点击显示 收起
Division of Birth Defects and Developmental Disabilities
National Center on Birth Defects and Developmental Disabilities
Centers for Disease Control and Prevention
1600 Clifton Road
Mail-Stop E-86
Atlanta, GA 30333
E-mail: rjberry{at}cdc.gov
Dear Sir:
The recent article by Morris et al (1) reports perplexing findings about the association of folic acid, vitamin B-12, and cognitive functioning among healthy participants aged 60 y in the United States from the National Health and Nutrition Examination Survey (NHANES) 1999–2002. They found that participants with a combination of high serum folate concentrations (>59 nmol/L) and normal vitamin B-12 status were less likely to have cognitive impairment, whereas those with a combination of high serum folate and low vitamin B-12 status were more likely to have cognitive impairment. The article by Morris et al and the accompanying editorial by Smith (2) suggest that these findings have important implications for current folic acid fortification practices in the United States and for other countries considering folic acid fortification. Although the analysis is well done, the lack of sufficient information about the small subgroup of interest (n = 42) with high serum folate and low vitamin B-12 status, the cross-sectional nature of the data, and the lack of consideration of the relative contribution of different sources of folic acid, which include supplements and all fortified food products, including ready-to-eat (RTE) breakfast cereals, raise important questions that should be considered when drawing inferences from this study. In countries in both the developed and developing world, it is likely that the only significant source of folic acid is fortified flour. The findings of Morris et al might only be applicable to seniors in the United States whose use of supplements and consumption of folic acid–fortified RTE breakfast cereals is high.
It would have been useful to see a comparison between participants with high serum folate and normal serum folate. It would have been more useful to see a comparison between participants in the highest folate quintile with low and normal vitamin B-12 status. The subgroup with high serum folate and low vitamin B-12 status was a highly restricted population: 1.1% (42/3706) of the senior survey participants, 2.5% (42/1684) of eligible seniors, and only 3.2% (42/1302) of those seniors who could have been evaluated for cognitive function. The analytic and reporting guidelines of NHANES state that, "The more one deviates from the original analytic categories ... the more important it is to evaluate the results carefully and to interpret the findings cautiously" (3). Therefore, drawing appropriate inferences from this small subgroup might be problematic.
Another potential limitation of the study by Morris et al is inherent in any study in which the use of folic acid is not randomized. When blood folate concentrations and/or intakes of folic acid–fortified food products are used to create folate and folic acid exposure strata, most subjects in the highest stratum will be supplement users and most subjects in the lowest stratum will not be supplement users. To illustrate this stratification among seniors in the United States, we divided non-Hispanic white participants aged 60 y from NHANES (2001–2002) into serum folate quintiles and analyzed their use of supplements and their dietary intake data (4). We used SUDAAN (version 8.1; Research Triangle Institute, Cary, NC) to take into account the complex survey design and calculated for each quintile of serum folate the proportion of non-Hispanic white seniors who used a supplement containing folic acid, the geometric mean of serum folate, and the total daily intake of folic acid from supplements and food. We also calculated the total daily intake of synthetic folic acid from all fortified foods and estimated the total daily intake of enriched cereal-grain products by removing folic acid found in RTE breakfast cereals from all fortified foods. Finally, we calculated the proportion of supplements containing folic acid that also contained vitamin B-12 and the median content of vitamin B-12 in those supplements.
It is shown in Table 1 that 92% and 12% of non-Hispanic white seniors in the highest and lowest folate quintiles, respectively, used supplements containing folic acid. As documented in the recent National Institutes of Health State-of-the Science conference on multivitamin-mineral supplement use and chronic disease prevention, there appear to be important behavioral and health-related differences between individuals who report using a supplement (eg, elderly persons, those with higher education, those with higher income, those with healthier lifestyles and diets, and those who are seeking to prevent a serious disease) and those who do not report using a supplement (5). Such differences increase the possibility that unmeasured confounding related to unique differences between subjects who take supplements and those who do not could account, in part, for the findings reported by Morris et al.
View this table:
TABLE 1. Proportion of participants who used supplements containing folic acid, serum folate concentrations, total daily intake of folic acid, and total daily intake of folic acid from all folic acid–fortified food products and enriched cereal-grain products by serum folate quintiles in non-Hispanic white seniors aged 60 y from the National Health and Nutrition Examination Survey (NHANES), 2001–20021
Of all the non-Hispanic white seniors, the 80th percentile of serum folate, which defined our highest quintile, was 58.2 nmol/L, which was similar to the value (59 nmol/L) Morris et al used to define their highest quintile. It is also shown in Table 1 that, for non-Hispanic white seniors in the highest quintile, the mean serum folate concentration was 77.9 nmol/L, and the geometric mean daily total intake of folic acid from fortified food products plus supplements was 555 µg. Mean daily intakes of folic acid from all fortified food products and enriched cereal-grain products were 144 and 86 µg, respectively. In the highest folate quintile, daily intakes of folic acid from all fortified food products constituted 26% (144/555), and enriched cereal-grain products constituted only 15% (86/555) of the total daily intake of folic acid. It is unclear why both the article by Morris et al and the accompanying editorial focus most of their discussion on folic acid fortification. Only folic acid from enriched cereal-grain products should be considered when making inferences that affect other countries where folic acid fortification of flour is likely the only source of synthetic folic acid in the diet.
In addition, we used SAS (version 9.1; SAS Institute Inc, Cary, NC) to evaluate correlations with serum folate concentrations in our table. Across the entire serum folate distribution, we found a strong correlation between the serum folate concentration and both the total daily intake of folic acid (r = 0.293, P < 0.0001) and the daily use of supplements containing folic acid (r = 0.289, P < 0.0001). We found a very weak, but statistically significant, correlation between the serum folate concentration and the total daily intake of folic acid from all fortified food products (r = 0.080, P = 0.016), but found no correlation with enriched cereal-grain products (r = –0.026, P = 0.444). These data clearly show that high concentrations of serum folate in older adults are not due to the consumption of enriched cereal-grain products and suggest that the findings of Morris et al should not be considered as relevant outside of the United States.
Another intriguing question about the Morris et al study concerns the subgroup of seniors with high serum folate and low vitamin B-12 status. In NHANES (2001–2002), we found that 92% of non-Hispanic white seniors in the highest serum folate quintile took supplements containing folic acid and that 97% (95% CI: 90%, 99%) of these supplements also contained vitamin B-12; the median daily vitamin B-12 content was 16.6 µg (95% CI: 6.8, 23.1µg). Therefore, it is unclear why these 42 participants continued to have low vitamin B-12 status when almost all of them would have been consuming oral vitamin B-12. Could they have been unable to absorb vitamin B-12? Could they have been developing pernicious anemia? Could their cognitive impairment have been due to their low vitamin B-12 status? These are some of the questions that need to be addressed before the finding of an association between high serum folate and cognitive impairment can be generalized to the entire population of US seniors with low vitamin B-12 status. We agree with Morris et al that further study is necessary to clarify their findings.
Morris et al also attribute the presence today of unmetabolized (free) folic acid in blood to folic acid fortification. Kelly et al (6) showed that consumption of a supplement containing 200 µg folic acid would produce free folic acid. If free folic acid in the blood were responsible for any adverse effects, such effects, if detectable, would have been present before folic acid fortification began in the United States. Any person using a supplement containing folic acid should have measurable blood concentrations of free folic acid. The overall prevalence of dietary supplement use in the United States has increased from 23% in NHANES I to 40% in NHANES III (7). For almost half a century in the United States, women have been prescribed prenatal vitamins containing folic acid during pregnancy. Therefore, free folic acid is unlikely to be a new phenomenon. We disagree that the findings of Morris et al support the theory that folic acid directly exacerbates the neurologic and neuropsychiatric effects of low vitamin B-12 status.
Even though 20% of the seniors in the study by Morris et al had serum folate concentrations >59 nmol/L, we think it is unlikely that such high serum folate concentrations are easily achievable from consuming folic acid–fortified food products anywhere consumption of RTE breakfast cereals is low. We found that the mean folic acid intake from folic acid fortification was 125 µg/d for all non-Hispanic white seniors; this value is similar to earlier estimates of the daily intake of folic acid from all fortified food products in the United States, which ranged from 128 µg/d for women of childbearing age (8) to 200 µg/d for the entire population (9). Data from Chile's fortification program, which has the world's highest recommended level of folic acid used to fortify flour, shows that the mean daily intake of folic acid from fortified bread was 427 µg, which among women of reproductive age, produced a mean serum folate concentration of 37.2 nmol/L (95th and 99th percentiles of 56.7 and 61.7 nmol/L, respectively) (10). Even though Chilean women consumed >3 times our estimate of the daily intake of folic acid (125 µg) from all fortified food products, <5% of Chilean women had serum folate concentrations >59 nmol/L. These data reinforce the suggestion that the high serum folate concentrations found in 20% of seniors in the Morris et al study were likely achieved primarily by the consumption of supplements containing folic acid.
In summary, any observational study, such as that by Morris et al, which associates folic acid with any health outcome, has inherent methodologic issues from unmeasured confounding associated with stratification of supplement users. Such studies should recognize that high concentrations of serum folate are likely due to the use of supplements that contain folic acid or RTE breakfast cereals and should consider whether some underlying disease, such as undiagnosed pernicious anemia, might account for their findings. Any such study should be carefully evaluated so that appropriate inferences are made, especially when the findings are generalized to populations in whom the use of supplements and RTE breakfast cereals are low. Solid evidence from randomized trials or cohort studies would be useful to better understand the relation between folic acid, vitamin B-12, and cognitive impairment. The apparently paradoxical findings of Morris et al warrant continued investigation.
We must continue to be vigilant in identifying potential adverse effects of folic acid fortification, but this effort must be founded on appropriate inferences. We must continue to examine the potential positive and negative effects of public health intervention programs and make decisions that will be safe for the entire population. Nevertheless, these decisions should not compromise the effectiveness of proven interventions, such as folic acid fortification, to prevent neural tube birth defects. Even if the findings of Morris et al are confirmed, we believe that it would be inappropriate to generalize their findings to countries in which the use of supplements and consumption of folic acid–fortified RTE breakfast cereals is low. For now, wheat flour for bread, which is fortified with folic acid, is likely the only potential source of synthetic folic acid elsewhere in the world. Based on Morris et al's findings, the editorial by Smith cautions countries contemplating folic acid–fortification programs intended to prevent neural tube defects to delay such a decision. We believe that a delay in the implementation of folic acid–fortification programs elsewhere in the world on the basis of an inappropriate interpretation of Morris et al's findings is not warranted and would be detrimental to public health.
ACKNOWLEDGMENTS
No conflicts of interest were reported. The findings and conclusions in this letter are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
REFERENCES