Evidence for Benefits of Folic Acid Supplementation

One RCT, two cohort studies, and eight case-control studies met eligibility criteria, in addition to two publications from the prior review.^36,86 These 20 publications, comprising 11 primary studies and one systematic review, drew from eight data sources (Hungarian trial,^88-94 Hungarian cohort,⁸¹ New England study,^97,98 National Birth Defects Prevention Study,^9,87 Slone Birth Defects Study,^11,95,99 Texas Department of Health's Neural Tube Defect Project,²² California Birth Defects Monitoring Program,⁸² and NICHD Neural Tube Defects Study⁹⁶). Together they span births occurring over three decades, from 1976 through 2007.

After the publication of the Hungarian trial and other trials in women with recurrent NTDs, the clear evidence of benefit pointed to the need for large-scale public health interventions, and the United States added folate to grain products in 1998. The clear evidence of benefit also made the conduct of additional trials unethical. As a consequence, all subsequent studies relied on observational data.

Although all included studies in this review avoid fatal flaws, their designs contain inherent and unavoidable sources of bias. Prospective studies may not be able to ascertain all cases. Retrospective studies have a risk of recall bias. In the case-control studies included in this review, women were asked to recall frequency and dose of supplements over a relatively short period of exposure occurring between 13 months and 3 years prior to the interview. Compounding the risk of recall bias is the relatively widespread knowledge of the protective effect of folic acid supplementation on NTDs. Studies that compare cases with NTD malformations with controls with non-NTD malformations may have a lower risk of differential recall bias. Two such studies showed a clear and consistent protective effect of folic acid supplementation but were based on data collected in the prefortification era.^95,99

Both of the risks of bias described above (case ascertainment and recall) will reduce the differences between study groups. A further issue with the included study designs is the relatively rarity of the outcome and the difficulty of adequately powering studies to determine benefits.

Figures 4 and 5 present the results in a forest plot of the largest or more representative study from each data source. Figure 4 demonstrates that the older studies (with one exception) show a protective effect of folic acid supplementation on NTDs, and the newer studies, all conducted after the introduction of food fortification, do not show a protective effect. The same studies, when sorted by study design (Figure 5), show greater consistency in direction of effect in the cohort studies and greater variation in direction of effect in the case-control studies.

Figure 4

Folic Acid Supplementation and Neural Tube Defects by Earliest Year of Recruitment: Forest Plot.

Figure 5

Folic Acid Supplementation and Neural Tube Defects by Study Design: Forest Plot.

Although the incidence of NTDs has declined in conjunction with food fortification, 23 percent of U.S. women have suboptimal RBC folate concentrations.⁴⁵ Among women who consume supplements, the proportion is less than 10 percent; among women who do not consume supplements, it is 28 percent. Nearly 33 percent of women who consume mandatorily fortified foods alone (i.e., they do not consume folic acid supplements or voluntarily fortified foods) have suboptimal RBC folate concentrations. These statistics suggest a continued and important role for supplements.

Three case-control studies provide information about the effects of folic acid supplementation by racial/ethnic and other maternal characteristics.^11,82,87 One study suggested that folic acid supplementation may be less protective for Hispanic women, a second showed a higher risk of NTDs among Hispanic women, and the third did not show an effect. None of the studies are conclusive because of small numbers that could have resulted in chance findings.

One cohort study, set in New England (1984 to 1987) and described in two publications,^97,98 and six case-control studies^{11,22,82,87,96,99} provided evidence on dose and timing. All four studies on dose (one cohort⁹⁸ and three case-control studies^82,96,99) predate food fortification; none show a dose-response effect. Notably, the number of cases for varying levels of dosage was small. The five studies (one cohort⁹⁷ and four case-control studies^11,22,82,87) reporting on timing of folic acid supplementation did not consistently compare the same timing of exposure. The two older studies did not find statistically significant effects of folic acid supplementation by timing of supplementation.^22,82 One newer study, conducted entirely in the postfortification era, found more protective effects for women who started before pregnancy compared with during the first month of pregnancy for anencephaly; the protective effect of early timing of exposure did not appear to hold for spina bifida.⁸⁷ The other new study, focusing on spina bifida only, did not find a statistically significant effect of timing of folic acid supplementation on the odds of spina bifida.¹¹

Evidence for Harms of Folic Acid Supplementation

We included one RCT comparing folic acid supplementation with a multivitamin versus trace elements described in seven publications^88-94 that evaluated the harms of folic acid supplementation in the periconceptional period to prevent NTDs. We also included one fair-quality cohort study⁸³ of women with and without folic acid supplementation use that met our inclusion criteria. Although we could not rule out risk of an increase in higher-level multiple gestations (triplets or greater) in the RCT^88-90,93,94 due to a limited number of events, analyses focused on twinning among women with or without treatment with fertility drugs (clomiphene citrate) were reassuring, with point estimates and CIs that did not indicate a risk for multiple gestations. Among a general population of women in the RCT, there was no evidence of a higher risk of twin pregnancies or twin births (combination of live births and stillbirths). When the analyses were limited to only women who completed full or partial folic acid supplementation, there was no evidence of a higher risk of twin pregnancies. Suggestion of a higher likelihood of twin births (live births and stillbirths) among women who completed full or partial folic acid supplementation compared with those who completed full or partial trace element supplementation should be interpreted cautiously, given the similarity in point estimates and overlap in the calculated 95% CIs. Findings from the observational study⁸³ support the findings from the RCT. After adjustment for in vitro fertilization, the authors found no association between periconceptional folic acid supplement use and twinning.

Comparison of maternal symptoms between the multivitamin with folic acid and trace elements groups in the RCT^88-90,93,94 suggested a potential higher risk of maternal weight gain, diarrhea, and constipation at 12 weeks of gestation. These symptoms are common in pregnancy and, thus, limit our ability to draw definitive conclusions about the association of folic acid supplementation and these symptomatic outcomes.

With regard to childhood respiratory illness, three meta-analyses^101,108,109 provided no evidence of a higher risk of child asthma or other allergen-related illness (e.g., atopy, eczema, and atopic dermatitis) and no consistent variation in these outcomes by timing or dose of exposure.¹⁰¹

Applicability of Evidence

Most of the studies included in this review are applicable to primary care. One cohort study of women undergoing alpha-fetoprotein screening or amniocentesis is representative of older pregnant women but not all women.⁹⁷ The modal age for the cohort study ranged from 30 to 39 years; in contrast, the average age of women in other studies was in the 20s. A majority of studies reported that they excluded NTD cases with multiple malformations or known syndromal causes of NTDs.

Differences by Age

Studies consistently found that among women of reproductive age, supplemental folic acid intake increases with age. PRAMS 2009 found that rates of folic acid consumption (multivitamins, prenatal vitamins, or a folic acid supplement every day of the month before pregnancy) increased as follows: 42.4 percent among women ages 35 to 55years ; 34.5 percent among women ages 25 to 34 years; and 16.1 percent among women ages 18 to 24 years (p<0.05).⁴⁸ The National Birth Defects Prevention Study (1997 to 2005) found rates of compliant folic acid use (defined as ≥5 times per week during the 3 months before conception) as follows: 7.23 percent among women age 19 years or younger, 16.18 percent among women ages 20 to 24 years, 39.30 percent among women ages 25 to 34 years, and 51.96 percent among women age 35 years and older.¹¹³

The March of Dimes surveys and NHANES also provide data on folic acid consumption among women of reproductive age but do not focus on consumption prior to pregnancy. These studies also found that use of supplemental folic acid increases with age.^111,116,117

Differences by Race/Ethnicity

Significant improvements in RBC folate status have occurred among all racial/ethnic groups; among non-Hispanic black women, the prevalence of low RBC folate declined from 59.6 percent between 1998 and 1994 to 12.1 percent between 1999 and 2000. During the same time period, the prevalence of low RBC folate declined from 34.5 to 4.5 percent among non-Hispanic white women and from 38.7 to 1.6 percent among Mexican American women.⁴⁹

Although all women of childbearing age increased their median total folate intake by at least 100 μg/day since fortification, increases were larger for whites than for blacks and Mexican Americans.¹¹⁸ White women were also more likely to have reached the 400 μg/day threshold both pre- and postfortification (30% and 39%, respectively) than black women (20% and 26%, respectively) and Mexican American women (17% and 28%, respectively).¹¹⁸

Consumption of supplements containing folic acid also varies by race/ethnicity and other characteristics. According to the 2007 March of Dimes survey, women who were nonwhite, ages 18 to 24 years, had less than a high school education, or had a household income less than $25,000 were the least likely to report daily consumption of a supplement containing folic acid.¹¹⁹ Studies consistently found that among women of reproductive age, supplemental folic acid intake is higher among whites than blacks and among non-Hispanics than Hispanics. The National Birth Defects Prevention Study found higher rates of compliant folic acid supplement use among white women (43.73%) compared with black (16.89%), Hispanic (10.87%), or other women (27.98%).¹¹³ PRAMS 2009 found that folic acid consumption was highest among white women (34.2%), followed by other race/ethnicity (33.0%), Hispanic (22.5%), and black women (19.5%) (p<0.05).⁴⁸

The March of Dimes 2008 survey, NHANES,¹¹¹ and the California Women's Health Survey¹²⁰ found similar patterns by race/ethnicity. NHANES 2003 to 2006 data found non-Hispanic whites were most likely to take supplemental folic acid (37.2%), followed by other race/multiracial (24.7%), Mexican Americans (20%), and non-Hispanic blacks (20%).¹¹¹ The California Women's Health Survey 2006 found more white women reported daily use of folic acid containing supplements (50.6%) compared with blacks (39.5%) and Asians (40%). Non-Hispanic women were more likely to report folic acid supplement use (46.9%) than Hispanic women (30.2%).¹²⁰

The March of Dimes conducted a survey of Spanish-language–dominant Hispanic women in 2008. Overall, 21 percent of women took folic acid daily (multivitamins, folic acid supplements, or prenatal vitamins). Rates of daily folic acid supplement use varied by ancestry: Mexican (19%), Central American (22%), South American (35%), and Caribbean/other (25%).¹¹⁷

Differences by Education

Studies consistently found that the use of supplemental folic acid increases with education. The National Birth Defects Prevention Study found rates of compliant folic acid supplement use as follows: 9.04 percent among women with less than a high school education; 17.36 percent among women with a high school education; 30.98 percent among women with 1 to 3 years of college education; and 58.20 percent among women with 4 or more years of college education.¹¹³ State-level PRAMS data from Rhode Island¹¹⁵ and Texas¹¹⁴ found similar patterns. In Rhode Island (PRAMS 2004 to 2008), daily multivitamin use was as follows: 25.1 percent among women with less than a high school education; 23.1 percent among women with a high school education; and 44.5 percent among women with more than a high school education (p<0.0001). In Texas (PRAMS 2002 to 2010), no daily multivitamin use was highest among women with less than a high school education (80.3%) and high school graduates (82.2%) compared with those with more than a high school education (66.6%). Differences were statistically significant in multivariate models that included age, race/ethnicity, education, insurance status , Medicaid coverage, and pregnancy intention. March of Dimes surveys^111,116,117 and the California Health Information Survey¹²⁰ also found that folic acid supplement consumption increases with education.

Differences by Income

Studies consistently found that supplemental folic acid use increases with income. The National Birth Defects Prevention Study found rates of compliant folic acid supplement use as follows: 10.07 percent for households making less than $10,000; 26.05 percent for households making $10,000 to $50,000; and 55.21 percent for households making more than $50,000.¹¹³ Rhode Island PRAMS data (2004 to 2008) found a similar pattern: multivitamin use prior to pregnancy was 21.1 percent for households making less than $25,000; 30.2 percent for households making $25,000 to less than $50,000; and 50.5 percent for households making $50,000 or more (p<0.0001).¹¹⁵ March of Dimes surveys,^116,117 NHANES,¹¹¹ and other surveys¹²⁰ also found that folic acid supplement consumption increases with household income.

Differences by Insurance Status

State-level analyses of PRAMS data found differences in supplemental folic acid use by insurance status. Rhode Island PRAMS data (2004 to 2008) found significant differences in folic acid supplement use prior to pregnancy as follows: 17.3 percent among women with no insurance; 25.5 percent among women with public insurance; and 44 percent among women with private insurance (p<0.0001).¹¹⁵ Texas PRAMS data (2002 to 2010) found no daily multivitamin use was higher among women without health care coverage before pregnancy (83.2%) compared with those with coverage (67.9%) and among women on Medicaid (82.7%) compared with those not on Medicaid (64.7%). Differences were statistically significant in multivariate models that included age, race/ethnicity, education, insurance status, Medicaid coverage, and pregnancy intention.¹¹⁴

Differences by Employment Status

The National Birth Defects Prevention Study found higher rates of compliant folic acid supplement use among women who were employed (35.47%) versus those who were unemployed (24.93%).¹¹³

Differences by Marital Status

Rhode Island PRAMS data (2004 to 2008) found higher supplemental folic acid use prior to pregnancy among married women (44%) compared with nonmarried women (21.2%) (p<0.0001).¹¹⁵

Differences by Pregnancy Intention

State-level analyses of PRAMS data found that women intending pregnancy have higher rates of supplemental folic acid use than women not intending pregnancy. Rhode Island PRAMS data (2004 to 2008) found significant differences in folic acid supplement use prior to pregnancy by pregnancy intention (44.5% among women intending pregnancy, 19.4% among women not intending pregnancy; p<0.001).¹¹⁵ Texas PRAMS data (2002 to 2010) found that no daily multivitamin use was more common among women not intending pregnancy (85.6%) than among those intending pregnancy (66.5%). Differences were statistically significant in multivariate models that included age, race/ethnicity, education, insurance status, Medicaid coverage, and pregnancy intention.¹¹⁴

The National Birth Defects Prevention Study found higher rates of compliant folic acid supplement use among women not using birth control (33.51%)—who may be intending pregnancy—than among women using birth control (29.48%).¹¹³

Differences by U.S. Versus Foreign Born and Years in the United States

The National Birth Defects Prevention Study found higher rates of compliant folic acid supplement use among women born in the United States (35.94%) than among foreign-born women (18.5%). The 2008 March of Dimes survey of Spanish-language–dominant Hispanic women found rates of daily folic acid supplement use varied by number of years in the United States (<5 years, 10%; 5 to 10 years, 19%; ≥10 years, 24%; born in the United States, 18%).¹¹⁷

Differences by Parity

Rhode Island PRAMS data (2004 to 2008) found differences in folic acid supplement use prior to pregnancy by parity: 37.1 percent among women having their first birth and 32.6 percent among women having their second or higher birth (p=0.0012). NHANES 2003 to 2006 reports folic acid supplement use by parity: 32.4 percent for women with no live births, 30.3 percent for women with one live birth, and 31.3 percent for women with two or more live births.¹¹¹ The 2008 March of Dimes survey of Spanish-language–dominant Hispanic women found that daily folic acid supplement use was lowest among women who had never been pregnant (12%) compared with women who had been pregnant in the past 2 years or who were currently pregnant (20%), last pregnant 3 to 4 years ago (26%), or last pregnant 5 or more years ago (23%).¹¹⁷

Differences by Health Behaviors/Health Status

The National Birth Defects Prevention Study found higher rates of compliant folic acid supplement use among nonsmokers (35.13%) versus smokers (21.93%) and among women who had any alcohol intake in the 3 months prior to conception (38.01%) compared with no alcohol intake (27.65%).¹¹³ NHANES 2003 to 2006 found use of folic acid supplements was lower among women who have diabetes (17%) than among women who do not have diabetes (32.1%).¹¹¹

Diabetes

Offspring of women with pregestational diabetes have a two- to fourfold increased risk of a wide range of birth defects. Prior studies in humans and animal models show that glucose control is an important prevention method. Additionally, studies suggest that micronutrient levels, including folic acid, may play a key role in the pathogenesis of NTDs and other birth defects. Larger epidemiologic studies in this area have focused primarily on the joint effects of diabetes and obesity and folic acid intake to better determine the association of folic acid supplementation across key groups: no diabetes or obesity and folic acid use (reference group), no diabetes or obesity and no folic acid intake, diabetes and/or obesity and folic acid use, diabetes and/or obesity and no folic acid use (highest risk group). One of the largest population-based studies, the Birth Defects Prevention Study (1997 to 2004),^28,121 identified 14,721 cases (infants with cardiac or noncardiac birth defects, including spina bifida and anencephaly) and 5,437 controls and assessed the joint effects of maternal diabetes and folic acid consumption on birth defect development. Periconceptional folic acid supplementation was defined as intake in the month before conception or during the first 3 months of pregnancy. Women with pregestational diabetes and no use of folic acid supplementation had an increased odds of birth defects overall, but the findings for specific birth defects must be interpreted with caution because they are limited by a small number of cases. Regarding spina bifida, there was a nonsignificant twofold increase (OR, 2.37 [95% CI, 0.21 to 26.65]) in the risk of spina bifida among women with diabetes and no folic acid supplementation compared with women without diabetes who were taking folic acid supplements (reference group). There was a higher odds of anencephaly in this same group (OR, 31.56 [95% CI, 4.98 to 199.94]).

The Slone Epidemiology Birth Defects Study¹²² assessed the effect of periconceptional folic acid supplementation on NTDs within the setting of diabetes and obesity. Folic acid intake was ascertained through maternal interviews conducted within 6 months of delivery and was operationalized for analysis as 400 μg/day or more versus less than 400 μg/day. Spina bifida was more likely to occur in women who had diabetes (0.7% vs. 0.4%) or obesity (19% vs. 10.8%) than in those without either condition. In analyses stratified by folic acid use, pregnancies with diabetes and less than 400 μg/day of folic acid had a statistically significant odds of NTDs (aOR, 3.95 [95% CI, 1.56 to 10.00]). Among pregnancies with appropriate folic acid supplementation and diabetes, there was a lower but statistically significant odds of NTDs (aOR, 1.31 [95% CI, 0.7 to 10]).

A smaller study¹²³ among pregnant women with diabetes (n=31) and without diabetes (n=54) found no difference in dietary, serum, or RBC folate concentrations between the two groups, after adjusting for the extent of folic acid supplementation. These findings suggest that there is no difference in folic acid metabolism among women with and without diabetes, but the findings will need to be confirmed in a larger study. Additionally, while the level of glycemic control in early pregnancy is hypothesized to influence development of cardiac defects, the role of glycemic control on the conversion of homocysteine to methionine, the pathway thought to be most important to the development of NTDs, is not known, and there was no adjustment for glucose levels or the degree of glucose control in the analysis. Finally, it may be that the NTDs among women with diabetes may occur through multiple pathways.

Maternal Obesity

The epidemic of maternal obesity in the United States has drawn attention to whether the recommended daily dose of periconceptional folic acid supplementations is sufficient for women with an elevated BMI. It is unclear, however, whether lower concentrations of folate in overweight/obese pregnant women are associated with a higher risk of NTDs and, therefore, whether the effect of folic acid supplementation varies with obesity. Although there are no recommendations specifically for overweight or obese women in the United States, some countries have recommended higher dosages up to 5 mg daily for overweight women. Prior studies have reported an inverse association between BMI and concentrations of micronutrients, including folic acid. Lower folic acid levels have been documented with increasing BMI, although a direct effect on the incidence of NTDs has not been observed. In one study,¹²⁴ serum folate concentrations and dietary intake of folate were assessed across BMI quartiles at mid- and late pregnancy among 802 and 660 women, respectively. A statistically significant association between BMI and folate concentration was reported at mid-pregnancy (p=0.001 for trend). While these data suggest a decrease in folic acid levels during pregnancy, the time frame is mid- to late pregnancy, which is well beyond the first 28 days of gestation, in which maternal folate concentrations prevent NTDs.

Folic Acid Antagonists and Seizure Medications

We did not find conclusive evidence that the effect of folic acid supplementation on prevention of NTDs varies with the use of folic acid antagonists or seizure medications. Jentink and colleagues¹²⁵ conducted a registry-based case-control study of infants with spina bifida (cases) compared with infants with other congenital malformations unrelated to folic acid metabolism (controls) to determine the effect of periconceptional folic acid supplementation (4 weeks prior to conception through the first 8 weeks of pregnancy) on the risk of NTDs among women with epilepsy and treatment with valproic acid or carbamazepine. Over the study time period, 11,864 pregnancies with congenital malformations (197 pregnancies with spina bifida) were identified. Sixty-six women were exposed to antiepileptic drugs. There was no statistically significant effect of folic acid supplementation on pregnancy outcomes in women exposed to valproic acid or carbamazepine (OR, 0.9 [95% CI, 0.2 to 4.3]). In a subsequent analysis, Jentenk¹²⁶ examined the effect of folic acid supplementation on NTDs in pregnancies exposed and unexposed to valproic acid alone. Among pregnancies unexposed to valproic acid, there was a 50 percent reduction in NTDs with folic acid supplementation (OR, 0.5 [95% CI, 0.3 to 0.7]). Among pregnancies exposed to valproic acid, there was no effect of folic acid supplementation on the incidence of NTDs (OR, 1.0 [95% CI, 0.1 to 7.6]). There is little evidence to suggest that folic acid has a diminished protective effect on NTDs in women taking antiepileptic drugs and essentially no data on the effect of various dosages of folic acid supplementation on NTDs. First trimester exposure to valproic acid is associated with a 1 to 2 percent risk of NTDs, but the mechanism by which NTDs occur in the setting of valproic acid exposure is unclear, and further investigations are needed to better understand the biological pathways in which valproic acid may affect neural tube closure.

Several folic acid antagonists were evaluated in a large case-control study by Hernandez-Diaz.⁹⁵ Reported findings indicate a higher likelihood of NTDs (OR, 2.8 [95% CI, 1.7 to 4.6]) with exposure to any one of several folic acid antagonists (including carbamazepine, phenobarbital, phenytoin, primidone, sulfasalazine, triamterene, and trimethoprim) in the first or second month of pregnancy compared with no exposure, even after adjusting for maternal characteristics, reproductive history, and folic acid supplementation. Subgroup analyses show a higher adjusted odds of NTDs with exposure to carbamazepine (aOR, 6.9 [95% CI, 1.9 to 25.7]) or trimethoprim (aOR, 4.8 [95% CI, 1.5 to 16.1]). These findings should be interpreted with caution because the number of cases identified, particularly in subgroup analyses, was small. No studies were identified that evaluated the impact of folic acid supplementation in the setting of methotrexate exposure, a folic acid antagonist used in the treatment of ectopic (i.e., tubal) pregnancy.

Prior NTDs

Meta-analysis findings¹²⁷ indicated that periconceptional folic acid supplementation was effective in reducing the risk of recurrent NTDs in the included studies of women with prior pregnancies complicated by NTDs. Evidence from three studies^128-130 (n=1,650 total pregnancies) of folic acid (doses ranging from 360 μg/day^102,128 to 4 mg/day^129,130) with a multivitamin (10/817 cases) compared with multivitamin alone or iron and calcium alone (32/833 cases) indicated a 67 percent reduction in NTDs (RR, 0.33 [95% CI, 0.17 to 0.66]); folic acid of 400 μg/day combined with a multivitamin (10/817 cases) compared with multivitamin alone (32/833 cases) indicated a 67 percent reduction in NTDs (RR, 0.33 [95% CI, 0.17 to 0.66]). Heterogeneity of the pooled analysis was low (χ²=1.25; I²=0.0%; p=0.74). The pooled RR was largely driven by the Medical Research Council Study, which had a total of 27/1,195 cases. Individual RR estimates, however, were consistent and statistically significantly associated with a reduction in the recurrence of NTDs across all three studies (range, 0.29 to 0.42).