Overview

We found a total of 20 publications on the benefits of folic acid supplementation. Seven publications present results of the only eligible RCT.^88-94 The trial, conducted in Hungary, is an RCT that was initiated in 1984 and terminated in 1992, with information collected through 1993. Three publications relate to two cohort studies; one was a Hungarian cohort study of women recruited between 1993 and 1996,⁸¹ and the second was a cohort drawn from women who underwent alpha-fetoprotein screening or amniocentesis between 1984 and 1987.^97,98 All other studies were case-control studies and compared NTD cases with nonmalformed infants^{9,11,22,82,87, 96} or infants with non-NTD malformations.^95,99 Additionally, we drew on information from two publications in the previous update.^36,86

We present information from the RCT first, followed by the cohort studies and the case-control studies. Although the RCT and the cohort studies potentially offer greater control for potential sources of bias, they predate mandatory food fortification. The case-control studies span a period ranging from 1976 through 2008, including several relying exclusively on data collected after food fortification. Because these eight publications of case-control data draw from related, or in some cases, subsets of the same data, we present them by the broadest data source first (national or multistate followed by two-state or single-state studies) and recency of data collection within each section. Table 3 provides supporting descriptions of each study. Tables 4 and 5 provide results. Because of the heterogeneity across studies and the differences in food fortification over time, we did not pool the results.

Table 3

Study Characteristics of Studies on the Effect of Folic Acid Supplementation on Neural Tube Defects.

Table 4

Results of Prospective Studies on the Effect of Folic Acid Supplementation on Neural Tube Defects.

Table 5

Results of Retrospective Studies on the Effect of Folic Acid Supplementation on Neural Tube Defects.

Study Characteristics of the Included RCT

One RCT, described in seven publications,^88-94 randomized women to a vitamin supplement containing folic acid (0.8 mg folic acid and 12 vitamins, four minerals, and three trace elements) or a trace-element supplement (copper, manganese, zinc, and a low dose of vitamin C). Women started the supplement at least 28 days before conception and continued at least until the date of the second missed menstrual period.⁹¹ The trial, as part of the Hungarian Optimal Family Planning Programme, excluded women with delayed conception and infertility or with ongoing pregnancies. For the first 4 years, the program also excluded women older than age 35 years or with a prior wanted pregnancy.⁹⁰ The trial involved repeated contact with women at regular intervals. Women were asked to visit the clinic immediately after the first missed menstrual period. The staff administered a sensitive serum pregnancy test, followed within 2 weeks by an ultrasound.⁹³ As a result, the authors note that they had “nearly total ascertainment of unsuccessful pregnancy outcomes, including fetal deaths and malformations.”⁹³ The trial included only “informative” cases in the analysis; that is, live births, terminations in the second trimester, and stillbirths (late fetal deaths). It did not account for loss to followup, which constituted 0.9 percent of each arm (26 cases of 2,819 confirmed pregnancies in the multivitamin arm and 23 cases of 2,863 confirmed pregnancies in the trace-element supplement arm), or loss due to first-trimester losses, chemical pregnancy, ectopic pregnancy, or miscarriage (395 cases [14%] in the multivitamin arm and 504 cases [17.6%] in the trace-element supplement arm).

The trialists ascertained compliance with the supplement by 1) asking women; 2) checking women's record of supplement use, recorded daily with basal body temperatures; and 3) checking boxes of supplements for unused tablets.⁹⁰ Women who became pregnant before starting the supplement or during the first month were considered unsupplemented and were referred to prenatal care immediately. Women who became pregnant after a period of supplementation were referred at 12 weeks to prenatal care.

Results of the Included RCT

The trial reported no cases of NTDs in the experimental arm and 6 cases in the control arm. Based on a denominator of “informative” cases only (live births, stillbirths, and second-trimester terminations only; counting each of twin and triplet births separately), the p-value for the Fisher exact test was 0.014. We calculated the Peto odds ratio (OR) as 0.131 (95% CI, 0.0263 to 0.648; p=0.013).

Study Characteristics of Included Cohort Studies

At the conclusion of the RCT described above, no additional RCT was considered ethically possible. The authors continued their investigation using the same intervention (multivitamin supplement containing 0.8 mg of folic acid) in women drawn from the Hungarian Periconceptional Service (1993 to 1996), with supplementation provided before conception.⁸¹ The comparison group comprised unsupplemented pregnant women at their first visit in the regional antenatal care clinic between the 8th and 12th week of gestation; women who were determined (on a one-page, personally administered questionnaire) to have taken multivitamins or folic acid supplements during the periconceptional period were excluded. Unexposed women were matched to exposed women for age, socioeconomic status, employment status, and residence during the first year of pregnancy. Informative offspring included malformed fetuses, antenatally diagnosed and terminated in the second or third trimester; stillborn fetuses (late fetal death after the 28th week of gestation and/or weighing >1,000 g); and live-born infants. Informative offspring were ascertained in three ways: 1) by antenatal diagnoses of terminated fetuses, supported by a description of pathology; 2) through records at birth; and 3) by examination by a blinded pediatrician at 1 year of age or from pediatricians' records. NTDs included anencephaly and spina bifida. The study had a potential risk of selection bias because women in the supplemented cohort had a higher rate of comorbid conditions and unsuccessful pregnancies. The latter, in particular, likely prompted them to seek entry into the Hungarian Periconceptional Service. Supplemented women, however, were likely to have planned their pregnancies, had healthier behaviors in the periconceptional period, and received better prenatal care than unsupplemented women.

We also identified an eligible cohort study based on 23,491 women undergoing alpha-fetoprotein screening or amniocentesis between 15 and 20 weeks of gestation (1984 to 1987).^97,98 Most of these women lived in Boston (33%), elsewhere in Massachusetts (48%), elsewhere in New England (5%), and outside New England (14%). Most of the samples were analyzed at the Boston University School of Medicine facilities. Nurses contacted women at the time that their tests were received by the laboratory; 93 percent did not know the results of their tests at the time of the interview. Nurses then asked women to recall their use of multivitamins in the first 3 months before pregnancy and the first 3 months of pregnancy. Exposure was defined as the use of at least one multivitamin containing folic acid per week between weeks 1 and 6 following conception. The maximum period of recall was 8 months (3 months prepregnancy and 5 months of pregnancy).

Results of Included Cohort Studies

The Hungarian cohort study reported 1 case of an NTD in 3,056 supplemented women and 9 cases in 3,056 unsupplemented women.⁸¹ The selection bias arising from higher rates of unsuccessful pregnancies in the supplemented cohort likely biased the results toward the null, while the selection bias from intentional pregnancies likely biased the results toward an effect of the intervention on NTDs. The authors adjusted the OR for birth order, chronic maternal disorder, and history of previous unsuccessful pregnancies. The study reported an adjusted OR (aOR) of 0.11 (95% CI, 0.011 to 0.91; p-value not reported).

The New England study reported 10 cases of NTDs among 10,713 women who took multivitamins containing folic acid in weeks 1 through 6 compared with 11 cases of NTDs among 3,157 women who did not take any supplements (OR, 0.27 [95% CI, 0.11 to 0.63]). By contrast, use of multivitamins containing folic acid from week 7 onward had no statistically significant effect on NTDs (25 cases in 7,883 supplemented women vs. 11 cases in 3,157 unsupplemented women; OR, 91 [95% CI, 0.45 to 1.80]) when compared with nonuse.⁹⁷

Study Characteristics of Case-Control Studies

Results of Included Case-Control Studies

Despite differences in definition of exposure, comparison, and timing across the two National Birth Defects Prevention Study publications, both are consistent in demonstrating a lack of effect of folic acid supplementation on benefits (aOR for anencephaly and spina bifida, 0.93 [95% CI, 0.82 to 1.06] and aOR for anencephaly, 1.2 [95% CI, 0.8 to 1.9], respectively).^9,87 A potential explanation for the findings from this surveillance-based database is that in the postfortification era, the majority of cases of NTDs arising from folate deficiency have been averted, and the remainder of the cases represent other potential etiologies. A second explanation is that these findings could have arisen from bias. The National Birth Defects Prevention Study is a surveillance-based database in which eight of 10 sites recorded fetal deaths and elective pregnancy terminations in addition to live births. This case ascertainment approach mitigates the risk of bias from selection that otherwise occurs in studies focusing on live births only, where potentially eligible cases (i.e., fetal deaths and elective pregnancy terminations attributable to NTDs) are lost to analysis. If folic acid supplementation is protective, a sample that is selectively missing women who do not use folic acid supplements and have NTD-affected pregnancies that end in terminations or stillbirths will have higher odds of NTDs with folic acid supplementation than a sample without selection bias. The risk of recall bias, however, is a concern with all retrospective studies. An additional risk of differential recall bias may occur if study participants are generally aware of the potential benefits of folic acid and case mothers systematically overreport its use. A “yes/no” categorization of folic acid supplementation further risks misclassifying exposure.⁹ One publication attempted to address recall bias by focusing on consistent use,⁸⁷ but the risks stemming from recall over the course of up to 3 years persist.

The 2011 Slone Birth Defects Study found no effect of folic acid supplementation on the risk of spina bifida, regardless of the level of supplementation. Consistent users, when compared with nonusers, had an adjusted odds of 1.11 (95% CI, 0.74 to 1.65). Early pregnancy initiators had an adjusted odds of 0.79 (95% CI, 0.54 to 1.16). Inconsistent users had an odds of 2.20 (95% CI, 0.64 to 7.62). These results could be explained by the ceiling effect—all cases of NTDs preventable through supplementation were averted by food fortification, and the remainder constitute a population with a different etiology. Alternatively, these results could arise from bias. Specifically, the sources of bias include 1) differential recall of supplementation, particularly in an era with more widespread knowledge of the support and claims for the use of folic acid supplementation in pregnancy, and 2) selection bias from incomplete case ascertainment, because the study did not consistently include terminated spina bifida cases, which were available consistently from only one site.

The two prefortification Slone Birth Defect Studies with overlapping time periods consistently demonstrate that daily use of supplements reduces the risk of NTDs compared with nonuse (aOR, 0.7 [95% CI, 0.5 to 0.8] in the 2001 study;⁹⁵ adjusted relative risk [RR], 0.6 [95% CI, 0.4 to 0.8] in the 1993 study⁹⁹). The Slone Birth Defect Study, particularly in the early years, had a potential risk of selection bias by not including stillbirths and elective terminations. Later rounds of analysis included a more complete case ascertainment process. The risks of recall bias were somewhat mitigated by having a shorter recall period and a calendar aid highlighting the woman's last menstrual period. Additionally, these two studies attempted to correct for the issue of differential recall of periconceptional exposure in cases and controls^95,99 by comparing cases of NTDs with controls of other malformed infants. The 1993 study found a much higher rate of knowledge of the folic acid hypothesis among NTD case mothers than among control mothers of other malformed infants (74/432 [17%] vs. 65/2,561 [2.5%]), suggesting that the knowledge of the hypothesis could skew recollections of folic acid supplement intake.⁹⁹ However, all cases in this study belong in the prefortification era and do not address the effect of folic acid supplementation in the current environment.

Notably, the 2001 study also offered, through the analysis of the effect of folic acid antagonists, a perspective on what would occur when women are folate deficient because of a folate antagonist. The study found that the adjusted odds of NTDs among women exposed to folic acid antagonists was 2.8 (95% CI, 1.7 to 4.6).

The study of Mexican Americans, spanning the pre- and postfortification era, found a nonsignificant reduction in the odds of NTDs associated with daily consumption of multivitamins containing folic acid (0.77 [95% CI, 0.19 to 3.22]); when adjusted for maternal age, education, obesity, and previous stillbirth or miscarriage, the direction of effect altered (aOR, 1.12 [95% CI, 0.22 to 5.78]; p-value not reported).²² Of note are the extremely low levels of folic acid supplement use in both arms (3 cases of daily use in the 3-month preconceptional period vs. 66 cases of no use in the 6-month periconceptional period among cases; 4 cases of daily use in the 3-month preconceptional period vs. 68 cases of no use in the 6-month periconceptional period among controls).

Two other studies were conducted in the prefortification era. Both studies drew on data from the California Birth Defects Monitoring Program, using cases from 1989 to 1991⁸² and 1985 to 1987. The Shaw et al study found an OR of 0.65 (95% CI, 0.45 to 0.94) for any use in the 3 months before conception. A larger analysis comprising women reporting supplement use in the 3 months before and after conception found an OR of 0.6 (95% CI, 0.46 to 0.79). The NICHD Neural Tube Defects Study, using a combination of slightly older California data (1985 to 1987) and Illinois data (also from 1985 to 1987), reported no effect of supplements on NTDs (calculated OR, 1.00 [95% CI, 0.73 to 1.40]; p=0.97).⁹⁶ The Shaw et al study was able to ascertain the status of approximately 88 percent of eligible cases and controls.⁸² By contrast, the case ascertainment of the Mills et al study was estimated, based on a re-evaluation of the likely prevalence, to be as low as 43 percent.⁸²

Study Characteristics

Three case-control studies provide limited information about the effects of folic acid supplementation by racial/ethnic and other maternal characteristics.^11,82,87 Table 3 presents study characteristics and Table 6 provides results. The Slone Birth Defects Study provides the most recent data (1998 to 2008).¹¹ In this study, mothers of infants with and without birth defects were interviewed within 6 months of delivery about pregnancy exposures, including details of diet and vitamin intake. Periconceptional folic acid supplementation and dietary folate consumption were compared between 205 mothers of spina bifida cases and 6,357 mothers of nonmalformed controls. Women who reported folic acid supplement use of at least 4 days per week during at least 2 of the 3 periconceptional months (2 months before to 2 months after last menstrual period) were considered to be “consistent users.” A second case-control study analyzed the data from 1998 to 2003 from the National Birth Defects Prevention Study.⁸⁷ It used logistic regression to compute crude and aORs between cases and controls assessing maternal periconceptional use of folic acid supplements and intake of dietary folic acid. The third case-control study used data from the California Birth Defects Monitoring Program (1989 to 1991). Mothers of 549 cases and 540 controls were interviewed about vitamin supplements used in the 3 months before or after conception.⁸²

Table 6

Variations in the Effect of Folic Acid Supplementation on Neural Tube Defects by Race/Ethnicity.

Results

The Slone Birth Defects Study found that in the setting of folic acid fortification, folic acid supplementation does not appear to offer further benefit for reducing spina bifida risk.¹¹ Women who reported taking folic acid supplements at least 4 days per week during the months before neural tube closure did not have decreased risk of spina bifida compared with women who reported no supplementation. The lack of protective relationship was observed for white women. The study found a possible increased risk of spina bifida among consistent supplement users of Hispanic ethnicity compared with nonusers (aOR, 2.20 [95% CI, 0.98 to 4.92]); however, the authors note this finding may be due to chance.

The National Birth Defects Prevention Study found that periconceptional supplement use did not reduce the risk of having a pregnancy affected by an NTD, and there were no differences in the effects of folic acid supplementation by race/ethnicity.⁸⁷ Supplement use-race interactions were not significant for anencephaly (p=0.57) or spina bifida (p=0.08). However, the authors note that the number of cases among non-Hispanic black and Hispanic populations were relatively small, so findings should be interpreted with caution.

The California Birth Defects Monitoring Program found that women who used any folic acid–containing vitamin in the 3 months before conception had a lower risk of having an NTD-affected pregnancy.⁸² Reduction in risk for Hispanics was of smaller magnitude (OR, 0.96 [95% CI, 0.44 to 2.10]) than that observed for non-Hispanic whites (OR, 0.62 [95% CI, 0.35 to 1.10]) and blacks (OR, 0.54 [95% CI, 0.09 to 3. 20]), but these results were not statistically significant and could have occurred due to chance.

Although a study focusing on Mexican Americans²² does not provide information about differences by race/ethnicity, it provides an estimate of effect among Hispanic women, albeit in a limited geographical context. When adjusted for maternal age, education, obesity, and previous stillbirth or miscarriage, the OR was 1.12 (95% CI, 0.22 to 5.78; p-value not reported).

Study Characteristics

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 information on the effect of dosage and timing of folic acid supplementation on NTDs. Of these, the most recent case-control studies drew from the Slone Birth Defects Study (1998 to 2008)¹¹ and the National Birth Defects Prevention Study (1998 to 2003).⁸⁷ A third, focusing on Mexican Americans along the Texas-Mexico border, was conducted between 1995 and 1999.²² Two older case-control studies drew from the California Birth Defects Monitoring Program (1989 to 1991)⁸² and the Slone Birth Defects Study (1988 to 1991),⁹⁹ respectively. The oldest case-control study, the NICHD Neural Tube Defects Study, drew from both California and Illinois (1985 to 1987).⁹⁶ Four studies (one cohort⁹⁸ and three case-control studies^82,96,99) reported on dose of folic acid supplementation. Five studies (one cohort⁹⁷ and four case-control studies^11,22,82,87) reported on timing of folic acid supplementation.

Table 3 provides further details on study characteristics. We report the cohort study first and then report on the case-control studies in order of recency.

Results

Folic Acid Supplementation Variation by Dosage

All included studies on dose predate the food fortification era (Table 7). The New England cohort study (1984 to 1987)⁹⁸ found no statistically significant differences by dose (1 to 399 DFEs, 400–799 DFEs, and ≥800 DFEs vs. none). Although authors infer that the study provides no evidence of a dose-response relationship, the number of NTDs for each dose category was low.

Table 7

Variations in Effect of Folic Acid Supplementation on Neural Tube Defects by Dosage.

The case-control study using data from the California Birth Defects Monitoring Program compared the effect of three levels of dosage (<0.4 mg, 0.4 to 0.9 mg, and ≥1.0 mg) with no folic acid supplementation in the 3 months before or after conception.⁸² The study found no differences by dose for women reporting use in the 3 months before conception. In a larger sample of women reporting use in the 3 months after conception (including those who started before conception and continued), the effect of doses below 0.4 mg or above 0.9 mg was not statistically significant compared with no use (OR for 0.4 mg, 0.99 [95% CI, 0.56 to 1.80]; OR for ≥1.0 mg, 0.92 [95% CI, 0.54 to 1.60]); only the use of 0.4 to 0.9 mg had a statistically significant effect on NTDs compared with nonuse (OR, 0.54 [95% CI, 0.41 to 0.72]). Of note, however, is the small sample size for the higher and lower doses (54 cases and controls took <0.4 mg and 75 cases and controls took ≥1.0 mg).

Data from the Slone Birth Defects Study (1988 to 1991)⁹⁹ suggest lower odds of NTDs for daily use versus less than daily use (calculated OR, 0.57 [95% CI, 0.35 to 0.93]). A supplemental analysis in the same study of differences by dosage among women who did not know the hypothesis between folic acid supplementation and NTDs failed to find a dose-response effect. An older case-control study, the NICHD Neural Tube Defects Study, drawing from cases identified in California and Illinois between 1985 and 1987, reported on the number of NTDs in women receiving the RDA from supplements compared with those receiving less than the RDA.⁹⁶ The study reported no statistically significant differences between different levels of exposure (calculated OR, 1.84 [95% CI, 0.92 to 3.71]). Of note, this study likely had problems with case ascertainment.

Folic Acid Supplementation Variation by Timing

The single cohort study (drawing on cases from 1984 to 1987 and set in New England⁹⁷) reported that using multivitamins in weeks 1 through 6 resulted in a lower odds of NTDs than using multivitamins in weeks 7 and later (10/10,731 vs. 25/7,795; calculated OR, 0.29 [95% CI, 0.14 to 0.60]) (Table 8).

Table 8

Variations in Effect of Folic Acid Supplementation on Neural Tube Defects by Timing.

Of the four case-control studies, two were set in the postfortification era,^11,87 one spanned the pre- and postfortification era,²² and one predated the food fortification era.⁸² The most recent case-control study¹¹ reported the risk of consistent use (defined as ≥4 days of use per week in 2 of 3 periconceptional months) versus initiating use in the first month of pregnancy (≥4 days per week starting in the first or second month postconception). Consistent users had a higher but statistically nonsignificant risk of spina bifida (calculated OR, 1.23 [95% CI, 0.88 to 1.73]). The second postfortification case-control study⁸⁷ reported the risk of anencephaly and spina bifida separately. Women who started folic acid supplementation use before pregnancy had a lower risk of anencephaly compared with women who started during the first month of pregnancy (calculated OR, 0.61 [95% CI, 0.40 to 0.93]). No difference was found for the spina bifida cases (calculated OR, 0.95 [95% CI, 0.71 to 1.28]).

A study of Mexican American women along the Texas-Mexico border, drawing from cases from 1995 to 1999, did not find any statistically significant differences in the odds of NTDs by preconceptional versus postconceptional use (calculated OR, 1.84 [95% CI, 0.58 to 5.86]).²²

The case-control study using data from the California Birth Defects Monitoring Program (1989 to 1991) found lower odds but wide CIs for the use of folic acid supplements in the 3 months before conception compared with no use. It also found no statistically significant effect of NTDs compared with any use of folic acid supplements in the 3 months after conception (88 cases and 98 controls 3 months before vs. 322 cases and 384 controls 3 months after conception; calculated OR, 1.07 [95% CI, 0.77 to 1.48]).⁸²

Study Characteristics

We included one RCT comparing folic acid supplementation with a multivitamin versus trace elements described in seven publications^88-94 and one cohort study (Table 9).⁸³ Additionally, the previous review also reported on twinning.^{36,86,88-90,93,94} The trial characteristics are described under KQ 1a.¹⁰¹ As noted previously, the trial included only “informative” cases in the analysis (i.e., live births and stillbirths [late fetal deaths]). The authors noted that it was generally not possible to recognize multiple gestations in miscarriages or ectopic pregnancies.

Table 9

Harms of Folic Acid Supplementation: Study Characteristics of Included Twinning Studies.

In a retrospective, population-based cohort study in Norway (N=176,042) of births from December 1998 through the end of 2001,⁸³ the use of folic acid supplements and multivitamins was ascertained using a birth notification form submitted through the Norway Birth Registry. For multiple gestations, the registry received one form for each birth. Separate notification was made for pregnancies conceived through in vitro fertilization.

Three meta-analyses^101,108,109 met our initial inclusion criteria and evaluated the effects of periconceptional folic acid supplementation on childhood respiratory illness. One meta-analysis¹⁰¹ with a low risk of bias evaluated the association of folic acid supplementation during the specified time frame of 1 month prior to pregnancy or the first 12 weeks of pregnancy with childhood asthma or wheezing and allergy-related outcomes. However, because of heterogeneity in the type of folic acid supplementation (e.g., folate, combination of folate and dietary folate) and measure of exposure during the periconceptional period, authors limited the pooled estimate to five studies (three cohort, two nested case-control)^102-106 that assessed folic acid in the periconceptional period (from the month prior to pregnancy) or first trimester. Folic acid supplementation was operationalized as “yes/no.” In the three studies that reported the dose of folic acid, the range was 400 to 600 μg/day.^103-105 In one study, the average dose was not reported but the investigators suggested it was 400 μg/day.¹⁰² Asthma or wheezing was assessed through a structured parental interview or parental completion of a medical questionnaire. Two studies reported on asthma,^105,106 two studies reported on wheezing,^102,103 and one study reported on wheezing and asthma.¹⁰⁴

A second meta-analysis¹⁰⁸ with a medium risk of bias included five published studies and data from a review of a longitudinal cohort study of folic acid supplementation and asthma. We describe the two meta-analyses noted above in Table 10. Additionally, a third eligible meta-analysis¹⁰⁹ used a subset of the evidence in the other two meta-analyses; we focus on concordance of these results with other meta-analyses.

Table 10

Harms of Folic Acid Supplementation: Study Characteristics of Included Asthma/Wheezing Studies.

Results

Twinning in Women

In an analysis of informative pregnancies in the trial,⁹¹ the proportion of twin pregnancies and twin births (live and stillbirths) was not statistically significantly different between the multivitamin and trace-element arms (Table 11). Out of the total pregnancies in the multivitamin group, 1.9 percent (46/2,421) were determined to be twin gestations compared with 1.36 percent (32/2,346) of pregnancies in the trace element group (χ²=2.13; p=0.15). The RR (1.4 [95% CI, 0.87 to 2.26]) was not statistically significantly different between the two groups. The proportion of twin births (as opposed to pregnancies) was higher in the multivitamin group (93/2,468 [3.77%]) than in the trace element group (64/2,378 [2.69%]; RR, 1.42 [95% CI, 1.01 to 1.98]).

Table 11

Results of Prospective Studies on the Associations Between Folic Acid Supplementation and Twinning.

In a further analysis for the same trial, women who were not supplemented were excluded (i.e., women who became pregnant before or during the first month of supplementation). The analysis continued to demonstrate a lack of significant difference in the risk of twin pregnancies (calculated RR, 1.5 [95% CI, 0.94 to 2.39]). The study found an increased risk of twin births in the multivitamin group compared with the trace element group (RR, 1.53 [95% CI, 1.08 to 2.16]). The increased risk in the multivitamin group may be due to several factors, including differences in maternal characteristics and pregnancy-specific or delivery-related complications. Limited information is available on differences in maternal characteristics and essentially no data are available on pregnancy-related factors. The findings should be interpreted with caution because births include live births and stillbirths, and there are no data to discern the proportion of live or stillbirths in each treatment arm.

Childhood Asthma or Wheezing and Allergen-Related Outcomes

We identified eight eligible articles,^{100,102-107,110} which were synthesized in three systematic reviews.^101,108,109 All included primary studies were observational, with attendant risks of misclassification and recall bias. We discuss pooled estimates from the meta-analyses below (Table 12). With regard to asthma, the pooled estimate from one meta-analysis¹⁰¹ focusing on the prepregnancy period through the first trimester (N not reported) found no evidence from three studies^104-106 of an association between maternal folic acid supplementation compared with no use and childhood asthma (pooled RR, 1.01 [95% CI, 0.78 to 1.30]; I²=0.00; p=0.95 and 0.73, respectively). For the combined outcomes of wheezing in infants and toddlers and asthma in children, the pooled estimate from five studies^102-106 resulted in a slightly elevated risk with the use of folic acid supplements before pregnancy or during the first trimester (RR, 1.05 [95% CI, 1.02 to 1.09]; I²=0.00; p=0.01 and 0.68, respectively).

Table 12

Results of Meta-Analyses on the Associations Between Folic Acid Supplementation and Asthma/Respiratory Illness.

A second meta-analysis evaluating any exposure from the periconceptional period through pregnancy (n=14,438)¹⁰⁸ included five studies in the pooled estimate.^100,104-107 The meta-analysis found no association between folic acid supplementation during the periconceptional period or pregnancy and the development of child asthma (OR, 1.06 [95% CI, 0.99 to 1.14]), but the authors reported wide variations in the dose of folic acid supplementation across included studies. Other allergen-related outcomes included a combination of atopy, eczema, and atopic dermatitis. One meta-analysis¹⁰¹ evaluated these outcomes for periconceptional and first trimester exposure from four studies^102-104,107 and found two reports of elevated risk from one study¹⁰² among 13 reported associations of lower respiratory tract infections (adjusted RR, 1.09 [95% CI, 1.01 to 1.15]) and hospitalizations from lower respiratory tract infections (adjusted RR, 1.24 [95% CI, 1.09 to 1.41]) among infants ages 0 to 18 months.

A third meta-analysis, using a smaller subset of studies,^100,103-105 also found no statistically significant differences in the incidence of asthma, wheezing, atopic dermatitis, eczema, or sensitization.¹⁰⁹

Other Reported Harms in Women

The Hungarian trial also reported on other harms.⁹³ The presence or absence of these harms represents potential side effects of folic acid supplementation, many of which are common pregnancy symptoms, and provides reassurance of the safety of folic acid supplementation in the preconceptional period. The Hungarian trial⁹³ reported on differences between weight gain, body weight, gastrointestinal symptoms (hunger or increased appetite, lack of appetite, heartburn and indigestion, constipation, diarrhea, irregular and/or colic defecation [urge to defecate after a meal]), and exanthema (a skin disorder characterized by a rash and skin eruptions) after periconceptional multivitamin and trace element supplementation (Figure 3). The study found no statistically significant differences in the report of most of these symptoms between the two groups from before pregnancy through pregnancy confirmation. Women who continued supplementation through the first 12 weeks of pregnancy had an increased risk of weight gain (calculated RR, 1.78 [95% CI, 1.23 to 2.57]), diarrhea (calculated RR, 7.09 [95% CI, 2.72 to 18.47]), and constipation (calculated RR, 1.67 [95% CI, 1.06 to 2.63]) compared with the trace element group. They also had a lower risk of irregular and/or colic defecation compared with the trace element group (RR, 0.33 [95% CI, 0.16 to 0.68]). The study found no difference in the risk of exanthema, although two participants in the multivitamin group and one in the trace element group withdrew from the study because of this disorder.

Figure 3

Folic Acid Supplementation and Adverse Events: Forest Plot.

Study Characteristics

One meta-analysis¹⁰¹ with a low risk of bias evaluated the association of folic acid supplementation with childhood asthma or wheezing and allergy-related outcomes. Studies were grouped by the timing of exposure: early (preconceptional, periconceptional, first trimester) versus late (second and third trimesters).^100,102,107 A second meta-analysis also examined timing of supplementation (prepregnancy, early pregnancy, other period in pregnancy).¹⁰⁹ One meta-analysis found one study reporting on dose (<200 μg/day, 200 to 499 μg/day, and >500 μg/day).¹¹⁰

Results

As noted in KQ 2, the meta-analysis that analyzed folic acid supplement use as a dichotomous variable and reported its association with asthma in childhood showed a pooled RR of 1.01 (95% CI, 0.78 to 1.30; I²=0.00; p=0.95 and 0.73, respectively) (Table 13).¹⁰¹ Two of the cohort studies included in the meta-analysis examined the association between prenatal use of a supplement containing folic acid (compared with no use) in the second or third trimester and asthma or wheezing in childhood.^100,102

Table 13

Variation in Harms of Folic Acid Supplementation by Timing.

Of the 15 associations across two studies, only one association was significantly elevated. Specifically, one study showed that maternal use of folic acid supplements in the third trimester was associated with increased risk of maternal report of wheezing at age 1 year (adjusted prevalence ratio, 1.20 [95% CI, 1.04 to 1.39]).¹⁰⁰ Regarding other outcomes, three cohort studies examined the use of supplements containing folic acid during the second or third trimester and risk of other allergy outcomes.^100,102,107 The meta-analysis reported no significant findings in 38 reported associations across these three studies.

A meta-analysis examined the incidence of asthma and wheezing by timing of supplementation (prepregnancy, early pregnancy, other period in pregnancy).¹⁰⁹ Four of five reported associations showed no statistically significant effect of folic acid supplementation on asthma or wheezing in childhood. The one statistically significant effect on wheezing in childhood was associated with exposure in early pregnancy (RR, 1.06 [95% CI, 1.02 to 1.09]).^102-104 One study separated the study population into tertiles (<0.2 mg/day, 0.2–0.499 mg/day, and >0.5 mg/day) and compared the second and third tertiles with the first for the incidence of any allergic disease, sensitization, recurrent wheezing, eczema, food reactions, immunoglobin E–mediated food allergy, and sensitization to food allergens (Table 14). In all cases, the number of events was small, ranging from 16 to 69. All results had wide CIs spanning or overlapping the line of no difference.¹¹⁰

Table 14

Variation in Harms of Folic Acid Supplementation by Dose.