Secondary Logo

Journal Logo

Original Research

Maternal Global Methylation Status and Risk of Congenital Heart Diseases

van Driel, Lydi M. J. W. MSc; de Jonge, Robert PhD; Helbing, Willem A. MD, PhD; van Zelst, Bertrand D.; Ottenkamp, Jaap MD, PhD; Steegers, Eric A. P. MD, PhD; Steegers-Theunissen, Rėgine P. M. MD, PhD

Author Information
doi: 10.1097/AOG.0b013e31817dd058
  • Free

Each year, congenital heart diseases (CHDs) account for more than one million affected newborns worldwide.1 More than 85% of CHDs are thought to result from complex interactions involving genetic susceptibilities and environmental exposures.2 Because cardiogenesis takes place in the first weeks of pregnancy, maternal metabolic derangements during the periconception period and early pregnancy can detrimentally affect the embryonic processes involved.3 We and others have shown that maternal hyperhomocysteinemia is associated with an increased risk of CHD offspring, especially with outflow tract defects.4–7 In line with this observation, periconception use of multivitamins containing folic acid decreases the homocysteine concentration and reduces the risk of CHD offspring.8 Folic acid seems to be the key factor, because the use of folate antagonists by the mother increases risk of CHD offspring, which can be prevented by a folic acid supplement.9 So far, it is not clear whether homocysteine itself is teratogenic or just an epiphenomenon of a deranged one-carbon metabolism.

Hyperhomocysteinemia also leads to the accumulation of S-adenosylhomocysteine, a potent inhibitor of methyltransferases using S-adenosylmethionine as the main methyl donor.10,11 These methyltransferases are important in DNA methylation, which is the best known epigenetic mechanism regulating gene expression without changing DNA sequences.12 Hypomethylation of DNA is believed to initiate chromosome instability and to alter gene expression, cell differentiation, and apoptosis during embryogenesis.13 It thereby affects many biologic pathways at different times in several tissues by gene silencing and also by chromosome segregation. As a consequence, major and subtle genetic aberrations can be differentially expressed resulting in various CHDs. A recent study identified total homocysteine, S-adenosylhomocysteine, and methionine as the most important predictive biomarkers for the included CHDs.14 However, it is not clear whether this accounts for isolated CHDs or perhaps also for nonisolated CHDs.

In the present study, we conducted a case–control triad study in the western part of the Netherlands to test the hypothesis that maternal hypomethylation reflected in S-adenosylmethionine/S-adenosylhomocysteine ratio, S-adenosylmethionine and S-adenosylhomocysteine concentrations rather than hyperhomocysteinemia is a risk factor for CHD offspring. We investigated the CHD cases as a total group in comparison with the controls, and in subgroups.


The subjects are enrolled in the Hartafwijkingen, vasculaire status en nutriënten (HAVEN) study, of which the name is an acronym for the Dutch name of the ongoing study designed to investigate determinants in the pathogenesis and prevention of CHDs. It is a case–control triad study that has been conducted from June 2003 at the Department of Obstetrics and Gynecology, Division of Obstetrics and Prenatal Medicine at the Erasmus University Medical Center in Rotterdam. Case children and both parents are enrolled in collaboration with the Departments of Pediatric Cardiology of four collaborating university medical centers in the Netherlands. Control children, together with their parents, are recruited in collaboration with the child health centers of “Thuiszorg Nieuwe Waterweg Noord” in the western part of the Netherlands. These child health centers are part of the Dutch Health Care system, where all newborns are regularly checked on for health, growth, and development by physicians specialized in child health care. The domain population comprised both case children and control children living in the western part of the Netherlands. All children were between 11 and 18 months of age, and there was no familial relationship between cases and controls. The materials and methods of this study have been described before and are summarized hereafter.7

Cases are included if they had a CHD diagnosed by a pediatric cardiologist and confirmed by echocardiography and/or cardiac catheterization and/or surgery after birth. The CHD phenotypes included (n=231) comprised tetralogy of Fallot (n=35), transposition of the great arteries (n=31), atrioventricular septal defect (n=21), perimembranous ventricular septal defect (n=66), coarctation of the aorta (n=26), aortic valve stenosis (n=4), pulmonary valve stenosis (n=39), and hypoplastic left heart syndrome (n=9). These phenotypes were selected because experimental and epidemiologic studies showed that hyperhomocysteinemia and related gene–environment interactions are involved in their causation.4,8,14 The total case group consisted of 180 isolated and 51 nonisolated defects. The 51 nonisolated CHDs consisted of 20 nonsyndromic cases, 19 with Down syndrome, and 12 with other syndromes: 22q11 deletion syndrome (n=5), insertion 1>3 (n=1), Noonan (n=1), Turner (n=1), Alagille (n=1), Saethre-Chotzen (n=1), CHARGE association (n=1), and Beckwith-Wiedemann syndrome (n=1).

Children were eligible as controls if they did not have a major congenital malformation or chromosomal abnormality according to the medical records and regular health checks by the physician at the child health centers. Cases and controls were excluded if the index pregnancy was a plural birth, if they were not familiar with the Dutch language in writing and reading and if the mother was pregnant, breast-feeding, or reported to have a different diet at the study moment than in the periconception period. Each group had corresponding proportions of female and male children. The study protocol was approved by the Central Committee on Research involving Human Subjects (Centrale Commissie Mensgebonden Onderzoek) and the institutional review boards (medical ethics committees) of all participating hospitals. All participants gave their written informed consent.

At a fixed study moment of approximately 16 months after the index pregnancy, mothers filled out a general questionnaire at home, which was checked for completeness and inconsistency at the hospital visit. Information was obtained about educational level, ethnicity, and both current use of alcohol, tobacco, medication, and B vitamin supplements and use of these products during the periconception period. Current use was defined as any use in the previous 4 weeks. The periconception period was defined as 4 weeks before until 8 weeks after the conception. Educational level was classified according to the definitions of Statistics Netherlands.15 Mothers were classified as Dutch Natives, European Others and Non Europeans. Maternal height (anthropometric rod; seca GmbH, Hamburg, Germany) was measured up to 0.1-cm accuracy and weight (weighing scale; seca) up to 0.5-kg accuracy. Body mass index (BMI) was defined as weight divided by the square of length.

Ethylenediaminetetraacetic acid blood was taken for determination of total homocysteine, S-adenosylmethionine, and S-adenosylhomocysteine. After withdrawal, blood was kept on ice and centrifuged at 4°C within 2 hours. Plasma aliquots were stored at –80°C until analysis. Total homocysteine was determined using isotope-dilution liquid chromatography tandem mass spectrometry (LC-MS/MS; Waters Acquity UPLC Premier XE, Milford, MA) by a method adapted from Ducros et al.16 For chromatographic separation we used a Waters Symmetry C8 column (2.1×100 mm, reference WAT 058961, Waters, Etten-Leur, the Netherlands) with a precolumn (Waters, reference 205000343). The column was eluted at 0.25 mL/min, and no splitter was used. Calibration was performed with aqueous standards because the results were similar to those of plasma-based standards. S-adenosylmethionine and S-adenosylhomocysteine were also determined using LC-MS/MS by a method adapted from Gellekink et al.17 In short, nonacidified ethylenediaminetetraacetic acid plasma was stored at –80°C, and 200 microliters of plasma was used for sample clean-up. Samples (10 microliters) were injected on a 50×2.1–mm Atlantis C18 column (Waters) and eluted in a gradient of methanol in aqueous acetic acid (0.1%). The retention times were 0.6 minutes (S-adenosylmethionine) and 1.4 minutes (S-adenosylhomocysteine). Standards were dissolved in 1 mmol/L hydrochloric acid; pool sera were S-adenosylmethionine and S-adenosylhomocysteine depleted by solid phase extraction and spiked with the calibrator. Calibration curves for S-adenosylmethionine and S-adenosylhomocysteine were linear until 500 nmol/L.

Time after pregnancy, maternal age, and BMI are presented as medians and compared between the different case groups and controls using the Mann-Whitney U test. Differences in frequencies were tested using the χ2 test. The biochemical characteristics are presented as medians and ranges, and because the distributions were positively skewed, all biomarker data were log-transformed (natural log) before analysis. Normality of the transformed data was verified for all case groups and the control group by using the Kolmogorov-Smirnov test. Differences between the case groups and controls were tested with a general linear model (univariate analysis of variance). Current B vitamin supplement use of the mother was considered a confounding factor and was therefore included in the model. Probability values of P<.05 were considered statistically significant. Moreover, all comparisons were adjusted for multiple testing by the method of Bonferroni. All analyses were performed using SPSS 15.0 for Windows software (SPSS Inc., Chicago, IL). The results on maternal homocysteine levels in association with congenital heart defects have already been published within our study, although in a much smaller sample size and only in the total case group.7


The general characteristics of the total case and control group were not significantly different (Table 1). The general characteristics of the subgroups of cases were also determined (data not shown). Mothers of nonsyndromic cases used significantly more medication in the periconception period than mothers of controls (cases: number of users 8 (40%); controls: number of users 66 (21%); P<.05). However, after adjusting for multiple testing, it became nonsignificant. Mothers of children with Down syndrome were significantly older (median 35.0 years, range 31.3–41.2 years) compared with control mothers (median 32.6 years, range 28.5–34.9 years; P<.01). All other general characteristics of the subgroups of cases compared with the control group were not significantly different. We also examined maternal medical diseases in the periconception period associated with CHDs, such as diabetes and epilepsy. Both diseases were present in two case mothers and four control mothers. Also, morbidly obese BMI (40 kg/m2 or more) was considered a potential confounder; there were three case mothers and five control mothers with a morbid BMI. These were very small numbers and not different between cases and controls (data not shown). After adjusting for multiple testing, all variables became nonsignificant, except for the variable maternal age in the Down Syndrome group compared with the controls.

Table 1
Table 1:
General Characteristics of Mothers of a Congenital Heart Disease Child (Cases) and Nonmalformed Child (Controls)

In Tables 2 and 3, comparisons of plasma biomarker concentrations are presented between the different case subgroups and the controls. Table 2 reveals a significant difference in total homocysteine concentration between the total case group and the nonisolated cases compared with the controls. In Table 3, the nonisolated cases are divided into three different subgroups: Nonsyndromic, Down Syndrome, and Other Syndromes. The Other Syndrome group presented significantly higher levels of total homocysteine than the control group. Moreover, mothers of children with Down syndrome presented not only significantly higher levels of total homocysteine but also of S-adenosylhomocysteine and a lower S-adenosylmethionine/S-adenosylhomocysteine ratio than the control group. All comparisons remained significant after adjusting for multiple testing, except for the comparison of total homocysteine concentrations between the total case group and the Down Syndrome group compared with the controls.

Table 2
Table 2:
Plasma Biomarkers Concentrations of Cases Compared With Controls
Table 3
Table 3:
Plasma Biomarker Concentrations of Nonisolated Cases Compared With Controls


In this study we confirmed our previous finding that maternal total homocysteine concentrations were significantly higher in the total group of CHD offspring than in the control group. The nonisolated CHD cases, however, and more specifically the subgroups of Down Syndrome and Other Syndromes seem to be entirely responsible for this finding. Of most interest is the new observation that a maternal status of hypomethylation, reflected by a high total homocysteine and S-adenosylhomocysteine level and low S-adenosylmethionine/S-adenosylhomocysteine ratio, is significantly associated with an increased risk of having a child with CHD and Down syndrome.

So far, only Hobbs et al14 associated maternal hypomethylation with CHD risk. The children in that study had a nonsyndromic CHD comprising a septal, conotruncal, or right- or left-sided obstructive defect. The most important difference with our results is that we found this profile of hypomethylation in the syndromic case group only, and especially in mothers of a child with Down syndrome and CHD, but not in the nonsyndromic case group.

There could be several explanations for these different results. Firstly, Hobbs et al14 included mothers at various points within a period of 0.1 to 52.2 months after the index pregnancy, whereas we included mothers at a fixed study point at around 15 months within a much shorter period of 11 to 18 months after the index pregnancy. The wide period of sampling and data collection might have introduced bias by covariates that affected the total homocysteine levels and were not reported, such as breast-feeding and age.18 If more case mothers than control mothers were breast-feeding at the time of sampling, the total homocysteine levels could have been higher and the S-adenosylmethionine/S-adenosylhomocysteine ratio lower, thereby producing an overestimation of the associated risk estimates. This might have been the case, because fewer case mothers participated more than 6 months after the index pregnancy, whereas in our study all breast-feeding mothers were excluded. Another explanation can be that the CHD phenotypes and the omitted numbers in the case groups are different, because they were not reported.

Nevertheless, our main finding was the altered S-adenosylmethionine/S-adenosylhomocysteine ratio as a marker of hypomethylation in mothers of a child with CHD and Down syndrome. A higher S-adenosylhomocysteine concentration and a lower S-adenosylmethionine/S-adenosylhomocysteine ratio are indicators of a lower methylation capacity.19 Hypomethylation has been associated with chromosome instability.20–22 Moreover, folate deficiency also reduces synthesis of S-adenosylmethionine, leading to DNA hypomethylation.23 Only one case report described a child with Down syndrome and neural tube defect.24 The mother of this child had a threefold higher S-adenosylhomocysteine level; this led to a markedly reduced S-adenosylmethionine/S-adenosylhomocysteine ratio and to a significant hypomethylation of lymphocytes. Moreover, both mother and child were homozygous for the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism, which is associated with mildly increased total homocysteine and decreased folate levels, particularly in situations of deprived folate status. In other studies, this MTHFR polymorphism has also been associated with Down syndrome.25,26 In an additional analysis we therefore evaluated the methylation biomarkers together with the maternal MTHFR C677T genotypes. In the total case group, mothers with the homozygous variant not only had lower S-adenosylmethionine/S-adenosylhomocysteine ratios (data not shown and not significant), they also showed significantly higher total homocysteine concentrations (median [range]: 14.8 [4.0–43.8] micromols per liter; n=19) than the heterozygous case group (10.4 [5.1–26.1] micromols per liter; n=95) and than wild-type carriers (10.1 [6.5–31.2] micromols per liter; n=97; P=.007, Kruskal Wallis Test). The group of mothers of a child with Down syndrome and a CHD was small, and only one mother carried the homozygous variant. She had a lower median S-adenosylmethionine (homozygous 74.8 nmol/L compared with heterozygous 83.1 nmol/L compared with wild-type 77.4 nmol/L), a lower S-adenosylmethionine/S-adenosylhomocysteine ratio (homozygous 4.5 compared with heterozygous 6.1 compared with wild-type 5.6), a higher S-adenosylhomocysteine (homozygous 16.5 nmol/L compared with heterozygous 16.1 nmol/L compared with wild-type 14.5 nmol/L), and also a higher total homocysteine level (homozygous 43.8 micromols per liter compared with heterozygous 11.3 micromols per liter compared with wild-type 11.1 micromols per liter).

These results are substantiated by the literature and suggest an association between the hypomethylation status of the mother and the risk of having a child with Down syndrome and CHD. Because of our definition of the case group, we were unable to distinguish between the risk of Down syndrome without a CHD and risk of a CHD in children with Down syndrome. Therefore, further studies are needed to unravel the role of hypomethylation in these separate groups.

Low maternal folate status, MTHFR 677TT carriership, and hyperhomocysteinemia might result in a decreased S-adenosylmethionine/S-adenosylhomocysteine ratio, thereby inducing general hypomethylation and increasing the risk of offspring with Down syndrome and CHD. The absence of this association in the other CHD phenotypes suggests that maternal hypomethylation may be less important. As shown by Becker et al,27 the folate status is a strong determinant of total homocysteine levels, but not of the S-adenosylmethionine/S-adenosylhomocysteine ratio. This may suggest that the treatment of hyperhomocysteinemia by folic acid does not reduce the risk of having offspring with Down syndrome and CHD, because it does not affect the hypomethylation status. On the other hand, Ingrosso et al28 showed that hypomethylation of lymphocytes could be restored by folic acid treatment. Although several studies suggest that periconception intake of folic acid may reduce the risk of a broad range of CHDs, the results are not conclusive.8,29 Therefore, we suggest that, due to differences in cardiac genes sensitive to folate shortage, hyperhomocysteinemia or hypomethylation, folic acid treatment may not reduce the risk of all CHD phenotypes.

Some methodologic limitations have to be considered. Despite our efforts to minimize the potential biases inherent in a case–control study focusing on congenital malformations, our results do not allow concluding that the alterations in maternal biomarkers were also present during the embryogenesis of the heart. However, our findings are substantiated by the fact that biomarkers of the total homocysteine pathway are rather stable in the periconception period and over a period of 1–2 years.30 Moreover, the subgroup biomarker analyses not showing a difference with controls are likely underpowered, which means that there could be a difference.

A strength of our study is the standardized study moment of around 15 months after the index pregnancy. It does not significantly interfere with the maternal physiology, metabolism, and endocrinology, and it diminishes the chance of misclassifying cases and controls, because most malformations are completely diagnosed in the first year of life. Moreover, compared with other studies, our study moment is relatively soon after pregnancy, which is important to minimize recall bias. Another strength is that we selected CHD phenotypes on the basis of evidence from experimental and epidemiologic studies showing associations with maternal hyperhomocysteinemia and related gene–environment interactions.4,8,14 This is reinforced by the hypothesis underlying the HAVEN study, which is that most human teratogenic exposures are derived from maternal endogenous metabolic or endocrine derangements or exogenous harmful exposures, which adversely affect the embryogenesis of her child.3 This adverse effect is dependent on the time window and sensitivity of the exposures. By analyzing the CHD phenotypes separately and after pooling, we produced comparable results that substantiated the homogeneity of the CHD group.

In conclusion, our findings suggest an association between maternal hypomethylation status reflected in high total homocysteine and S-adenosylhomocysteine levels and a low S-adenosylmethionine/S-adenosylhomocysteine ratio and offspring with Down syndrome and CHD. Because this subgroup was very small, this needs to be confirmed in a much larger population, focusing on mothers whose children have Down syndrome with or without a CHD. It is important to extend this knowledge with new risk factors for CHD, because this is the only manner to improve the preconception counseling and care of future women who want to become pregnant.


1. March of Dimes Birth Defects Foundation. Global report on birth defects: the hidden toll of dying and disabled children. White Plains (NY): March of Dimes Birth Defects Foundation;2006. p. 28.
2. Botto LD, Correa A. Decreasing the burden of congenital heart anomalies: an epidemiologic evaluation of risk factors and survival. Prog Pediatr Cardiol 2003;18:111–21.
3. Steegers-Theunissen RP, Steegers EA. Nutrient-gene interactions in early pregnancy: a vascular hypothesis. Eur J Obstet Gynecol Reprod Biol 2003;106:115–7.
4. Boot MJ, Steegers-Theunissen RP, Poelmann RE, van Iperen L, Gittenberger-de Groot AC. Cardiac outflow tract malformations in chick embryos exposed to homocysteine. Cardiovasc Res 2004;64:365–73.
5. Hobbs CA, Malik S, Zhao W, James SJ, Melnyk S, Cleves MA. Maternal homocysteine and congenital heart defects. J Am Coll Cardiol 2006;47:683–5.
6. Kapusta L, Haagmans ML, Steegers EA, Cuypers MH, Blom HJ, Eskes TK. Congenital heart defects and maternal derangement of homocysteine metabolism. J Pediatr 1999;135:773–4.
7. Verkleij-Hagoort AC, Verlinde M, Ursem NT, Lindemans J, Helbing WA, Ottenkamp J, et al. Maternal hyperhomocysteinaemia is a risk factor for congenital heart disease. BJOG 2006;113:1412–8.
8. Botto LD, Mulinare J, Erickson JD. Do multivitamin or folic acid supplements reduce the risk for congenital heart defects? Evidence and gaps. Am J Med Genet A 2003;121:95–101.
9. Hernandez-Diaz S, Werler MM, Walker AM, Mitchell AA. Folic acid antagonists during pregnancy and the risk of birth defects. N Engl J Med 2000;343:1608–14.
10. Yi P, Melnyk S, Pogribna M, Pogribny IP, Hine RJ, James SJ. Increase in plasma homocysteine associated with parallel increases in plasma S-adenosylhomocysteine and lymphocyte DNA hypomethylation. J Biol Chem 2000;275:29318–23.
11. James SJ, Melnyk S, Pogribna M, Pogribny IP, Caudill MA. Elevation in S-adenosylhomocysteine and DNA hypomethylation: potential epigenetic mechanism for homocysteine-related pathology. J Nutr 2002;132:2361S–6S.
12. Perna AF, Ingrosso D, De Santo NG, Galletti P, Zappia V. Mechanism of erythrocyte accumulation of methylation inhibitor S-adenosylhomocysteine in uremia. Kidney Int 1995;47:247–53.
13. Ehrlich M. Expression of various genes is controlled by DNA methylation during mammalian development. J Cell Biochem 2003;88:899–910.
14. Hobbs CA, Cleves MA, Melnyk S, Zhao W, James SJ. Congenital heart defects and abnormal maternal biomarkers of methionine and homocysteine metabolism. Am J Clin Nutr 2005;81:147–53.
15. Statistics Netherlands. Classification of educational level. Available at: Retrieved June 17, 2008. Voorburg/Heerlen, the Netherlands; Netherlands.
16. Ducros V, Belva-Besnet H, Casetta B, Favier A. A robust liquid chromatography tandem mass spectrometry method for total plasma homocysteine determination in clinical practice. Clin Chem Lab Med 2006;44:987–90.
17. Gellekink H, van Oppenraaij-Emmerzaal D, van Rooij A, Struys EA, den Heijer M, Blom HJ. Stable-isotope dilution liquid chromatography-electrospray injection tandem mass spectrometry method for fast, selective measurement of S-adenosylmethionine and S-adenosylhomocysteine in plasma. Clin Chem 2005;51:1487–92.
18. Milman N, Byg KE, Hvas AM, Bergholt T, Eriksen L. Erythrocyte folate, plasma folate and plasma homocysteine during normal pregnancy and postpartum: a longitudinal study comprising 404 Danish women. Eur J Haematol 2006;76:200–5.
19. Caudill MA, Wang JC, Melnyk S, Pogribny IP, Jernigan S, Collins MD, et al. Intracellular S-adenosylhomocysteine concentrations predict global DNA hypomethylation in tissues of methyl-deficient cystathionine beta-synthase heterozygous mice. J Nutr 2001;131:2811–8.
20. Harrison JJ, Anisowicz A, Gadi IK, Raffeld M, Sager R. Azacytidine-induced tumorigenesis of CHEF/18 cells: correlated DNA methylation and chromosome changes. Proc Natl Acad Sci U S A 1983;80:6606–10.
21. Leyton C, Mergudich D, de la Torre C, Sans J. Impaired chromosome segregation in plant anaphase after moderate hypomethylation of DNA. Cell Prolif 1995;28:481–96.
22. Almeida A, Kokalj-Vokac N, Lefrancois D, Viegas-Pequignot E, Jeanpierre M, Dutrillaux B, et al. Hypomethylation of classical satellite DNA and chromosome instability in lymphoblastoid cell lines. Hum Genet 1993;91:538–46.
23. Das PM, Singal R. DNA methylation and cancer. J Clin Oncol 2004;22:4632–42.
24. Al-Gazali LI, Padmanabhan R, Melnyk S, Yi P, Pogribny IP, Pogribna M, et al. Abnormal folate metabolism and genetic polymorphism of the folate pathway in a child with Down syndrome and neural tube defect. Am J Med Genet 2001;103:128–32.
25. Hobbs CA, Sherman SL, Yi P, Hopkins SE, Torfs CP, Hine RJ, et al. Polymorphisms in genes involved in folate metabolism as maternal risk factors for Down syndrome. Am J Hum Genet 2000;67:623–30.
26. James SJ, Pogribna M, Pogribny IP, Melnyk S, Hine RJ, Gibson JB, et al. Abnormal folate metabolism and mutation in the methylenetetrahydrofolate reductase gene may be maternal risk factors for Down syndrome. Am J Clin Nutr 1999;70:495–501.
27. Becker A, Smulders YM, Teerlink T, Struys EA, de Meer K, Kostense PJ, et al. S-adenosylhomocysteine and the ratio of S-adenosylmethionine to S-adenosylhomocysteine are not related to folate, cobalamin and vitamin B6 concentrations. Eur J Clin Invest 2003;33:17–25.
28. Ingrosso D, Cimmino A, Perna AF, Masella L, De Santo NG, De Bonis ML, et al. Folate treatment and unbalanced methylation and changes of allelic expression induced by hyperhomocysteinaemia in patients with uraemia. Lancet 2003;361:1693–9.
29. Jenkins KJ, Correa A, Feinstein JA, Botto L, Britt AE, Daniels SR, et al. Noninherited risk factors and congenital cardiovascular defects: current knowledge: a scientific statement from the American Heart Association Council on Cardiovascular Disease in the Young: endorsed by the American Academy of Pediatrics. Circulation 2007;115:2995–3014.
30. Nurk E, Tell GS, Vollset SE, Nygard O, Refsum H, Nilsen RM, et al. Changes in lifestyle and plasma total homocysteine: the Hordaland Homocysteine Study. Am J Clin Nutr 2004; 79:812–9.
© 2008 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.