Is the Jury Still Out on Folic Acid and Congenital Anomalies?

In this issue of Epidemiology , Turner et al argue that another randomized clinical trial of folic acid alone may be necessary to determine whether this nutrient specifically protects against neural tube defects (NTDs). 1 Their argument is based on two premises: that the epidemiologic evidence (including that from a recent large-scale public health campaign in China) is weak and does not allow for the separation of a folic acid effect from that of other vitamins, and that the principal randomized trial supporting a protective effect of folic acid alone was subject to a striking type of non-compliance bias that threatens the validity of the results. The Strength of the Evidence for an Effect of Folic Acid Alone During the 1980s and 1990s, there were several case-control studies of multivitamin use and NTDs. 2–7 Most found a strong protective effect of taking multivitamins containing folic acid during the periconceptional period. One large prospective cohort study demonstrated a 64% reduced risk of a NTD (prevalence ratio = 0.36; 95% CI = 0.15–0.83) in the offspring of women who took a multivitamin during the first 6 weeks of pregnancy compared with those who took no multivitamins. 8 More importantly, the offspring of those who took a multivitamin containing folic acid had a 73% reduced risk of a NTD while the offspring of those who took a multivitamin that did not contain folic acid had only a 7% reduced risk of a NTD. Another epidemiologic study demonstrated a rise in the prevalence of NTDs following the reduction in intake of folic acid-containing foods associated with a natural disaster. 9 The evidence from the recent non-randomized intervention study in China adds important weight to the evidence from the above studies. 10 Beginning in 1993, all women in one northern province and two southern provinces in China were pregnant or preparing for marriage who were invited to purchase supplements containing 400 μg of folic acid and to take those supplements every day through the first trimester of pregnancy. Women who did not agree to take the supplements or who were already pregnant at the time of registration served as controls. In the northern region, with a high baseline rate of NTDs, there was a 79% reduction in NTD risk associated with folic acid use during the periconceptional period. In the south, with much lower baseline rates of NTDs, there was a 41% reduction in risk. Turner et al suggest that the study is weakened by the lack of randomization and the requirement for the women to purchase supplements, since those who purchased supplements may have differed systematically with respect to unmeasured risk factors from women who did not. First, a randomized placebo-controlled trial during the mid-1990s given the existing evidence at the time would have been unethical. Furthermore, the requirement to purchase the pills must be considered in its cultural context before determining whether such a requirement would enhance or reduce bias. The authors report that they stratified by age, number of previous pregnancies, education, and occupation and that this analysis did not alter the study’s results. While we cannot completely rule out other unknown factors biasing the results, it is difficult to imagine factors of sufficient magnitude in both regions to eliminate the observed protective effect of folic acid. Another fact that strengthens the China study is the change in the population rates of NTD during the intervention period. The overall baseline rate of NTD occurrence (during the 9 months before the onset of supplementation) in the northern region was 5.5/1,000 pregnancies of at least 20 weeks gestation. This rate declined to 3.3/1,000 pregnancies during the study period, making it highly unlikely that systematic population differences alone (between the supplemented and non-supplemented groups of women) could explain the NTD risk reduction. Similarly, the overall baseline NTD rate of 1.0/1,000 pregnancies in the south declined to 0.8/1,000 during the folic acid intervention. In fact, the folic acid supplementation program essentially wiped out the geographic difference in NTD prevalence among women who were compliant with supplementation (prevalence of NTDs among women who were more than 80% compliant with folic acid supplementation was 0.7/1,000 pregnancies in the north vs 0.6/1,000 pregnancies in the south). Thus, this non-randomized intervention provides strong additional evidence for a protective effect of folic acid alone. There have been several randomized clinical trials demonstrating a protective effect of folic acid-containing multivitamins on NTD risk. 11–15 As Turner et al point out, the supplementation trial carried out by the Medical Research Council (MRC) 16 perhaps weighed most heavily in the decision of the United States Food and Drug Administration (FDA) to embark on a folic acid food fortification program. 17 The MRC trial was carried out among women with a previous pregnancy affected by a NTD and used a factorial design to assess the effects of folic acid alone (4 mg/day), multivitamins alone (without folic acid), and folic acid plus multivitamins compared with a control group taking a supplement with iron and calcium alone. Compliance was measured by ascertainment of serum folate concentrations at the visit immediately before the women became pregnant (median time to 1st day of last menstrual period was 7 days). As Turner et al note, the 90th percentile of serum folic acid in this study is very high (an average of 194 ng/ml in the two groups exposed to folic acid at 4 mg/day), suggesting that some women may have been doubling up on missed doses before the blood test. Folate status in the body is measured as either serum folate or red blood cell folate, although serum folate has not been as closely correlated with NTD risk. 18 Serum folate reflects very recent folic acid consumption and levels fall quickly after a reduction in intake. Red cell folate, however, reflects longer-term intake since folate remains in the red blood cell for its entire life span of 120 days. Turner et al suggest that doubling up on supplements in all arms of the trial may have exposed the offspring of women in both multivitamin exposure groups to dangerously high levels of other vitamins, particularly vitamin A, thereby increasing the risk of congenital anomalies in those groups. In addition, if double dosing did occur, the controls may have been exposed to excessively high doses of supplemental iron and calcium, both of which are known to inhibit absorption of dietary zinc, thus putting the offspring of women in the control group at increased risk of a NTD due to zinc deficiency. It does appear that some women in the MRC study may have doubled up on missed doses before their blood tests. It is not at all clear, however, that this action invalidates the primary results of the study for several reasons. First, there is no evidence that such double dosing occurred during the critical period for closure of the neural tube since these blood tests were generally taken before pregnancy. Secondly, there is no real evidence from human studies that the teratogenic effect of vitamin A extends to NTDs; in the large prospective study cited by Turner et al, there was no such adverse effect shown for NTDs. 19 (Nevertheless, I agree all women of childbearing age should be cautioned against doubling up on multivitamins to prevent malformations associated with exposure to teratogenic doses of some vitamins.) Finally, the rate of NTDs in the control group in the MRC study is within the range of what would have been expected (for women with a previous history of having a baby with a NTD). Thus, it seems unlikely that the observed protective effect of folic acid is simply a consequence of a spuriously high prevalence of NTDs among the controls due to zinc deficiency. Overall, the available evidence for a protective effect of folic acid is strong and consistent. Turner et al have not convincingly demonstrated that the methodologic issues cited may have invalidated the earlier published results. Despite their recommendation that a randomized controlled trial of folic acid alone may be necessary, the authors acknowledge that it would be difficult to convince clinicians to enroll patients in such a placebo-controlled trial. More importantly, such a trial would be clearly unethical given the existing evidence. The authors also consider the possibility of a randomized trial of folic acid alone vs multivitamins containing folic acid, but they conclude that such a trial would be impractical for reasons of power and size. Does Current Fortification with Folic Acid Go Far Enough? In 1996, the U.S. FDA decided to add 140 μg of folic acid to every 100 g of grain beginning in January of 1998, with the goal of adding 100 μg/day of supplemental folate to the diets of all Americans. This dose was thought to have the potential to reduce the prevalence of NTDs by about 20% while not increasing the risk, particularly among the elderly, of undetected consequences of pernicious anemia. Some have argued that the current fortification amount is too conservative and should be increased to 400 μg/day. 20,21 Oakley has been an outspoken advocate for increasing the folic acid fortification levels in food. 22 He argues that the beneficial effects of increased fortification far outweigh the risks associated with the masking of pernicious anemia and he cites data from the Framingham Study in which 16 of 694 subjects were found to have both anemia and low plasma vitamin B-12 levels. Of those, only one was projected to consume a potentially harmful dose above 1.0 mg of total folate per day. The Framingham investigators, however, conclude that vitamin B-12 deficiency and perhaps its associated neurologic or neuropsychologic symptoms may be more widespread than previously thought. 23 Fortification at the present level may reduce plasma homocysteine levels and, hence, heart disease risk in a large number of elderly. Despite these benefits, Tucker et al conclude the much more data are needed to understand the clinical course of various types of vitamin B-12 deficiency and to measure the potential effects of high folate intake on the course of this condition. In our own data from a prospective cohort of more than 23,000 pregnant women, 3.0% of women consumed at least 1.0 mg/day of supplemental folic acid in the month before pregnancy, while 7.1% consumed at least 1.2 mg/day of total folate. These levels were present before the current food fortification program was in place so the proportion of individuals with high total folate will be considerably greater now. The goal of the U.S. Department of Health and Human Services (Healthy People 2010) was to increase the median red blood cell folate level in women aged 15–44 years in the United States to 220 μg/L. 24 In October of 2000, the National Center for Health Statistics (NCHS) released its first report of the changes in folate status for women of childbearing age associated with the current level of food fortification. 25 NCHS compared data on red blood cell folate levels from the Third National Health and Nutrition Examination Survey conducted between 1988 and 1994 (NHANES III) with that from NHANES 1999. The median red blood cell folate concentration increased from 160 μg/L during the pre-fortification period to 293 μg/L during the 1999 survey, an 80% increase in red blood cell folate. Red cell folate status at the 10th percentile increased from 92 μg/L to 174 μg/L and at the 90th percentile from 296 μg/L to 474 μg/L. The recent results from NCHS are encouraging. Based on these data, about 75% of women of childbearing age have already reached the 2,010 folate goals. Fortification at higher levels would raise red blood cell folate levels even more and perhaps reduce the risk of NTDs even further. 26 Nevertheless, the jury is still out on a number of important questions that must be addressed before embarking on a course of higher fortification. What Questions Remain? The question of the minimum effective dose required for the prevention of first occurrence and recurrence of NTDs (and other defects) is unknown. 27 Clinical and epidemiologic data demonstrating that higher doses of supplemental folic acid provide additional NTD protection are sparse. While most scientists acknowledge the need for such data, the general consensus seems to be that ethical and feasibility issues would preclude the studies. I disagree. Certainly, placebo-controlled trials would be unethical. If the relation between folic acid dose and NTD risk is truly linear, however, then carefully designed studies of the full range of the dose-response curve are needed. The current fortification program provides the perfect opportunity for epidemiologic studies of this question, including studies designed to assess the interaction of supplemental folic acid at different doses with plasma and dietary indicators of vitamin B-12 levels. Studies of the possible adverse effects of high-dose folic acid are also needed. In addition to the unresolved issues associated with anemia, it is not known whether high intakes of folic acid could have a teratogenic effect on the fetus or lead to other adverse fetal or maternal outcomes. The possibility of harmful fetal effects from excessive folate has been addressed in a small number of animal studies. Middaugh et al found similar teratogenic effects of folic acid deficiency and excess in pregnant mice. 28 Other animal studies suggest that folic acid and other nutrients may act synergistically in the production of malformations. 29,30 Hook and Czeizel have also suggested that the apparent protective effect of folic acid in pregnant women may be a consequence of folate-induced spontaneous abortion of affected fetuses. 31 Another possible concern stems from findings that high folic acid intake may stimulate seizure activity. 32 For those with autoimmune-related disorders, there may be greater permeability of the blood-brain barrier, providing more opportunity for high-dose folic acid to stimulate seizure activity. 33 At this time, there are insufficient data to determine whether any of the above adverse outcomes are of concern for pregnant women or others. While the detection of rare outcomes (particularly when associated with imprecisely measured exposures) is difficult, such rare outcomes will effect large numbers of individuals when millions are exposed. Conclusions The evidence that folic acid protects against the development of NTDs is strong. Future studies designed to answer whether such an effect exists would be unethical to conduct anywhere in the world. The minimum effective dose of folic acid and the potential harm associated with high doses in the population as a whole, however, are largely unknown. Thus, pushing the upper limit of food fortification beyond its current level cannot be supported at this time. Epidemiologic studies of the many important questions associated with the dose, timing, and interaction of folic acid with other nutrients are needed.