Association Between Maternal Folic Acid Supplementation and Congenital Heart Defects in Offspring in Birth Cohorts From Denmark and Norway

Background Evidence linking individual‐level maternal folic acid supplementation to offspring risk of congenital heart defects is lacking. We investigated whether folic acid supplementation in early pregnancy reduces offspring risk of heart defects in 2 large birth cohort studies. Methods and Results Women recruited in early pregnancy within the DNBC (Danish National Birth Cohort), 1996–2003, and MoBa (Norwegian Mother and Child Cohort Study), 2000–2009, were followed until delivery. Information on periconceptional intake of folic acid and other supplements was linked with information on heart defects from national registers. Among 197 123 births, we identified 2247 individuals with heart defects (114/10 000). Periconceptional (4 weeks before through 8 weeks after conception) use of folic acid plus other supplements (54.8%), folic acid only (12.2%), and non–folic acid supplements (5.0%) were compared with no supplement use (28.0%); the adjusted relative risks of heart defects were 0.99 (95% CI, 0.80–1.22), 1.08 (95% CI, 0.93–1.25), and 1.07 (95% CI, 0.97–1.19), respectively. For initiation of folic acid in the preconception period weeks −4 to −1 (33.7%) and the postconception periods 0 to 4 weeks (15.5%), 5 to 8 weeks (17.8%), and 9 to 12 weeks (4.6%), compared with no or late folic acid intake (29.1%), relative risks of heart defect were 1.11 (95% CI, 1.00–1.25), 1.09 (95% CI, 0.95–1.25), 0.98 (95% CI, 0.86–1.12), and 0.97 (95% CI, 0.78–1.20), respectively. Relative risks of severe defects, conotruncal defects, and septal defects showed similar results. Conclusions Folic acid was not associated with offspring risk of heart defects, including severe defects, conotruncal defects, or septal defects.

C ongenital heart defects are the most common birth defects worldwide, but their etiology is largely unknown. 1,2 Gene defects contribute to the occurrence of heart malformations, 3 but most heart defects occur in isolation in a family. 4 Consequently, a large proportion of heart defects presumably arise in susceptible individuals who carry low-penetrance genes or gene combinations, possibly in interaction with maternal or intrauterine factors. Although it is well established that maternal intake of folic acid around the time of conception reduces the risk of neural tube defect in offspring, 5,6 its association with reduced risk of congenital heart defects is only suggestive. 7,8 Although the precise role of folic acid supplementation on cardiac morphogenesis remains unclear, folic acid may have a role in the migration of the cardiac neural cells that contribute to the development of the embryonic heart. 9,10 It is important to corroborate this hypothesis with evidence from clinical and population-based cohort studies with prospectively collected information on individual-level maternal folic acid intake. To date, this hypothesis is supported by results from 1 randomized trial of high-dose folic acid supplementation, 11 2 casecontrol studies 7,8 and 1 hospital-based case-control study. 12 However, no significant associations were found in 4 other case-control studies of birth defects including cardiac defects, in studies of cardiac outflow defects, [13][14][15][16] or in a nationwide cohort study of supplement use in pregnancy and offspring risk of heart defect. 17 The inconsistencies among these studies may be related to design differences, recall bias of retrospectively collected exposure data in case-control studies, or less accurate information on dose or timing of folic acid supplementation in pregnancy; therefore, the evidence for a preventive effect of folic acid supplementation on heart defects is still inconclusive.
Knowledge of whether folic acid supplementation may reduce the risk of heart defects is important in strategies for the prevention of congenital malformations. 18 Heart defects account for a third of infant deaths caused by malformations in industrialized countries, 19 and severe heart defects require costly and complex surgical intervention. 20 To date, Denmark and Norway have a policy of no food fortification with folic acid.
We have utilized 2 unique large prospective birth cohorts, the DNBC (Danish National Birth Cohort) 21 and MoBa (Norwegian Mother and Child Cohort Study) 22 to investigate whether periconceptional folic acid supplementation measured as individual-level information of folic acid exposure from up to 4 weeks before through 8 weeks after conception, including the timing of folic acid initiation, reduces the risk of congenital heart defects.

Methods
We can make tabulated data and analytical methods available to other researchers on reasonable request sent to the corresponding author.

Independent Data Access and Analyses
All authors had the ability to query any aspect of the data either directly or through independent analysis. Authors N.Ø., S.B., and J.W. had full access to all data in the study and take responsibility for its integrity and the data analysis.
Birth Cohorts DNBC, 1996DNBC, -2003 More than 100 000 Danish women in early pregnancy were recruited to the DNBC. 21 By linkage, all births were updated with information on heart defects in their children, retrieved from the National Patient Register (NPR) and the Causes of Death Register, classified into cardiac phenotypes as published previously. 4,23 In addition, information on chromosomal aberrations from the Danish Cytogenetic Central Register, University of Aarhus, and familial heart defect cases from the Danish Familial Relational Database were included. The NPR contains information on in-patient diagnoses assigned since 1977 and outpatient diagnoses from 1995 onward. Diagnoses of heart defects in the NPR have been validated against clinical records. 24 The Causes of Death Register contains death certificate information, including underlying cause of death and up to 3 contributing causes of death since 1970, with high validity for infant death causes. 25 The Danish Cytogenetic Central Register was established in 1968 and contains reports on all pre-and postnatal chromosomal analyses performed in Denmark since 1970 and 1960, respectively.

MoBa, 2000-2009
By June 2009, >100 000 pregnant women were included in MoBa. 22 We used the MoBa analytic data file version 7, linked with information from the Medical Birth Registry of Norway. All births in MoBa were updated with information on heart defect Clinical Perspective What Is New?
• In 2 independent prospective birth cohorts of almost 200 000 births, we identified 2247 children with congenital heart defects. Our results do not support the hypothesis that maternal intake of supplements containing folic acid before or after conception reduces the risk of congenital heart defects in offspring. • Our null association between maternal folic acid supplementation and offspring congenital heart defects was consistent across various definitions of folic acid exposure and types of cardiac defects.
What Are the Clinical Implications?
• Although most likely not harmful, the effect of maternal folic acid supplementation with respect to preventing cardiac birth defects may be questioned, at least in regions with sufficient intake of dietary folate. • Our finding of no association between individual-level maternal folic acid supplementation and offspring heart defect is in contrast to time trend studies claiming a causal relationship between folic acid and heart defects when comparing trends of birth prevalence of heart defects before and after folic acid fortification of staple food.
from a national project database on heart defects, Congenital Heart Defects in Norway, Cardiovascular Disease in Norway Project (https://cvdnor.w.uib.no/), in which individuals with heart defects were ascertained from 4 data sources: the Medical Birth Registry of Norway, the hospitals' patient administrative systems, Oslo University Hospital's clinical database for children with heart disease, and the Causes of Death Registry. 26 Individuals with heart defects had been assigned specific cardiac phenotypes before linkage with MoBa. Information on chromosomal aberration and familial heart defect was ascertained from the Medical Birth Registry.

Dietary Supplement: Assessment of Folic Acid Exposure
In the DNBC, information on folic acid brand names, dosage, and frequency of use was collected at enrollment covering the period from 4 weeks before the last menstrual period until enrollment (around the 12th gestational week). At enrollment the women answered a questionnaire including information on supplement use. This recruitment form had 2 formats during the study period: initially a format with a floating time window and then, halfway through the study period, a format with a fixed time window (each week specified from gestational week À4 to week 14). Information from the first recruitment form was manually set to the same format as the second recruitment form. 27 In MoBa, the women answered a questionnaire including information on folic acid supplement use and intake of specific nutrients from 3 months before pregnancy and throughout pregnancy until enrollment (18th gestational week). More detailed information on exposure matrixes in both cohorts has been published. [28][29][30] Data have also been used to evaluate health campaigns for the intake of folic acid during pregnancy. 29 Self-reported folic acid intake has been validated and found to mirror biological levels of folate in erythrocytes 31 and in plasma. 32 Therefore, use of folic acid is one of the most well-defined exposures so far characterized collaboratively between the 2 cohorts with the original objective of studying its influence on neural tube defects. The embryonic heart forms from the second through the eighth weeks after conception. 33 Therefore, the presumed window of susceptibility from insufficient maternal folate regarding embryonic heart development was defined as 4 weeks before through 8 weeks after conception. Exposure information of maternal folic acid supplement use was classified by 2 approaches (supplement type and timing), as suggested by Roth and colleagues. 34,35 First, we assessed any supplement use 4 weeks before through 8 weeks after conception for 4 mutually exclusive exposure categories: folic acid (0.4 mg) plus other micronutrient supplements, only folic acid, other non-folic acid supplements, and no supplements (as the reference). Next, we evaluated initiation of any folic acid intake (folic acid plus other supplements were combined with folic acid only) for 4 time windows: 4 weeks before conception, 4 weeks after conception, 5 to 8 weeks after conception, and 9 to 12 weeks (9-11 weeks in Denmark) after conception, with no folic acid supplementation (no supplements was combined with non-folic acid supplements) in the period 4 weeks before to 12 weeks (11 weeks in Denmark) after conception (as the reference).

Case Ascertainment and Classification of Congenital Heart Defects
In both birth cohorts, infants with International Classification of Diseases, 10th Revision (ICD-10) codes for heart defects were classified into cardiac phenotypes in a hierarchical system 23,26,36 : heterotaxia, conotruncal defects, atrioventricular septal defects, anomalous pulmonary venous return, left ventricle outflow tract obstruction, right ventricle outflow tract obstruction, isolated septal defects (ventricular septal defects, atrial septal defects registered after 6 weeks, and ventricular plus atrial septal defects), complex defects, patent ductus arteriosus registered after 6 weeks or with surgical correction in preterm live births (gestational age <37 weeks) or live births at term (gestational age ≥37 weeks), other specified heart defects, unspecified heart defects (Table S1). In tables, we report relative risk estimates for any type of heart defect (not counting patent ductus arteriosus in preterm births), severe heart defects (heterotaxia, conotruncal defects, atrioventricular septal defects, anomalous pulmonary venous return, left or right ventricle outflow tract obstruction, complex defects combined), conotruncal defects, and septal defects (ventricular septal defects, atrial septal defects registered after 6 weeks, and ventricular plus atrial septal defects).

Potential Confounders and Stratification Variables
We decided a priori on the following potential confounders (variables associated with maternal folic acid supplement use and most likely also with the risk of heart defect in offspring) for the association between maternal folic acid supplementation and offspring heart defect risk: year of birth and birth order, maternal age, prepregnancy body mass index (kg/m 2 ), and epilepsy (or intake of folic acid antagonists for maternal epilepsy). In addition, we also considered maternal socioeconomic status and education, pregestational diabetes mellitus, smoking, and alcohol consumption as possible confounders. We did not, however, include these factors in the final model because they were either associated only with the exposure (maternal education, smoking, alcohol consumption) or only with heart defects in the offspring (pregestational diabetes mellitus). We also explored the main effects when restricting the analyses by information on live births, singletons, mother without heart defect, or planned pregnancy. Individuals with missing data for the covariates among all births (10 352 of 197 213, 5.5%) and births with cardiac defect (117 of 2247, 5.2%) were not included in the multivariable regression analyses.

Study Cohort and Statistical Analysis
The 2 study cohorts consisted of live births, and MoBa also included stillbirths and pregnancy termination for fetal abnormalities. We excluded births with chromosomal aberrations (353 in DNBC and 140 births in MoBa). The association between maternal folic acid intake and the risk of heart defect was calculated as risk ratios (RRs) comparing offspring risk of heart defect among those exposed and not exposed to supplementation. The RRs were estimated with 95% CIs in loglinear binominal regression analyses with adjustment for the a priori confounders in the separate analyses of the Danish and Norwegian cohorts, and then with adjustment for interactions between country and the a priori country-specific confounders when combining the 2 cohorts.

Study Size and Power Calculation
A priori, we estimated that we would be able to detect a reduced risk of heart defect (all types combined) and conotruncal defects among infants exposed to folic acid supplementation of at least 16% and 41%, respectively, relative to unexposed infants. Estimates were based on 80% power and a 5% significance level, the Danish prevalence of heart defects without chromosomal aberrations in 93 per 10 000 and in 12 per 10 000 for conotruncal defects, 23 and a mean prevalence of folic acid supplementation of 23% (Denmark 13.5% 37 and Norway 32.0% 28 ).

Ethics
We have approval from the DNBC's steering committee (ref.

Denmark: DNBC
In the study cohort from Denmark of 94 228 births without chromosomal aberrations, we present information on mothers' periconceptional intake of supplements (Table 1). In the period 1996-2003, maternal intake of folic acid plus other supplements not containing folic acid was reported in 59.0% of the births, intake of only folic acid was reported in 4.1%, use of other non-folic acid supplements was reported in 3.7%, and no maternal supplement intake was reported in 33.2%.

Norway: MoBa
The study cohort from Norway consisted of 102 985 births without chromosomal aberrations ( Table 2). In the period 1999-2009, the mother reported intake of folic acid plus other supplements in 50.3% of the births, folic acid only in 19.0%, other supplements without folic acid in 6.2%, and no supplement intake in 24.5%.
Maternal use of folic acid or other supplements is shown for maternal and birth characteristics in both cohorts (Tables 1 and 2).

Cardiac Defects and Birth Prevalence in the Danish and Norwegian Birth Cohorts
Country-specific numbers of heart defects and birth prevalence are presented in Table 3. The 2 birth cohorts were combined, and among 197 213 births, there were 2247 infants with any type of heart defect, not counting patent ductus arteriosus among preterm births (birth prevalence 114 per 10 000 births). There were 583 infants with severe heart defect (30 per 10 000), of which 201 had conotruncal defect (10 per 10 000 births), and 1191 infants had septal defect (60 per 10 000).

Folic Acid Supplementation and Risk of Congenital Heart Defects in Denmark and Norway
Excluding 5.2% with missing values, the combined cohort consisted of 186 861 births. In the period from 4 weeks before through 8 weeks after conception, 54.8% of the mothers used folic acid plus other types of non-folic acid supplements, 12.2% took only folic acid supplements, 5.0% used non-folic acid supplements, and 28.0% did not use any supplements ( Table 4). The adjusted RRs for any type of heart defect in offspring were 0.99 (95% CI, 0.80-1.22) for maternal intake of folic acid plus other supplements, 1.08 (95% CI, 0.93-1.25) for only folic acid supplements, and 1.07 (95% CI, 0.97-1.19) for non-folic acid supplements, compared with no supplement use. RR estimates were similarly nonsignificant for conotruncal defects, severe heart defects, and septal defects for any combination of intake of folic acid and nonfolic acid supplementation. We next analyzed initiation of any folic acid supplementation within the period from 4 weeks before to 12 weeks after conception (Table 4). Among the 186 861 births, 33.7% of the mothers started folic acid supplementation during weeks À4 to À1, 15.3% started during weeks 0 to 4, 17.6% started during weeks 5 to 8, and 4.6% started during weeks 9 to 12, whereas 29.1% did not take folic acid supplementation or started to take folic acid supplementation after week 12. The adjusted RRs of any heart defect in offspring were 1.11 (95% CI, 1.00-1.25), 1.09 (95% CI, 0.95-1.25), 0.98 (95% CI, 0.86-1.12), and 0.97 (95% CI, 0.78-1.20) for maternal folic acid intake in weeks À4 to À1, 0 to 4, 5 to 8, and 9 to 12, respectively, compared with the reference group with no or late folic acid intake. RR estimates were similarly nonsignificant for severe heart defects, conotruncal defects, and septal defects for any timing of folic acid use compared with no folic acid use.
Numbers and relative risks of offspring heart defect for maternal intake of folic acid and other non-folic acid supplements were estimated separately for Denmark and Norway (Tables S2 and S3). In each country, the relative risk of heart defects in offspring was nonsignificant when comparing folic acid with no folic acid in early pregnancy.
In Table S4, adjusted RRs in the combined cohort showed similar null findings for the association between folic acid and any heart defect or conotruncal defects using the final model, model 2 with further adjustment for maternal education, or model 3 with  the following covariates added to model 1: maternal diabetes mellitus, maternal smoking before conception, and maternal alcohol consumption 3 months before conception.
In sensitivity analyses, we restricted the combined cohort to live births, singleton births, births to mothers who did not have a heart defect, or births of a planned pregnancy, and we estimated sets of RRs for offspring risk of any heart defect ( Table S5). The RRs showed no association between folic acid and any heart defect regardless of various restrictions of the birth cohort.

Discussion
In 2 independent national birth cohorts in Denmark and Norway of almost 200 000 pregnancies with prospectively collected information on folic acid supplement intake with follow-up of all births, we identified 2247 births with congenital heart defects in the offspring. We did not find support for the hypothesis that maternal periconceptional supplementation containing folic acid would reduce the risk of congenital heart defects. The findings were consistent using 2 exposure approaches, any intake of folic acid supplementation during the period from 4 weeks before and 8 weeks after conception, and timing of such intake in the periconceptional period. Specifically the risks of conotruncal defects, severe heart defects, or septal defects were not reduced by such supplementation.

Previous Findings From Studies of the Association Between Folic Acid and Congenital Heart Defects
There are studies comparing birth prevalence trends before and after staple food folic acid fortification, 38,39 claiming a causal relationship between folic acid and heart defect. 38 However, a prerequisite for discussion of causal inference are studies designed with individual-level information of maternal folic acid supplement use and infant risk of heart defect. Although several studies have individual-level information of folic acid intake, these studies have conflicting findings most likely related to different design, bias, or confounding, they could also be explained by regional differences in dietary vitamin insufficiency. A detailed discussion of previous reports in the context of regions-for example, the United States, 7,15,16,[40][41][42] Hungary, 11,43 Western Australia, 14 the Netherlands, 8,13 China, 12 and Norway 17 -can be found in Data S1. To summarize previous findings from studies of the association between folic acid and congenital heart defects, we found 1 randomized controlled trial reporting a nonsignificant protective effect of high-dose folic acid supplements, 11 whereas 1 cohort study 17 and most other case-control studies 7,13-16,40,42 except one of the Dutch studies 8 do not support the hypothesis that folic acid supplementation prevents congenital heart defects. The overall lack of a protective effect from folic acid supplementation for heart defect risk in previous studies supports the null findings in the present study and is in stark contrast to inference drawn from trend studies of birth defects, including heart defects, beforeand-after folic acid fortification, or folic acid recommendation to fertile women. 38,44 In support of the null findings in the present birth cohort study from Norway and Denmark, in particular for conotruncal defects, is a report using a Mendelian randomization approach that genotyped >3000 individuals with conotruncal defects for the variants in the MTHFR (methylenetetrahydrofolate reductase) gene. 45 The variant 677TT has previously been found to be associated with reduced folate levels in blood and could be associated with offspring risk of heart defects. The multicenter study found no association between the variant 677TT and the risk of conotruncal defects compared with healthy controls. The authors also performed a meta-analysis of 16 published studies on the relationship between conotruncal heart defects and the variant 677TT and found a weak positive association, likely explained by publication bias: namely, that the studies with nonsignificant or null associations are less frequently published.

Interpretation of the Null Association Between Periconceptional Folic Acid and Cardiac Defects
There could be several explanations for the null findings between maternal folic acid intake in the periconceptional period and offspring risk of cardiac malformations in this study. First, the officially recommended folic acid dose of 0.4 mg/d could be insufficient to prevent cardiac malformations. The feature of the 2 studies from Hungary, 11,43 although of less optimal design, was the use of high-dose folic acid (0.8 and 5.6 mg/d), which was associated with reduced risk of cardiac malformation. Second, the null findings in the present study could also be explained by sufficient prepregnancy folate status among the participating mothers. In less advantaged regions or countries with lower baseline folate levels among fertile women, folic acid supplementation could be beneficial for the prevention of cardiac defects, as has been suggested in parts of China. 12 Third, mothers who take folic acid supplements are in general healthier than nonusers, indicated by lower body mass index, less smoking, more pregnancy planning, higher age at delivery, lower parity, or higher education, 32 with some of these factors also reported to be associated with lower offspring risk of cardiac malformation. However, we were able to adjust for potentially negative confounding. Fourth, the timing when women start taking folic acid may be too late in the periconceptional period to prevent development of cardiac malformations. However, initiation of folic acid classified in 4-week intervals in the period 4 weeks before through 11 to 12 weeks after conception all showed a null association with offspring risk of cardiac malformation. Fifth, it could be that folic acid supplements prevent neural tube defects 5,6 but not cardiac malformations. Maternal folate insufficiency may act differently during different time windows in embryonic organ development; folate insufficiency probably disturbs neural tube closure corresponding to an early period in embryo formation but not formation of the later embryonic heart development corresponding to a somewhat later period. Sixth, there could be nondifferential misclassification or measurement error of folic acid exposure and timing, which may bias the estimate to the null value. Interestingly, we found similar nonsignificant country-specific relative risks despite differences in collection and classification of folic acid intake. Although most likely not harmful, 46,47 the effect of maternal folic acid supplementation with respect to preventing cardiac birth defects may be questioned, at least in regions with sufficient intake of dietary folate.

Study Strengths and Limitations
The use of prospectively collected information on individuallevel folic acid intake in early pregnancy precluded recall bias of exposure. We have complete follow-up of all births (except for 204 mothers who gave birth abroad) through the nationwide Medical Birth Registers. Before the analyses, we used an established method for classifying folic acid exposure by type (folic acid plus other supplements, only folic acid, non-folic supplements, no supplements) in the period from 4 weeks before through 8 weeks after conception and by timing (initiation of supplement use) in the period from 4 weeks before through 11 to 12 weeks after conception, 34,35 Tables S2 and S3. ¶ Relative risk with 95% CI comparing supplement use with no use (reference) 4 weeks before to 8 weeks after conception (upper panel) or comparing initiation of folic acid supplements with no use or non-folic acid supplements (reference) 4 weeks before to 12 weeks after conception (lower panel). RRs adjusted for interaction between country (Denmark/Norway) and the following factors: year of birth (Denmark: 1996(Denmark: -1997(Denmark: , 1998(Denmark: , 1999(Denmark: , 2000(Denmark: , 2001(Denmark: , 2002(Denmark: -2003Norway: 1999Norway: -2000Norway: , 2001Norway: , 2002Norway: , 2003Norway: , 2004Norway: , 2005Norway: , 2006Norway: , 2007Norway: , 2008Norway: -2009, maternal age (≤24, 25-29, 30-34, ≥35 years), birth order (first, second, third or higher), maternal body mass index (<20, 20-24, 25-29, 30-34, ≥35), maternal heart defect (yes/no), maternal epilepsy before pregnancy (yes/no). Categories for adjustment variables combined for severe and conotruncal defects. # The 4 exposure categories in supplement use collapsed into no/yes; no is no folic acid use (no supplements and other supplements, no folic acid), and yes is folic acid use (folic acid only and folic acid plus other supplementation. **In DNBC, including week 11. corresponding to fetal cardiac development (third to eighth weeks after conception). Folic acid is a well-characterized exposure in the 2 cohorts 27,28,48 and is found to mirror blood plasma folate. 49 We did not have information on perigestational dietary folate, but from previous studies, supplemental folic acid outweighs dietary folate, as women reporting folic acid supplement use have higher plasma folate levels than nonusers regardless of dietary intake. 32 Low ascertainment of severe heart defects, including conotruncal defects, is unlikely because these defects almost always come to medical attention, either because of the need for surgery or at death. In Denmark there is good agreement between severe heart defect diagnoses in the NPR compared with hospital records, 24 and in Norway, nationwide health registries and databases, as well as a clinical register, were combined to ensure virtually complete ascertainment of all heart defects among live births, stillbirths, and terminated pregnancies, including information on chromosomal aberrations and familial heart defect cases. All cases from the clinical register at Oslo University Hospital have been coded by senior pediatric cardiologists.
By combining country-specific findings in 2 national birth cohorts, our study cohort was of sufficient size to produce narrow CIs, although we could not estimate RRs for the very rare outcomes of severe defects. A large set of potential confounders were available for evaluation, for adjustment or restriction of the study population for sensitivity analyses. Finally, although exposure data were collected somewhat differently in the 2 cohorts, we decided not to harmonize exposure among the 2 birth cohorts. 50 The combined countryspecific estimates with corresponding country-specific confounders illustrate that despite the variation in exposure collection and country-specific birth prevalence of heart defect, we found no significant association between maternal folic acid supplementation and heart defect risk in either birth cohort.
In conclusion, maternal periconceptional use of folic acid supplements was not associated with offspring risk of congenital heart defects, including conotruncal defects or other severe defects, as well as septal defects, after combining 2 national birth cohorts with prospectively collected information of periconceptional supplement intake. The possible effect of maternal folic acid supplementation with respect to preventing cardiac birth defects may be questioned, at least in regions with sufficient intake of dietary folate.

Acknowledgments
Tatiana Fomina, PhD, Department of Global Health and Primary Care, University of Bergen, Norway, programmed the algorithm that maps congenital heart defects into embryologically related defect phenotypes. Tomislav Dimoski, the Norwegian Knowledge Center for Health services, Oslo, Norway, developed software necessary for obtaining data from Norwegian hospitals, conducting the data collection, and quality assurance of data. Dr Petter Hagemo constructed and maintained the clinical database for congenital heart defects at the Oslo University Hospital, Norway. Christine Roth, PhD, shared valuable information on folic acid exposure collected in MoBa (Norwegian Mother and Child Cohort Study). We are grateful to all families in Denmark and Norway who participate in this ongoing cohort study. found no preventive effect of folic acid supplementations before pregnancy, a finding also supported by null findings for folic acid intake in the 5 weeks pregestational period, and for the lowest quartile of total folate intake. 4 In a study from the US National Birth Defects Prevention Study, 1997-2002, with the intention to investigate the association between maternal smoking and offspring heart defect, 5 the authors also reported the potential confounder maternal folic acid intake in the periconceptional period one month before and 2 months after conception. By our calculation of the authors' numbers (2284 exposed heart defect out of 3,067 all heart defects; 2,935 non-exposed controls out of 3947 controls), the unadjusted OR was 1.01 (95% CI 0.90-1.12), i.e. the association between folic acid and heart defect risk was null. In multi-center study, 6 we calculated an unadjusted OR of 0.95 (95% CI 0.85-1.06) of any heart defect for maternal use of multivitamins with folic acid (5,205 exposed with heart defects out of 5,979 individuals with heart defects; 4,737 exposed controls out of 5,408 controls).

Sources of Funding
The Hungarian randomized controlled trial of neural tube defects, 7 containing other defects as well, were extended with 8 months follow-up, 8 and reported a reduced risk of heart defect, but the difference was non-significant comparing the group receiving multivitamin supplements containing high dose of 0.8 mg folic acid with the group receiving trace element supplements other than folic acid. The numbers were small, with only 10 exposed and 17 unexposed heart defect cases. In a later follow-up, but in case-control design, 598 children with heart defect born in Hungary 1980-1996 were compared to 902 matched controls, 9 The authors reported a reduced risk of conotruncal heart defect in the group exposed to a very high dose of folic acid, on average 5.6 mg/d. From China, a hospital-based case-control study with 358 prenatally identified heart defect and 422 controls recruited consecutively among women receiving prenatal investigation in the provinces of Guangdong, Hubei, Fujian, and Shanxi, folic acid supplements reduced heart defect risk by 65%. 13 However, there may be limitations of the study, not discussed by the authors, such as design, selection into the study, and classification of exposure.
In the recent study of around 520,000 births registered in the Medical Birth Registry of Norway, 1999-2009, with pre-birth registered use of folic acid and multivitamin supplements before pregnancy and during pregnancy, 14 Table 4.
ǁ The number of events includes the entire cohort, whereas in the adjusted analyses, individuals with missing values for covariates were excluded (total births n=7,556; any heart defect n=85; severe defect n=22, conotruncal defects n=10; septal defects n=36).
** The four exposure categories in supplement use collapsed into no/yes; "no" is no folic acid use (no supplements and other supplements, no folic acid), and "yes" is folic acid use (folic acid only and folic acid plus other supplementation). Table S3. Relative risk (RR) of congenital heart defect* (overall), severe heart defect †, conotruncal defect ‡, and septal defect § by maternal intake of folic acid supplements in the periconceptional period among 102,985 births in the Norwegian Mother and Child Cohort Study (MoBa), Norway, 1999-2009.  Table S4. Relative risks (RR) of congenital heart defect* ( overall), severe heart defect †, conotruncal defect ‡, and septal defect § by maternal intake of folic acid supplements, adjusting for covariates using three models, combining 94,228 birthsǁ in The Danish National Birth Cohort (DNBC), Denmark, 1996-2003, and 102,985 birthsǁ in The Norwegian Mother and Child Cohort Study (MoBa), Norway, 1999-2009. * † ‡ § Congenital heart defects: see definitions in Table 4.
ǁ Individuals with missing values of covariates were excluded in the adjusted analyses, see numbers in Table 4 Any congenital heart defect* Severe heart defect † Conotruncal defect ‡ Septal defect §