| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
The National Research Centre for Growth and Development and Liggins Institute (H.L.M., P.L.H., M.H., D.W., P.D.G., W.S.C.) and Department of Community Health (E.M.R.), University of Auckland, Auckland 1010, New Zealand; and Fertility Associates (J.P.), Auckland 1051, New Zealand
Address all correspondence and requests for reprints to: Associate Professor Wayne Cutfield, The Liggins Institute, University of Auckland, Private Bag 92019, Auckland, New Zealand. E-mail: w.cutfield{at}auckland.ac.nz.
| Abstract |
|---|
|
|
|---|
Methods: We recruited healthy, prepubertal children born at term after singleton pregnancy. The children in the study group were conceived using IVF with fresh embryo transfer, whereas controls were naturally conceived. Anthropometric measurements, bone age, dual-energy x-ray absorptiometry, fasting serum glucose, insulin, lipid profile, IGF-I and -II, and IGF-binding proteins 1, 2, and 3 were performed.
Results: There were 69 IVF children aged 5.9 ± 0.2 yr and 71 control children aged 6.9 yr. IVF children were taller than controls when corrected for parents heights (height SD score of 1.05 ± 0.1 vs. 0.51 ± 0.11, P = 0.001) with higher levels of serum IGF-II (850 ± 24 vs. 773 ± 24 µg/liter, P = 0.03), higher IGF-I to IGF-binding protein 3 ratio (P = 0.04), and a trend toward higher IGF-I (105 ± 4 vs. 92 ± 4 µg/liter, P = 0.06). IVF children had higher high-density lipoprotein (1.67 ± 0.04 mmol/liter vs. 1.53 ± 0.04 mmol/liter, P = 0.02), lower triglycerides (0.65 ± 0.04 mmol/liter vs. 0.78 ± 0.04 mmol/liter, P = 0.02), and a lower total to high-density lipoprotein cholesterol ratio (2.58 vs. 2.86, P = 0.01). There were no differences in body composition.
Conclusions: IVF children are taller with higher IGF-I and IGF-II levels and have a slightly more favorable lipid profile. We speculate that IVF results in epigenetic change through altered methylation of genes involved in growth and metabolism. IVF programs should consider long-term longitudinal follow-up of IVF offspring.
| Introduction |
|---|
|
|
|---|
Beckwith-Wiedemann and Angelman syndromes are due to abnormal methylation patterns of imprinted genes. Methylation of cytosine in islands of cytosine paired with guanine switches off gene transcription (6). Imprinted genes are methylated according to whether they are inherited from the mother or the father. There are over 50 known imprinted genes in the human, and many of these genes are involved in cellular proliferation and growth. Methylation of nonimprinted genes can also occur, altering gene expression. Culture of mammalian embryos results in a change in the expression of several imprinted genes (1, 7). Furthermore, nutritional manipulation early in fetal life has been shown to lead to reduced methylation and overexpression of nonimprinted genes such as the glucocorticoid receptor (8).
Phenotypic features seen in children with Beckwith-Wiedemann syndrome are variable, with the full syndrome the extreme end of a continuum (9). We hypothesize that subtle differences may exist in the previously under-investigated IVF population manifested as measurable changes in phenotype and hormonal profile in mid-childhood.
| Subjects and Methods |
|---|
|
|
|---|
We recruited healthy prepubertal children between the ages of 4 and 10 yr, born at term (>36 wk gestation) after singleton pregnancy. Subjects were recruited into IVF and control groups. IVF subjects living in Auckland were recruited from Fertility Associates, the largest IVF provider in New Zealand. All IVF children were conceived using IVF with fresh embryo transfer between January 1995 and December 2000. All IVF fathers were documented as the sperm donors for their IVF child. Exclusion criteria for all subjects included multiple pregnancy identified on early scan, a known medical syndrome, chronic illness, or receiving regular medications. Although breastfeeding rates were not collected, virtually all subjects are likely to have been breastfed. Rates of breastfeeding in New Zealand are high at 92% of Caucasian infants (the ethnicity of almost all subjects studied), with the highest rates seen in mothers in the 30- to 40-yr age group (which matches the maternal age range of the subjects we studied) (from the "Report on maternity: maternal and newborn information 2002"; www.nzhis.govt.nz/publications/maternity/report.pdf).
Commercial embryo culture was sourced from Medicult (52 subjects) or Scandinavian IVF (17 subjects); the protein supplement was human serum albumin. Embryos were transferred on d 2 with two to six cells per embryo. Fertility Associates annual pregnancy rates, corrected for percentage of eggs retrieved progressively, increased through the study period from 15 to 40%, placing them within the top five New Zealand and Australian centers (10).
Controls were matched as closely as possible for socioeconomic background using school decile and residential address, age, sex, and ethnicity. Body mass index was recorded for both groups of parents, and all children were asked to complete a 3-d dietary diary before the study for later comparison. In addition to peers recruited by the IVF subjects, a small number of naturally conceived siblings of IVF children were also included in the control group. Ethics approval was provided by the Auckland Ethics Committee, and written informed consent was obtained for all subjects.
Methods
Demographic information was collected including mode of conception and birth data. Supplementary data were collected for children conceived using IVF, including protocol for ovulation induction, mode of conception [conventional IVF or intracytoplasmic sperm injection (ICSI), defined as the direct injection of sperm into oocytes], and embryo culture medium. Parental height and weight were recorded and midparental height calculated and corrected to a SD score (SDS) (11). Standing height was measured using a Harpenden stadiometer and head circumference recorded as an SDS (12). Weight and body composition were assessed using both body mass index (BMI) [weight (kg)/height (m)2] and whole-body dual-energy x-ray absorptiometry (DEXA) (Lunar Prodigy 2000, General Electric, Madison, WI). Bone age was assessed by a single blinded pediatric radiologist using standards established by Greulich and Pyle (13).
A fasting blood sample was obtained for assessment of metabolic and growth factors. Plasma glucose was measured using an automated random-access analyzer (model 911; Hitachi, Tokyo, Japan) with an interassay coefficient of variation (CV) of 1.2%. Insulin levels were determined by enzyme immunoassay (IMX microparticle assay; Abbott Laboratories, Chicago, IL) with an interassay CV of less than 5%. Total cholesterol, high-density lipoprotein (HDL)-cholesterol, and triglyceride were measured using a Roche (Indianapolis, IN) modular analyzer. Low-density lipoprotein (LDL)-cholesterol was calculated using the Friedwald formula. IGF-I was measured using a Diagnostic Products Corp. (Los Angeles, CA) Immulite analyzer (interassay CV 9.3%). Commercially available ELISAs (Diagnostic Systems Laboratories, Webster, TX) were used to evaluate plasma IGF-II (DSL-10-9100, intraassay CV 1.7%, interassay CV 6.2%), IGF-binding protein (IGFBP)-1 (DSL-10-7800, intraassay CV 1.7% interassay CV 6.2%), IGFBP-2 (DSL-10-7100, intraassay CV 1.5%, interassay CV 3.4%), IGFBP-3 (DSL-10-6600, intraassay CV 7.3%, interassay CV 8.2%) and GH-binding protein (DSL-10-48100, intraassay CV 4.8%, interassay CV 5.1%).
Statistical analysis
Linear regression analyses were used to investigate the differences in baseline characteristics and anthropometric, endocrine, and metabolic measures between the study groups of children. Gestation, birth weight, midparental height, maternal age, bone age, age, and sex were included in models of anthropometric data in the calculation of adjusted height data. Gestation, birth weight, midparental height, maternal age, standing height, weight, age, and sex were included in models of lipid and insulin data and age, sex, BMI, DEXA percent fat, and height for the remaining endocrine data. Means and SEM are reported for baseline data, means and SE adjusted for other variables in the model are reported for other measures. SAS version 9.1 was used for all statistical analyses. A P value of < 0.05 was used to define significance.
| Results |
|---|
|
|
|---|
Birth and parental clinical characteristics of the two groups are summarized in Table 1
. Parental anthropometry was comparable between the two groups. Mothers of the IVF children were older than mothers of control children; however, maternal age did not influence any of the outcome parameters measured. There was no difference in midparental heights between the groups; in particular, there was no difference in maternal height SDS values between IVF and control groups (0.55 vs. 0.4, P = 0.45). All families in the study were from higher socioeconomic groups. Dietary diaries were similar for IVF children and controls, and over 50% of control children were firstborn or singleton children. Although the IVF group had slightly lower birth weight, only two children fulfilled the criteria for being small for gestational age (SGA) (14). Predictably, IVF children were more likely to be firstborn; however, birth order did not influence any of the outcome measures, including height SDS, bone age, and BMI, or any of the serum hormone measurements.
|
0), whereas 44% of control subjects had a height – midparental height SDS of
0 (P < 0.03). Similar increases in sitting and standing heights of IVF children indicate that this increase in stature is proportionate. There were no other sex differences between IVF and control groups found for any other variables examined. The corrected BMI in the IVF group was lower, but there was no difference in percent fat assessed by DEXA.
|
|
|
IVF children had a more favorable lipid profile with higher HDL (P = 0.02), lower triglyceride level (P = 0.02), and lower total to HDL-cholesterol ratio (P = 0.01). There was no difference in fasting insulin or glucose between the two groups.
There were no differences between children conceived by conventional IVF and those conceived by ICSI, nor did ovarian stimulation protocol or the type of culture media influence any of the outcome measures.
| Discussion |
|---|
|
|
|---|
In our IVF cohort, we observed higher serum IGF-II and increasingly higher serum IGF-I with age, without any differences between the ICSI and non-ICSI IVF children. Conversely, Kai et al. (15) found lower serum IGF-I levels in ICSI singleton males and non-ICSI IVF females at 3 months of age; however, these differences were not present in their cohort of 5-yr-old children. The slightly lower birth weight and taller childhood stature in our IVF group suggests a postnatal rather than prenatal growth effect. In a cross-sectional study such as ours, it is difficult to ascertain whether the higher serum IGF-II or pattern of higher IGF-I with increasing age played a role in promoting taller stature in the IVF group. More precise clarification of the role of IGF-I or IGF-II as potentially causative in growth promotion in IVF would require a longitudinal study with detailed auxology and sequential serum hormone measurements.
Although all children studied had lipid parameters within the normal range, the IVF children displayed a slightly more favorable lipid profile with higher levels of HDL and lower triglyceride levels. The long-term significance of subtle improvement in lipid profile in IVF children is unclear but warrants further evaluation of adults born after IVF. There are many frequently cited examples of maladaptive responses to an adverse fetal environment; however, the presence of taller stature in IVF children may be the first described example of a better outcome as a result of events during fetal life (19, 20, 21).
There are several possible etiological mechanisms that could explain taller stature with hormonal and biochemical differences in IVF children, which include programmed endocrine changes due to the IVF process, selection of the best quality embryos for implantation, or selection bias in the control population. The alterations in growth and lipid profile suggest that manipulation during the periconceptual period can result in persistent later differences. Observations in animal studies lend support to this concept with both nutritional manipulation at the time of conception or in vitro culture affecting later outcome (1, 7, 22, 23, 24). Thus, we propose that in vitro manipulation has resulted in persistent and possibly permanent metabolic alterations in the IVF children we studied.
Although the etiology of the growth and metabolic differences in IVF children is yet to be determined, it is possible that the IVF process leads to epigenetic alteration of imprinted and nonimprinted genes. Beckwith-Wiedemann syndrome has been reported to occur up to nine times more commonly after IVF conception than in the general population (5, 25, 26, 27). In a metaanalysis, 23 of the 24 cases of Beckwith-Wiedemann syndrome after IVF were found to be due to hypomethylation of the Beckwith-Wiedemann syndrome imprinted region KvDMR1 (1). Given that our cohort was taller, it is possible that these children have more subtle alterations in DNA methylation patterns in imprinted genes associated with growth such as IGF-II at H19 or KvDMR1. Mild phenotypic features, such as taller stature in childhood with normal birth weight, have been described in relatives of children with Beckwith-Wiedemann syndrome (9).
A potential risk to drawing conclusions in any study of this kind is the applicability of the control group. Potential confounders influencing growth are genetic potential (or parental heights), socioeconomic background, nutritional status, and pubertal stage. We have tried to address this in our study design by attempting to ensure the two groups are similar in all respects. IVF children in New Zealand tend to be born to parents of higher socioeconomic background. To ensure appropriate matching, control subjects were of the same school decile code and similar residential address to remove any influence of socioeconomic status. Parental heights between the groups were similar as was nutritional status (based on body composition). All children were confirmed to be prepubertal, eliminating the effects of sex steroids on stature and body composition. We are aware that the IVF group represents a cohort of much wanted children and that parental care may reflect this; however, we are not aware of any studies suggesting that stature is affected by increased parental attention in families of higher socioeconomic status. The number of single-parent families was the same in both groups (one per group). In those firstborn, a small increase in adult height in women of 2 mm and in men 7 mm has been reported (28). These differences in height may be due to other stronger influences such as differences in parental heights and environmental factors. The authors stated that birth order was powerfully associated with all other environmental ratings and encapsulated much of their influence. Conversely, we found that birth order did not influence any of the parameters measured within either the IVF or control groups. Dietary records revealed similar eating patterns in both IVF and control children. Collectively, the extensive exclusion criteria used and lack of effect with the few different characteristics between the groups suggests that our findings do not reflect selection bias. Although every effort has been made to match IVF and control subjects, it remains conceivable that unrecognized potential confounders may still exist that could have influenced postnatal growth. The ideal study design would be to randomize couples to IVF or naturally conceived pregnancies and study the growth patterns of the offspring. However, this would be impossible to achieve; thus, all studies of IVF children may be confounded by unrecognized factors. Furthermore, the availability of cross-sectional rather than longitudinal growth data from birth prevents us from fully understanding the early growth pattern of IVF children.
In summary, we have shown that prepubertal children born after IVF are taller when adjusted for parents heights with higher serum IGF-II, IGF-I, and IGF-I to IGFBP3 ratio and more favorable lipid profile than naturally conceived control children. We propose that the differences in stature, growth factors, and lipid metabolism observed in our study may be due to subtle epigenetic alteration of imprinted genes or other genes that undergo epigenetic modification that are involved in growth and development. Despite the birth of 2 million IVF children and substantial data from birth registries, little is known about IVF children beyond the risk of congenital abnormalities. Comparisons of our data with other groups using different IVF technology will be important to define whether the changes we observed are modified with different methodologies. This study demonstrates the need to monitor the health of IVF offspring through childhood into adulthood.
| Acknowledgments |
|---|
| Footnotes |
|---|
H.L.M., P.L.H., M.H., D.W., E.M.R., P.D.G., and W.S.C. have nothing to declare. J.P. is employed as scientific director for Fertility Associates. None of the authors have a competing interest to declare.
First Published Online June 12, 2007
Abbreviations: BMI, Body mass index; CV, coefficient of variation; DEXA, dual-energy x-ray absorptiometry; HDL, high-density lipoprotein; ICSI, intracytoplasmic sperm injection; IGFBP, IGF-binding protein; IVF, in vitro fertilization; LDL, low-density lipoprotein; SDS, SD score.
Received November 9, 2006.
Accepted June 6, 2007.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
D. J. Amor and J. Halliday A review of known imprinting syndromes and their association with assisted reproduction technologies Hum. Reprod., December 1, 2008; 23(12): 2826 - 2834. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |