associated with assisted conception after the multivariate adjustment, however, this association was only observed in births conceived by ICSI but not by IVF, after adjustments.Reference Davies, Moore and Willson ¹⁵ In comparison, two meta-analysis reported the risk of birth defect was not significantly different between children conceived by IVF and ICSI.Reference Wen, Jiang and Ding ^{45 ,} Reference Massaro, MacLellan and Anderson ⁵¹

Ovulation induction alone has also been associated with increased risk of birth defects.Reference El-Chaar, Yang and Gao ^{39 ,} Reference Klemetti, Gissler and Sevon ⁵² Evidence shows that exogenous gonadotrophins may impair oocyte and embryo development as well as endometrial receptivity, increase chromosomal aneuploidy, alter epigenetic modifications, thus have detrimental effects on perinatal outcomes and long-term health.Reference Santos, Kuijk and Macklon ^{53 ,} Reference Vialard, Boitrelle and Molina-Gomes ⁵⁴ In Finland, ART singleton girls from ovulation induction had more major heart anomalies than controls conceived naturally.Reference Klemetti, Gissler and Sevon ⁵² Similarly, the risks of birth defects were higher in ovulation induction v. natural conception, whereas the risk was even higher in IVF v. ovulation induction.Reference El-Chaar, Yang and Gao ³⁹ Taken together, this data suggest that singletons conceived by ART procedures are at increased risk for birth defects. This is at least partly due to the underlying infertility, and parental characteristics, but may be further increased by ovulation induction and ART procedures.

Growth and development in ART singleton pregnancies

A number of studies have examined the growth patterns of ART children with conflicting resultsReference Wennerholm, Albertsson-Wikland and Bergh ^{55 –} Reference Olivennes, Kerbrat and Rufat ⁶⁹ (summarized in Table 3). The majority of studies have not observed any differences in the growth of ART children v. naturally conceived children. For instance, recent prospective follow-up studies in the United States compared 969 singletons conceived by infertility treatment including ART and ovulation induction with or without intrauterine insemination with 2471 singletons conceived naturally, and found the growth and development of children up to 3 years of age was comparable.Reference Yeung, Sundaram and Bell ^{60 ,} Reference Yeung, Sundaram and Bell ⁷⁰ Similar findings have been observed in ART children v. the general population up to 13 years of age in European countries.Reference Bonduelle, Wennerholm and Loft ^{61 ,} Reference Basatemur, Shevlin and Sutcliffe ^{62 ,} Reference Olivennes, Kerbrat and Rufat ⁶⁹ A study in the United States also reported IVF young adults exhibited normal pubertal development.Reference Beydoun, Sicignano and Beydoun ⁷¹ A handful of studies have found ART children had impairedReference Kai, Main and Andersen ^{63 ,} Reference Koivurova, Hartikainen and Sovio ⁶⁴ or enhanced childhood growth.Reference Miles, Hofman and Peek ^{65 –} Reference Green, Mouat and Miles ⁶⁷ Notably, some studies recruited children born prematurely, small for gestational age, with low birth weight or from multiple pregnancies, which may confound the results and few of these studies have controlled for subfertility. Ceelen et al. examined the growth data from birth to 4 years of age in a small follow-up study that included 233 IVF children aged 8–18 years and 233 spontaneously conceived controls born to subfertile parents. They showed IVF children had significantly lower weight, height and BMI standard deviation scores (SDSs) at 3 months, and weight SDS at 6 months of age compared with controls. IVF children demonstrated a catch-up growth during late infancy (3 months to 1 year) v. controls, such that no differences were observed in weight, height and BMI after 1 year of age between groups.Reference Ceelen, van Weissenbruch and Prein ⁶⁸ This is a small study, but potentially of concern, given evidence that rapid catch-up growth is associated with increased risk of disease later in life.Reference Ong and Loos ^{72 ,} Reference Ekelund, Ong and Linne ⁷³

Table 3 Growth and development in assisted reproductive technologies (ART) singleton pregnancies

IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; aOR, adjusted odds ratio; OR, odds ratio; IGF, insulin-like growth factor; BMI, body mass index.

A meta-analysis including four studies in singletons reported an increased risk of cerebral palsy for IVF children v. those conceived naturally.Reference Hvidtjorn, Schieve and Schendel ⁷⁴ This risk may be largely due to multiple births, low birth weight and preterm births among ART children.Reference Ericson, Nygren and Olausson ^{75 –} Reference Zhu, Hvidtjorn and Basso ⁸⁰ In Australia, an increased risk of cerebral palsy was observed in ART infants overall and for ART singletons, even after adjusting for parental and fetal factors.Reference Davies, Moore and Willson ¹⁵ Similarly, Zhu et al.Reference Zhu, Hvidtjorn and Basso ⁸⁰ found that ART infants had an increased risk of cerebral palsy after controlling for preterm birth and multiplicity, and there was no association between parental subfertility and the risk of cerebral palsy, indicating that the increased risk of cerebral palsy for ART infants was due to the effect of ART treatment.

No differences were observed in the cognitive and motor development in large cohort studies between ART children and controls examined at 3 or 5 years of age who were recruited in Europe,Reference Ponjaert-Kristoffersen, Bonduelle and Barnes ⁸¹ Great BritainReference Carson, Kurinczuk and Sacker ^{82 ,} Reference Carson, Kelly and Kurinczuk ⁸³ or the United States.Reference Yeung, Sundaram and Bell ⁷⁰ Similarly, a systematic review of 59 studies reported that children born following ART are not at increased risk of severe cognitive impairment compared with naturally conceived children.Reference Middelburg, Heineman and Bos ⁸⁴ There is also no increased risk of autism in singletons conceived by ART,Reference Grether, Qian and Croughan ^{85 –} Reference Lyall, Baker and Hertz-Picciotto ⁸⁸ but studies including IVF multiple births and autism spectrum disorders have shown conflicting results.Reference Grether, Qian and Croughan ^{85 ,} Reference Lehti, Brown and Gissler ⁸⁷ In contrast, Kissin et al.Reference Kissin, Zhang and Boulet ⁸⁹ found that the incidence of autism diagnosis in ART-conceived children during the first 5 years of life was higher when ICSI was used compared with IVF. Notably, Belva et al. showed that 54 ICSI-conceived adults had significantly lower sperm concentration, lower total sperm count as well as lower total motile sperm count but comparable mean levels of follicle-stimulating hormone, luteinizing hormone, testosterone and inhibin B in comparison to 57 spontaneously conceived peers, possibly reflecting inherited fertility problems.Reference Belva, Bonduelle and Roelants ^{90 ,} Reference Belva, Roelants and De Schepper ⁹¹ Taken together, the available data on the growth and development of ART children is generally reassuring, although an increased risk of cerebral palsy has been observed. This needs to be confirmed in large studies focussing on ART singletons born at term with normal birth weight. In addition, more follow-up studies in adults are warranted to determine if ART is associated with increased risk of impaired cognitive development and psychological adjustment, later in life.

Cancer risk in ART singleton pregnancies

Concerns are turning towards the longer-term health implications of IVF. A number of studies have been undertaken to examine the cancer risk of children conceived by ART procedures (summarized in Table 4). Most of the earlier studies demonstrate that ART procedures are not associated with increased risk of cancers.Reference Bruinsma, Venn and Lancaster ^{92 –} Reference Lerner-Geva, Toren and Chetrit ⁹⁵ For instance, one data linkage study that included 3528 ART singletons with a median follow-up period of 4 years showed that ART children did not have a significantly increased incidence of cancer.Reference Bruinsma, Venn and Lancaster ⁹² Similar results have been noted in the Netherlands over an average follow-up period of 6 yearsReference Klip, Burger and de Kraker ⁹³ and in a meta-analysis of 11 cohort studies.Reference Raimondi, Pedotti and Taioli ⁹⁶

Table 4 Cancer risk in assisted reproductive technologies (ART) singleton pregnancies

IVF, in vitro fertilization; HR, hazard ratio.

In the last decade, more studies have reported an increased risk of certain cancers in ART children. In Sweden, although there was no overall increase in cancer risk in >16,000 ART children compared with naturally conceived children, more cases of Langerhans histiocytosis were reported.Reference Kallen, Finnstrom and Nygren ⁹⁷ After 5 years, the same group reported a moderately increased risk for all cancers in 26,692 children conceived by ART during the years 1982–2005.Reference Kallen, Finnstrom and Lindam ⁹⁸ Notably, the increased cancer risk was associated with high birth weight, premature delivery and the presence of respiratory diagnoses as well as low Apgar score.Reference Kallen, Finnstrom and Lindam ⁹⁸ A large retrospective Nordic population-based cohort study found ART children had an increased risk for central nervous system tumours and malignant epithelial neoplasms v. children born spontaneously between 1982 and 2007.Reference Sundh, Henningsen and Kallen ⁹⁹ This cohort was matched for parity, year of birth and country, and controlled for maternal age, sex, gestational age and birth defects.Reference Sundh, Henningsen and Kallen ⁹⁹ Similarly, ART was associated with an increased risk of hepatoblastoma and rhabdomyosarcoma, but not the overall risk of cancer in the United Kingdom.Reference Williams, Bunch and Stiller ¹⁰⁰ This study also controlled for confounding factors such as sex, age at diagnosis, birth weight, singleton v. multiple birth, parity, parental age, type of assisted conception and cause of parental infertility. A meta-analysis, published in 2013, that included 25 cohort and case-control studies reported that children born after ART are at increased risk for all cancers (RR=1.33; 95% CI, 1.08–1.63), and specifically for leukaemias (RR=1.65; 95% CI, 1.35–2.01), neuroblastomas (RR=4.04; 95% CI, 1.24–13.18) and retinoblastomas (RR=1.62; 95% CI, 1.12–2.35). It should be noted that the majority of these studies did not control for confounders such as socioeconomic status, maternal smoking and perinatal health status, which may affect incidence.Reference Kallen, Finnstrom and Lindam ^{98 ,} Reference Schmidt, Schuz and Lahteenmaki ¹⁰¹ Further, it is unclear that whether the increased risk for cancers is related to underlying subfertility of the parent, or the ART procedure itself.Reference Hargreave, Jensen and Toender ¹⁰² More follow-up studies are needed to determine risk in children, as well as later in life.

Does ART increase the risk of chronic disease?

The long-term health implications of IVF are under-studied. Over the past decade, speculation is increasing that individuals conceived by ART may be at risk of developing metabolic syndrome, type 2 diabetes and cardiovascular disease, later in life.Reference Ceelen ^{103 –} Reference Chen, Wu and Zhao ¹⁰⁷ To date these studies are small, and this evidence is not conclusive (summarized in Table 5). Discrepancies between studies may be due to differences in the ages investigated, study period, inclusion criteria of subjects, sample size, sampling of the comparison group, dietary intake and/or parental characteristics, as well as the ART technique employed. Ceelen et al. Reference Ceelen ¹⁰³ reported an increase in body fat as assessed by skinfold thickness in IVF children who were matched for BMI. Post-pubertal IVF children in this study also had a trend towards increased body fat assessed by Dual-energy X-ray absorptiometry (DXA). Importantly, the control group studied were children who were born to subfertile parents and controlled for current size, birth weight, gestational age and parental characteristics.Reference Ceelen ¹⁰³ Belva et al. reported pubertal ICSI singleton girls had increased central, peripheral and total adiposity assessed by circumferences, skinfolds and BMI, respectively, compared with naturally conceived controls. Furthermore, increased peripheral adiposity was observed in ICSI adolescent singleton boys with advanced pubertal stages v. controls.Reference Belva, Painter and Bonduelle ¹⁰⁸ Conversely, one study reported no difference in fat percentage by DXA between IVF children and controls at 4–10 years of age.Reference Miles, Hofman and Peek ⁶⁵ More studies in ART adults are required to assess the obesity incidence and the amount of liver and visceral fat which is clearly associated with increased risk of type 2 diabetes and cardiovascular diseases.Reference van Harmelen, Eriksson and Astrom ¹⁰⁹

Table 5 Metabolic risk in assisted reproductive technologies (ART) singleton pregnancies

IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection

There is some suggestion that ART may also impair glucose metabolism in the offspring, potentially as a result of increased adiposity. Ceelen et al.Reference Ceelen, van Weissenbruch and Roos ¹⁰⁴ reported IVF adolescents had elevated fasting glucose levels compared with controls, irrespective of any early life factors or parental characteristics. However, there was no significant difference in fasting insulin levels, and insulin sensitivity as measured by the homeostasis assessment model. Another study reported fasting glucose levels were higher among children aged 5–6 years old conceived through ovulation induction and ART compared with naturally conceived children from fertile couples.Reference Pontesilli, Painter and Grooten ¹¹⁰ Conversely,Reference Sakka, Loutradis and Kanaka-Gantenbein ¹⁰⁵ there was no difference in weight, glucose, insulin, leptin, adiponectin, interleukin-6 or C-reactive protein in IVF children and controls. However, this study may be confounded by neonatal and parental factors. Another study reported IVF children were taller with significantly higher IGF1 serum levels.Reference Miles, Hofman and Peek ⁶⁵ Reduced peripheral insulin sensitivity was also observed in a small cohort of IVF young adults by using gold standard assessment hyperinsulinemic-euglycemic clamp, without any significant differences in fasting glucose or insulin levels compared with naturally conceived young adults.Reference Chen, Wu and Zhao ¹⁰⁷ However, only 14 IVF young adults were studied.

A number of studies suggest that ART may increase the risk of cardiovascular diseases. Celeen et al.Reference Ceelen, van Weissenbruch and Roos ¹⁰⁴ reported that systolic and diastolic blood pressure levels were elevated in IVF children v. children conceived naturally from subfertile couples, after controlling for early life factors and parental characteristics. Sakka et al.Reference Sakka, Loutradis and Kanaka-Gantenbein ¹⁰⁵ also reported that children born by IVF had significantly higher systolic and diastolic blood pressure than controls. Elevations in blood pressure in IVF conceived individuals are not universally detected.Reference Belva, Roelants and De Schepper ¹¹¹

There is evidence that the process of ovarian induction may be a contributing factor to increases in blood pressure since systolic blood pressure and subscapular skinfold thickness were elevated in IVF children v. children conceived by natural IVF (without ovarian stimulation) and subfertile couples who conceived naturally.Reference Seggers and Haadsma ¹¹² Blood pressure was also higher in children born to subfertile v. fertile couples.Reference Pontesilli, Painter and Grooten ¹¹⁰ In a more detailed investigationReference Scherrer, Rimoldi and Rexhaj ¹⁰⁶ ART children displayed systemic and pulmonary vascular dysfunction, that could not be explained by subfertility or ovulation stimulation because vascular function was not altered in children conceived naturally after ovulation stimulation and in siblings of ART children who were conceived naturally.Reference Sakka, Margeli and Loutradis ¹¹³ Further, another group conducted a prospective cohort study and found signs of cardiovascular remodelling in ART fetuses, and ART infants as compared with controls conceived spontaneously.Reference Valenzuela-Alcaraz, Crispi and Bijnens ¹¹⁴ Right ventricular dysfunction has also been detected in children and adolescents conceived by ART.Reference von Arx, Allemann and Sartori ¹¹⁵ Taken together, ART treatment, ovulation induction and subfertility may all contribute to adverse cardiovascular outcome in childhood,Reference Pontesilli, Painter and Grooten ¹¹⁰ and further pre-clinical studies are necessary to resolve some of the discrepancies reported, as these enable better control of the confounding factors.

Few studies have examined the effects of ART on lipid metabolism. Sakka et al. Reference Sakka, Loutradis and Kanaka-Gantenbein ¹⁰⁵ found that IVF children had significantly higher triglycerides, without differences in total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein, uric acid, apolipoprotein-A1, apolipoprotein-B or lipoprotein(a) values.Reference Sakka, Loutradis and Kanaka-Gantenbein ¹⁰⁵ Conversely, a group in New ZealandReference Miles, Hofman and Peek ^{65 ,} Reference Green, Mouat and Miles ⁶⁷ found more favourable lipid profiles in a prepubertal IVF children with higher HDL levels and lower triglyceride levels than in controls. More prospective follow-up studies in ART adults are required to determine if ART treatment alters lipid profiles in the offspring.

ART may also alter thyroid function. Sakka et al.Reference Sakka, Malamitsi-Puchner and Loutradis ¹¹⁶ reported thyroid-stimulating hormone (TSH) levels were significantly higher in 106 IVF children v. 68 naturally conceived children aged 4–14 years. Seven IVF children, but no controls also had subclinical primary hypothyroidism. It is of note that four of these children were born prematurely with low birth weight, although statistically TSH levels was not associated with birth weight and gestational ages. Similarly, subclinical hypothyroidism was observed in 2–4-week-old IVF infants, born at term.Reference Onal, Ercan and Adal ¹¹⁷ A cross-sectional study in China found that the levels of thyroid hormones including T4, FT4 and TSH were significantly increased in singleton IVF v. naturally conceived newborns and children aged 3–10 years old. Notably, the levels of T4 and FT4 of IVF children positively correlated with maternal serum levels of oestradiol during the first trimester of pregnancy.Reference Lv, Meng and Lv ¹¹⁸ Further, no statistical difference was observed between IVF children born from frozen embryo transfer and naturally conceived individuals.Reference Lv, Meng and Lv ¹¹⁸ This suggests that a high oestradiol maternal environment, resulting from ovarian stimulation, may increase the risk of thyroid dysfunction in offspring born following IVF. Further study is necessary.

How does ART increase the risk of adverse outcome?

In humans, it is difficult to separate out the effects of ART procedures themselves with the underlying subfertility, paternal characteristics as well as postnatal environmental exposure. Furthermore, if increases in risk are the result of ART techniques, which of these processes increase risk? Animal models suggest that ART procedures contribute to altered fetal and placental growth and development.Reference Bloise, Feuer and Rinaudo ¹¹⁹ The developmental origins of health and disease hypothesisReference Barker ¹²⁰ proposes that suboptimal periconceptional and perinatal environment can impair fetal and postnatal growth, followed by catch-up growth, predisposes offspring to increased risk of developing hypertension, obesity, type 2 diabetes and coronary heart disease in later life.Reference Eriksson, Forsen and Tuomilehto ^{121 –} Reference Forsen, Eriksson and Tuomilehto ¹²³ Epigenetic is likely to be the reprogramming mechanism, but there may be other contributing factors including transcription changes, oxidative stress and mitochondrial dysfunction, and endoplasmic reticulum stress as reviewed elsewhere.Reference Rinaudo and Wang ¹²⁴

Epigenetics is defined as heritable changes in gene expression without alterations in DNA sequence.Reference Waddington ¹²⁵ Epigenetic modifications, including DNA methylation, histone modifications, micro-RNAs and higher-order packaging of DNA around nucleosomes, regulate the temporal and spatial gene expression patterns and are essential in embryonic, fetal and postnatal development.Reference Le Bouc, Rossignol and Azzi ¹²⁶ DNA methylation is the most widely studied epigenetic mechanism and occurs through the enzymatic addition of a methyl group to the carbon-5 position of the cytosine of the cytosine–phosphate–guanine dinucleotide sequence. The methyl group interferes with the binding of particular transcription factors to DNA and attracts methyl-binding proteins that also regulate transcriptional repression.Reference Swales and Spears ¹²⁷ Hence, gene expression is generally inhibited by DNA methylation, but is activated by DNA demethylation.Reference Dennis ¹²⁸ DNA methylation also contributes to the preservation of chromosomal integrity and the inactivation of X-chromosome.Reference Laprise ¹²⁹ There are two waves of DNA methylation and demethylation during gametogenesis and early preimplantation,Reference Reik and Walter ¹³⁰ thus periconceptual manipulation of oocyte or blastocyst during IVF and ICSI treatment may impair the establishment of the DNA methylation in gametes and/or with the maintenance of DNA methylation within preimplantation embryos.Reference Laprise ¹²⁹

In humans, there is evidence that ART procedures may alter epigenetic modifications during the preimplantation period of development. A high frequency of imprinted methylation errors was observed in ART human preimplantation embryos.Reference White, Denomme and Tekpetey ¹³¹ Altered DNA methylation and/or gene expression of a number of genes in the fetus, cord blood, placenta, neonatal bloodspots and buccal cell have been reported in ART children.Reference Katari, Turan and Bibikova ^{132 –} Reference Gomes, Huber and Ferriani ¹³⁹ Notably, some of these genes whose expression altered by ART have been implicated in imprinting diseases and metabolic disorders such as obesity and type 2 diabetes. Song et al.Reference Song, Ghosh and Mainigi ¹³³ showed ART itself results in significant differences in placental DNA methylation levels by using donor oocyte from fertile young women compared with fertile control groups. Another study suggests DNA methylation levels of 23 genes can explain around 80% of the variance in infant birth weight and six of these are associated with growth phenotypes in human or mouse models.Reference Turan, Ghalwash and Katari ¹⁴⁰ Therefore, altered DNA methylation in ART offspring may contributes to low birth weight which is a marker of impaired fetal growth and adverse long-term health outcomes.

Animal models also support that ART alters epigenetic modifications in preimplantation embryos and offspring, thereby altering embryonic growth, fetal and placental growth, growth trajectory, increases risk of metabolic and cardiovascular diseases later in life, and shortens lifespan.Reference Chen, Wu and Zhao ^{107 ,} Reference Chen, Sun and Huang ^{141 –} Reference Rexhaj, Paoloni-Giacobino and Rimoldi ¹⁴⁹ Metabolic profiling in mice serum and microarray analysis of pancreatic islets and insulin-sensitive tissues (liver, skeletal muscle and adipose tissue) indicated systemic oxidative stress and mitochondrial dysfunction, which is associated with increased expression of thioredoxin-interacting protein (TXNIP) and enrichment for H4 acetylation at the Txnip promoter in blastocysts and adipose tissue in adult mice.Reference Feuer, Liu and Donjacour ¹⁴⁸ As TXNIP plays an important role in regulating peripheral glucose metabolism and integrating cellular nutritional and oxidative states with metabolic response, the data suggest that IVF results in epigenetic and gene expression changes in blastocysts that persist in adulthood.Reference Feuer, Liu and Donjacour ¹⁴⁸ Rexhaj et al.Reference Rexhaj, Paoloni-Giacobino and Rimoldi ¹⁴⁹ reported that ART mice offspring show endothelial dysfunction, increased arterial stiffness and arterial hypertension as well as shortened life span fed with a high-fat diet. Moreover, male ART mice transmit vascular dysfunction to their progeny and the methylation of imprinted genes such as H19 in the aorta is altered in ART mice and their progeny. Further, ART mice display increased DNA methylation of the promoter of the eNOS gene, decreased eNOS expression in the aorta and decreased plasma nitric oxide concentration. Importantly, all these alterations can be normalized by administration of the deacetylase inhibitor butyrate or addition of antioxidant melatonin to culture media, suggesting that altered epigenetic modification by ART causes vascular dysfunction in mice.Reference Rexhaj, Paoloni-Giacobino and Rimoldi ^{149 ,} Reference Rexhaj, Pireva and Paoloni-Giacobino ¹⁵⁰

Placenta plays an important role in fetal development by transporting nutrients and oxygen, adapting morphologically and functionally to adverse environmental stress and minimizing their impact on the fetus.Reference Burton, Fowden and Thornburg ^{151 ,} Reference Sferruzzi-Perri and Camm ¹⁵² Placenta size can predict cardiovascular diseases and insulin resistance.Reference Barker, Bull and Osmond ^{153 ,} Reference Eriksson, Kajantie and Thornburg ¹⁵⁴ ART may also impair placental development and function and thus fetal growth in utero.Reference Chen, Sun and Huang ^{141 ,} Reference Li, Chen and Tang ^{142 ,} Reference de Waal, Vrooman and Fischer ^{155 ,} Reference Bloise, Lin and Liu ¹⁵⁶ Increased placental thickness and placental haematomas as well as pathological findings were reported in ART pregnancies.Reference Lalosevic, Tabs and Krnojelac ^{157 –} Reference Zhang, Zhao and Jiang ¹⁵⁹ IVF impairs placental nutrient transport and metabolism in mice.Reference Chen, Sun and Huang ^{141 ,} Reference Tan, Zhang and Miao ^{144 ,} Reference Bloise, Lin and Liu ¹⁵⁶ Placental weight and placental:fetal weight ratio was significantly higher in ART pregnancies than in naturally conceived pregnancies in humans and mouse models.Reference Bloise, Feuer and Rinaudo ^{119 ,} Reference de Waal, Vrooman and Fischer ^{155 ,} Reference Bloise, Lin and Liu ^{156 ,} Reference Daniel, Schreiber and Geva ^{160 ,} Reference Haavaldsen, Tanbo and Eskild ¹⁶¹ This was associated with reduced methylation levels and altered genomic imprinting and developmental gene expression by ART treatment in the placenta in mice and humans.Reference Katari, Turan and Bibikova ^{132 ,} Reference Turan, Katari and Gerson ^{138 ,} Reference Chen, Sun and Huang ^{141 ,} Reference Li, Chen and Tang ^{142 ,} Reference Tan, Zhang and Miao ^{144 ,} Reference de Waal, Vrooman and Fischer ¹⁵⁵ These altered DNA methylation levels may impair a number of biological processes and functions during IVF placentation, including actin cytoskeleton organization, haematopoiesis, placental growth and vascularization, energy metabolism and nutrient transport.Reference Tan, Zhang and Miao ^{144 ,} Reference Choux, Carmignac and Bruno ¹⁶² Improper adaptive responses of placenta throughout pregnancy may result in adverse outcomes such as abortion, preeclampsia or intra-uterine growth restriction.Reference Choux, Carmignac and Bruno ¹⁶² Although successful placental adaptation leads to normal progress of the pregnancy, the memory of epigenetic adaptation mechanisms established during pregnancy increases the risk of developing metabolic diseases later in life.Reference Choux, Carmignac and Bruno ¹⁶² Taken together, impaired placental function and development due to altered DNA methylation may play a key role underlying the adverse outcomes in ART offspring. However, more studies are needed to examine whether other epigenetic mechanisms such as histone modifications and micro-RNAs during the development of preimplantation embryos are altered by ART.

Future directions to improve health of ART children

The perinatal outcomes in children born after ART have improved over time, mainly as a result of single embryo transfer and frozen thaw embryo transfer.Reference Henningsen, Gissler and Skjaerven ^{163 ,} Reference Hansen and Bower ¹⁶⁴ As already practiced in many countries, single embryo transfer clearly reduces many of the risks associated with ART procedures, including improved perinatal outcomes, without compromizing live birth rates.Reference Hodgson and Wong ^{165 ,} Reference Pandian, Marjoribanks and Ozturk ¹⁶⁶ The impact of hormone stimulation on perinatal and longer-term outcomes is under increasing scrutiny. The available evidence indicates that frozen embryo transfer may improve outcomes for both patients, and especially women with polycystic ovary syndrome and infants.Reference Chen, Shi and Sun ^{24 ,} Reference Evans, Hannan and Edgell ¹⁶⁷ However, large well-controlled trials to determine if freeze-all protocols have benefits to the general infertile population and the later metabolic health in ART children are still needed. Further pre-clinical and large epidemiological studies from around the globe, that collect data to control for as many potential confounders as possible, are needed to compare ART techniques including frozen embryo transfer, stimulated v. unstimulated IVF cycles and the embryo culture lengths for not only the optimal perinatal outcomes, but for long-term health.