Maternal periconceptional use of multivitamins and reduced risk for conotruncal heart defects and limb deficiencies among offspring

American Journal of Medical Genetics 59536-545 (1995)

MATERNAL PERICONCEPTIONAL USE OF MULTIVITAMINS AND REDUCED RISK FOR CONOTRUNCAL HEART DEFECTS AND LIMB DEFICIENCIES AMONG OFFSPRING

Gary M. Shaw, Cynthia D. O'Malley, Cathy R. Wasserman, Marie M. Tolarova, Edward J. Lamer

March of Dimes Birth Defects Foundation, California Birth Defects Monitoring Program, Emeryville, CA (G.M.S, C.D.0, C.R.W, M.M.T): Division of Medical Genetics, Children's Hospital, Oakland, CA (E. J.L) .

We investigated whether a woman's periconceptional use of a multivitamin containing folic acid was associated with a reduced risk for delivering offspring with a conotruncal heart defect or a limb deficiency. Data were derived from a population-based case-control study of fetuses and liveborn infants with conotruncal or limb defects among a 1987-88 cohort of births in California. Telephone interviews were conducted with mothers of 207 (87.0% of eligible) conotruncal cases, 178 (82.0%) limb defect cases, and of 481 (76.2%) randomly selected liveborn nonmalformed control infants. Reduced risks were observed for maternal use of multivitamins containing folic acid from one month before until two months after conception. Odds ratios and 95% confidence intervals for any compared to no multivitamin use were 0.70 (0.46-1.1) for conotruncal defects and 0.64 (0.41-1.0) for limb defects. Controlling for maternal race/ ethnicity, age, education, gravidity, alcohol use, and cigarette use resulted in a further reduction to the odds ratio for conotruncal defects, 0.53 (0.34-0.85), but not for limb defects. Among non-vitamin using

women, consumption of cereal containing folic acid was also associated with reduced risks for both defects. Women who take multivitamins have 30-35% lower risk of delivering offspring with either conotruncal or limb defects. This association may not be attributable to folic acid specifically, but may be a consequence of other multivitamin components, or some unknown behaviors that highly correlate with regular use of a multivitamin. However, should the association prove causal, it offers an important opportunity for preventing thousands of serious birth defects.

KEY WORDS: birth defects, heart defects, conotruncal defects, limb defects, epidemiology, folic acid, vitamins

0 1995 Wiley-Liss, Inc.

INTRODUCTION

Substantial evidence shows that women who use multivitamin supplements periconceptionally have a reduced risk for neural tube defect-affected pregnancies [Laurence et al., 1981; Smithells et al., 1983; Mulinare et

Address reprint requests to Dr. Gary Shaw, California Birth Defects Monitoring Program, 1900 Powell St. Suite 1050, Emeryville, CA 94608 0 1995 Wiley-Liss, Inc.

Limb and Heart Anomalies and Maternal Vitamin Use 537

al., 1988; Bower and Stanley, 1989; Milunsky et al., 1989; MRC Vitamin Study, 1991; Czeizel and Dudas, 1992; Werler et al., 1993; Shaw et al., 1995al. The common vitamin component in these findings appears to be folic acid. There is also evidence that multivitamin use may result i n a reduced risk for orofacial cleft defects [Tolarova, 1982; Shaw et al. , 1995333. From a randomized controlled trial , Czeizel [19931 recently reported a reduced prevalence of still other congenital anomalies among offspring whose mothers took a multivitamin supplement pericon- ceptionally compared with mothers who took only a trace element supplement. Although the trial data were sparse for most individual congenital anomalies, the reduced prevalence overall in the treatment group was primarily attributable to lower frequencies of hypertrophic pyloric stenosis, urinary system anomalies , limb deficiencies, and cardiovascular anomalies. Based on this suggestion of a reduced prevalence for the latter 2 groups of anomalies, we investigated whether maternal periconceptional use of multivitamins, or cereal , containing folic acid, reduces the risk of delivering offspring with limb deficiencies or with conotruncal defects, a pathogenetic subgroup of cardiovascular anomalies [Clark, 19901.

METHODS

For this case-control study, all infants or fetuses with conotruncal heart or limb defects were ascertained by the California Birth Defects Monitoring Program [Croen et al., 19911. This birth defects registry routinely reviews medical records at all hospitals and genetic centers in a known geographic population base. Eligible were infants and fetuses (gestational age > 20 weeks) diagnosed within one year after birth among the cohort of 344,214 births and fetal deaths that occurred between January 1987 and December 1988 to women residing in selected count ie s

throughout California. Ineligible were infants with a trisomy or monosomy X syndrome (45, X) .

Diagnoses of conotruncal heart defects were confirmed by reviewing echocardiography, cardiac cathe- t e r i zation, surgery , or autopsy reports. Each case with a heart defect was documented by at least echocardiographic evidence. For most cases, diagnostic information was available from more than one of the

findings were in conflict, we accepted the autopsy diagnoses, if one was performed, or the diagnoses derived from the most invasive procedure performed. Using these diagnoses, a medical geneticist (EJL) classified each case with complex heart defects using the pathogenetic scheme developed by Clark [19901. Considered as conotruncal defects were all malformations whose pathogenesis was thought to be due to abnormal aorticopulmonary septation: tetralogy of Fallot, dextro-transposition of the great arteries, and other (double outlet right ventricle, truncus arteriosus communis, pulmonary valve atresia with ventricular septal defect, subaortic ventricular septal defect type I, and aorticopulmonary window). Ascertained were 252 infants or fetuses with conotruncal defects.

reviewed diagnostic modalities. If

Diagnoses of limb deficiencies were confirmed by radiography, surgery, or autopsy. Medical records of the procedures were reviewed by a medical geneticist (MMT) who first classified defects of the lower or upper limb as having either a "known" or unknown cause. Those considered to be of "known" cause mostly included monogenic conditions. Defects of unknown cause were further classified into the groups longitudinal or transverse based on the concept of embryologic failures [ Swanson et al. , 19831. Among transverse defects, those of amniotic band origin were considered separately in analyses. Ascertained were 219 infants or fetuses with limb defects.

538 Shaw et al.

A control infant was eligible if he or she was live born during 1987- 88, his or her mother was a resident of the same counties in which cases were ascertained, and he or she did not have a reportable birth defect [Croen et al., 19911 prior to the first birthday. Using a pseudorandom number generator among all eligible liveborn infants (n=341,839), 652 control infants were electronically selected from California vital records.

Telephone interviews were conducted with mothers of cases and controls in English (92%) or Spanish. Women who only spoke languages other than English or Spanish (15 cases and 21 controls), and 1 case mother who died prior to interview contact, were not eligible. Interviewed were 207 (87.0%) of 238 eligible conotruncal case mothers, 178 (82.0%) of 217 limb defect case mothers, and 481 (76.2%) of 631 control mothers. The conotruncal cases included 77 with dextrotransposition of great arteries, 83 with tetralogy of Fallot, and 47 with other anomalies. The limb defect cases included 81 with longitudinal defects, 50 with transverse defects, 34 with defects due to amniotic bands, and 13 with "known" cause. Interviews were completed an average of 3.7 years after the date of delivery for cases and 3.8 years for controls. Information was unavailable from 2.6% of case and 2.7% of control mothers who refused to be interviewed, and from 12.7% of case and 21.1% of control mothers who could not be located.

Considerable effort was made to keep interviewers uninformed of whether the woman being interviewed was a case or a control mother until the end of the interview. Once the interviewer obtained informed consent, she assisted each woman in establishing a 4-month period, 1 month before to 3 months after conception, that was referred to throughout the interview to elicit information on medical and family history, as well as environmental exposures related to her pregnancy. Women were asked if they

used vitamin or mineral supplements during the 4-month period. Women who responded affirmatively were asked about the types of supplements they used ( "prenatal vitamin" , "mu1 t ivit amin" , "vitamin A", 'I f ol i c acid", ''other type, specify"). For each supplement and for each month within the 4-month time period, women were asked about the frequency of use per week and quantity taken each time. Amount of folic acid in a supplement was assumed to be 0.8 mg in "prenatal" vitamins and 0.4 mg in other vitamin supplements determined by a nutritionist to contain folic acid. If a woman used a vitamin whose folic acid content could not be determined, the folic acid intake from that vitamin was considered zero. From supplement composition and frequency of use, we estimated a woman's average daily folic acid intake from all supplements she reported using. Women were also asked if they ate cold cereals in the 4-month period, and, if so, which brands and how frequently they were eaten. Folic acid in a cereal was determined by comparing a cereal to a nutrient database [Dept. Health Human Services, 19891 or to manufacturers' information. Women who consumed cereal that did not contain folic acid were classified as non- cereal users. Women who did not provide cereal type or brand information were classified as unknown. Consumption of one 28-30 g serving of a fortified cereal, is on average 0.1 mg folic acid, which is 25% of the folic acid content of a unit dose in a multivitamin supplement.

The odds ratio along with its 95% confidence interval was used to estimate risk. These estimates, including those derived from logistic regression models, were computed using EGRET [19911. For each case group, analyses were performed for folic acid intake from multivitamins for the period 1 month before through 2 months after conception. We chose the latter endpoint because heart and limb morphogenesis are complete by 2 months after conception. Women who started

Limb and Heart Anomalies and Maternal Vitamin Use 539

TABLE I. Maternal and Infant Characteristics (Percent)’ of Cases and Controls

Conotruncal Limb Reduction Defect Defect Controls Cases Cases (N=2 0 7 ) (N=17 8 ) (N=4 8 1 )

Race/ethnicity Hispanic 23.7 Non-Hispanic white 65.7 African American 3.4 Other 7.4

A g e < 20 20 - 24 25 - 29 30 - 34 35 - 39 > 39

Education < High school

graduate High school

24.7 28.7 62.9 57.6 6.2 2.9 6.2 10.6

7.7 12.9 9.8 24.6 27.0 26.8 34.3 30.3 33.5 22.2 20.2 22.0 10.1 7.3 6.7 0.5 1.1 0.8

19.3

31.9 graduate

31.4 Some college 17.4 College graduate

Employed 64.3

Gravidity 1 2 3

2 4

22.2 24.6 27.5 24.6

25.8

33.2

28.7 11.2

52.3

23.5

29.1

30.4 16.4

55.1

20.8 27.0 28.7 26.0 18.5 22.2 29.8 21.6

Smoked cigarettes” 29.0 25.8 23.7

Alcohol usea None

< Once/week > Once/week

Child sex Male

50.7 39.6 9.7

57.5

62.4 59.0 32.0 34.3 4.5 6.0

55.6 49.7 %May not add to 100 due to rounding or missing information for some individuals.

”Refers to use in the period 1 month before through 3 months after conception.

540 Shaw et al.

TABLE 11. Control Status

Maternal Use of Multivitamins Containing Folic Acid by Case and

Conotruncal Limb Maternal Vitamin Use defect cases de f e c t cases Controls None in period 1 month 47 45 86 before through 3 months after conception Any in period 1 month before through 2 months after conception Daily 113 Less than daily 11 Frequency not stated 1

Unknowna 3

Any starting third month 32 after conception

Total 207

97 13

22

1 178

288 32 6 63

6 481

"Includes women who didn't know if they took vitamins or women who took a vitamin with unknown folic acid content.

vitamin use in the third month after concept ion were excluded from comparisons. Folic acid intake was primarily considered as a dichotomous variable (use/no use), but also as a polytomous variable (no use, < O . 4, 0.4-0.9 and >0.9mg/day), approximating doses of less than, about the level, and more than the amount of folic acid usually found in unit doses of multivitamins. Analyses of cereal intake were restricted to women who did not use vitamin supplements containing folic acid.

RESULTS

Compared to control mothers, mothers whose infants had conotruncal defects were more likely to be non- Hispanic whites, older than 34 years, multigravidas, cigarette smokers, alcohol users, and employed. Mothers whose infants had limb defects were more likely to be non-Hispanic whites, under age 20 years, multigravidas, and less likely to have graduated from college (Table I). Case infants were more likely to be male.

Most case (61.8%) and con- trol (67.8%) mothers reported use of multivitamins containing folic acid in the period from 1 month before through the 2 months after conception (Table 11). Women who used multivitamins had a 30% lower risk of delivering offspring with conotruncal heart defects and a 36% lower risk of delivering offspring with limb defects (Table 111). However, the odds ratios reflecting these lowered risks had associated confidence intervals that included 1.0. Among conotruncal defect case subgroups, we found that much of the overall reduction in risk was accounted for by the lower risk of tetralogy of Fallot, but estimates for dextrotransposition or for other conotruncal defects were closer to unity and were imprecise as evidenced by overlapping confidence intervals. For limb defect case subgroups, no reduction in risk was observed for transverse defects, whereas the largest subgroup, longitudinal defects, was largely responsible for the lower risk observed for the overall group of limb defects. The risk estimates for the other two limb defect subgroups, "known cause" and

Limb and Heart Anomalies and Maternal Vitamin Use 541

"amniotic band, I' were imprecise in part due to small numbers of subjects. Odds ratios for 3 levels (<0.4, 0.4- 0.9, and - >1.0 mg) of average daily intake of folic acid from multivitamins by women were similar to the odds ratio for any use for both case groups (not shown).

The risk reductions associated with vitamin use were not substantially affected by excluding less than daily users (10% of all users), by removing from the non-user group 8 (3 cases and 5 controls) women

who used a vitamin that did not contain folic acid, nor by excluding from analyses cases (each defect group) and controls whose mothers had a history of diabetes or epilepsy/ seizures, or who used oral con- traceptives (suspected folate antagonists) post-conceptionally. For conotruncal cases, excluding cases and controls who had a family history of congenital heart disease in a first degree relative also did not substantially alter the risk reduction.

'TABLE 111. Any Maternal Use of Multivitamins Containing Folic Acid From 1 Month Before Through 2 Months After Conception and Risk for Conotruncal Heart, Limb Defects, and Their Component Anomalies

h Y

Use Ratio Interval Multivitamin Odds 95% Confidence

Contro 1 s All conotruncal defects Dextrotransposition Tetralogy of Fallot Other

All limb defects

Longitudinal Transverse "Known cause" Amn i o t i c b and

32 6

125

Reference 0.70 0.46-1.1

52 45 28

110

51 35 7 17

0.86 0.54 0.82

0.64

0.61 1.0 0.37 0.50

0.45-1.7 0.30-0.98 0.35-2.0

0.41-1.0

0.34-1.1 0.45-2.4 0.10-1.4 0.20-1.3

TABLE IV. Any Consumption of Cold Cereal among Women Who Did Not Use Multivitamins Containing Folic Acid From 1 Month Before Through 3 Months After Conception and Risk for Conotruncal Heart or Limb Defects.

Any Cereal No Cereal Odds 95% Confidence Consumption Consumption Ratio Interval

--- Controls 62 22 Reference

defects

defects

Conotruncal 31 14 0.79 0.36-1.7

Limb 28 16 0.62 0.29-1.3

542 Shaw et al.

Analysis of any use of a multivitamin containing folic acid was done controlling for mate rna 1 race/ethnicity (Hispanic, non-Hispanic white, African-American, other) , education (<high school graduate, high school graduate, some college, college graduate) , age (<20, 20-24, 25-29, 30- 34, 35-39, >39 years) , gravidity (0,1,2,3,4t previous pregnancies), smoking (0, 1-19, >19 cigarettes per day) , and alcohol use (no use, <once/week, lonce/week) . For cono- truncal defects, the adjusted risk estimate for multivitamin use dropped to 0.53 (0.34-0.85) , whereas for limb defects, the adjusted odds ratio was not substantially different from its crude counterpart, odds ratio=0.70 (0.43-1.1).

Any folic acid-fortified cereal consumption among the 178 women who did not use multivitamins containing folic acid in the 4-month period was associated with a reduced risk for both anomaly groups (Table IV) . However, these risks were imprecise with confidence intervals including 1.0. Daily cereal consumption was associated with a greater reduction in risk for conotruncal, odds ratio=0.17 (0.02-0.93) , but not limb defects, odds ratio=0.65 (0.21-2.0).

DISCUSSION

This population-based study shows that mothers who take multivitamins with folic acid have 30-35% less risk of delivering offspring with either conotruncal heart or limb defects. The study also suggests, but data were too sparse to draw a firm inference, that greater risk reductions may be associated with specific subgroups of these two defect groups. Despite the fundamental importance of folate in synthesis of nucleotides necessary for DNA replication and amino acid metabolism, relatively few studies have investigated whether peri- conceptional multivitiam/folic acid use is associated with reduced risks for congenital anomalies other than neural tube defects [Tolarova, 1982; Bower and Stanley, 1992; Czeizel,

1993; Shaw et al. , 1995333. The risk reduction observed in our study is consistent with the findings of Czeizel [19931, but not with those of Bower and Stanley [1992]. Czeizel [1993] observed risk reductions (40- 80%) for both heart and limb defects in a randomized controlled trial in which one group of women received a daily multivitamin supplement containing 12 vitamins, including 0.8 mg of folic acid, plus minerals and the other group of women received a supplement containing only trace elements. The trial had small numbers of infants born with each specific malformation, therefore the reduced risks among the treatment group may have been due to random variation. The Bower and Stanley [1992] study did not include limb defects and the number of cases with heart defects was too few to rule out risks of the magnitude observed in our study.

Our results can not discriminate whether the observed risk reduction is attributable to folic acid, to another of many vitamin and mineral constituents in multivitamin supple- ments (and fortified cereals), or to an unknown "behavior" highly cor- related with multivitamin/folic acid or cereal use. Too few mothers used a vitamin that only contained folic acid (n=4), or used a vitamin supplement without folic acid (n-8), to assess whether the risk reduction can be due solely to folic acid. Because of its randomized design, the trial by Czeizel [19931, provides some evidence against an uncontrolled "behavior" tightly associated with vitamin use as the explanation of the observed risk reduction. However, the trial data can not distinguish whether the preventive agent is folic acid or one of the 11 vitamins or 4 minerals contained in the trial's supplement.

Evidence in support of the protective agent being folic acid comes from studies conducted in experimental animals. Folic acid deficiency during gestation has been associated with multiple congenital anomalies in rats, including limb defects and conotruncal defects [Baird

Limb and Heart Anomalies and Maternal Vitamin Use 543

et al., 1954; Nelson et al., 1955; Asling et al., 19551. Evidence to suggest that the protective agent may not be limited to folic acid also exists. Numerous vitamin and mineral deficiencies other than folic acid have been associated with congenital anomalies in animals [Sharma, 19931. For example, Nelson et al. [1956] showed a relation between riboflavin deficiency and abnormal embryonic rat development. Moreover, one small human study of women's dietary intake has shown mothers of infants with congenital heart defects to have lower dietary intakes of protein, calories, iron, vitamin C, carbohydrate, and niacin [Pitt and Samson, 19611. Thus, the current evidence appears insufficient to conclude that folic acid is the agent responsible for the observed risk reduction for conotruncal and limb defects.

Chance or bias may be alternative explanations for our finding a reduced risk associated with multivitamin use. Many of the computed risks had associated confidence intervals that included 1.0, indicating limited precision in risk estimation. Further, we cannot exclude the possibility that the risk reduction was spurious due to recall bias. Such bias could result if case mothers underreported, or if control mothers overreported, vitamin use. However, 2 studies [Werler et al., 1993; Shaw et al., 1995al conducted in the same time period did not find evidence of biased maternal recall of vitamin use as a likely explanation for a decreased risk for neural tube defects. Although we can not directly measure recall bias, we can estimate the extent of reporting errors that would be necessary to create our results. If the odds ratio of 0.70 observed for conotruncal defects arose as a result of biased reporting, then either 8% of vitamin-using control mothers would have had to falsely report that they used vitamins when they had not, or 23% of nonvitamin-using case mothers would have had to falsely report that they did not use vitamins, when, in fact, they had. Further evidence against recall bias is the risk

reduction associated with maternal cereal use. Most women were probably unaware of our motives for asking about their cereal intake. With respect to potential selection bias, it is unlikely that the risk reduction was due to nonresponding control mothers having a substantially smaller proportion of vitamin users because the percentage of control mothers who reported use of vitamins in the first trimester, 82%, compared well with percentages observed in other studies [Werler et al., 1993; Shaw et al., 1995al. In addition, given the high percentage of case mothers who participated, it is unlikely that the risk reduction could be due to vitamin users being more common among nonresponding case mothers.

More than 4000 infants are born in the United States each year with conotruncal heart or limb defects given a combined prevalence of approximately 1 per 1000 livebirths [Adams et al., 1989; Lin et al., 19931. The morbidity, mortality, lifetime medical cost [Waitzman et al., 19941 and family anguish associated with these conditions are enormous. Our results, along with findings from one randomized controlled trial [Czeizel, 19931 suggest a risk reduction for conotruncal and limb defects among offspring of women who used a multivitamin periconceptionally. Whereas additional studies with larger numbers of cases and more information on dietary folate intake are needed to explore this association further, should this association prove causal, it opens an unprecedented opportunity for preventing thousands of serious birth defects.

ACKNOWLEDGMENTS

We are indebted to Donna Schaffer for her nutritional consultation. This work was partially supported by the Cigarette and Tobacco Surtax Fund of California Tobacco-Related Disease Research Program, University of

544 Shaw et al.

California, 1RT466. These results were presented at the 1994 Teratology Society Meeting, Puerto Rico.

RE FERENCE S

Adams MM, Mulinare J, Dooley K (1989) : Risk factors for conotruncal cardiac defects in Atlanta. J Am Coll Cardiol 14:432-442.

HM (1955): Congenital skeletal abnormalities in fetal rats resulting from maternal pteroylglutamic acid deficiency during gestation. Anat Rec 121:775- 800.

Baird CDC, Nelson MM, Monie IW, Evans HM (1954) : Congenital cardiovascular anomalies induced by pteroylglutamic acid deficiency during gestation in the rat. Circ Res 2:544-554.

folate as a risk factor for neural-tube defects: Evidence from a case-control study in Western Australia. Med J Aust 150:613-619.

Bower C, Stanley FJ (1992): Dietary folate and nonneural midline birth defects: No evidence of an association from a case-control study in Western Australia. Am J Med Genet 44:647-650.

Clark EB (1990) : Growth, morphogenesis and function: The dynamics of heart development. In: Moller JH, Neal WA (eds. ) "Fetal, Neonatal and Infant Cardiac Disease." Norwalk CT: Appleton and Lange, 3-23.

Croen LA, Shaw GM, Jensvold NJ, Harris JA(1991): Birth defects monitoring in California: A resource for epidemiological research. Pediatr Perinatol Epidemiol 5:423-427.

Czeizel AE (1993): Prevention of congenital abnormalities by periconceptional multivitamin supplementation. Br Med J 306:1645- 1648.

Czeizel AE, Dudas I (1992): Prevention of the first occurrence of neural- tube defects by periconceptional vitamin supplementation. New Engl J Med 327:1832-1835.

Services (1989): National Institutes

Asling CW, Nelson MM, Wright HV, Evans

Bower C, Stanley FJ (1989) : Dietary

Department of Health and Human

of Health, National Cancer Institute. Health habits and history questionnaire: Dietary history and other factors: Personal computer system packet. Version 2.3, Bethesda, MD.

EGRET (1991): Statistics and Epidemiology Research Corporation, Seattle, Washington.

Laurence KM, James N, Miller MH, Tennant GB, Campbell H (1981) : Double-blind randomized controlled trial of folate treatment before conception to prevent recurrence of neural-tube defects. Br Med J 282:1509-1511.

Lin S, Marshall EG, Davidson GK, Roth GB, Druschel CM (1993): Evaluation of congenital limb reduction defects in upstate New York. Teratology

Milunsky A, Jick H, Jick SS, Bruell CL, MacClaughlin DS, Rothman KJ, Willet W(1989):Multivitamin/folic acid supplementation in early pregnancy reduces the prevalence of neural tube defects. JAMA

47:127-135.

262:2847-2852. MRC Vitamin Study Research Group

(1991): Prevention of neural tube defects: Results of the Medical Research Council vitamin study. Lancet 338:131-137.

Mulinare J, Corder0 JF, Erickson JD, Berry RJ (1988): Periconceptional use of multivitamins and the occurrence of neural tube defects.

Nelson MM, Baird CDC, Wright HV, Evans HM (1956) : Multiple congenital abnormalities in the rat resulting from riboflavin deficiency induced by the antimetabolite galactoflavin. J Nutr 58:125-134.

Nelson MM, Wright HV, Asling CW, Evans HM (1955) : Multiple congenital abnormalities resulting from transitory deficiency of pteroylglutamic acid during gestation in the rat. J Nutr 56:349.

malformations and maternal diet. Aust Ann Med 10:268-274.

Sharma RP (1993): Dietary factors and birth defects. Association for the Advancement of Science. California Academy of Sciences.

JAMA 260:3141-3145.

Pitt DB, Samson PE (1961): Congenital

Limb and Heart Anomalies and Maternal Vitamin Use 545

Shaw GM, Lamer EJ, Wasserman CR, O'Malley CD, Tolarova MM (1995b) : Risks of orofacial clefts in children born to women using multivitamins containing folic acid periconceptionally. Lancet 345:393- 396.

Shaw GM, Schaffer D, Velie E, Morland KM, Harris JA (1995a): Periconceptional vitamin use, dietary folate, and the occurrence of neural tube defects. Epidemiology 6:219-228.

Smithells RW, Seller MJ, Harris R, Fielding DW, Schorah CJ, Nevin NC, Sheppard S, Read AP, Walker S, Wild J (1983): Further experience of vitamin supplementation for

prevention of neural tube defect recurrences. Lancet 1:1027-1031.

Swanson AB, Swanson G, Tada K (1983): A classification for congenital limb malformation. J Hand Surg 8:693-702.

Tolarova MM (1982) : Periconceptional supplementation with vitamins and folic acid to prevent recurrence of cleft lip. Lancet 2:217.

Waitzman NJ, Romano PS, Scheffler RM (1994): Estimates of the economic costs of birth defects. Inquiry 31:188-205.

Werler MM, Shapiro S, Mitchell AA (1993): Periconceptional folic acid exposure and risk of occurrent neural tube defects. JAMA 269: 1257- 1261.