Folate plays a major coenzymatic role in one-carbon metabolism and is a key participant in the biosynthesis of DNA, RNA and certain amino acids (Wagner 1995). The body's requirement for folate is thus related to the amount of cellular reproduction occurring at any particular time (Hibbard 1964). Pregnancy is associated with an enormous increase in cellular proliferation as a result of uterine enlargement, expansion of blood volume, placental development and fetal growth (Cunningham et al. 1993).

The recommended dietary allowance (RDA)⁶ for folate was reduced by approximately one half in the 10th edition from 800 and 400 µg/d (NRC 1980) to 400 and 180 µg/d (NRC1989) for pregnant and nonpregnant women, respectively. One rationale for reducing the RDA for nonpregnant women was the observation that folate intake estimated by population surveys including the Second National Health and Nutrition Examination Survey (NHANES II) (Life Sciences Research Office 1984) was ~50% lower than the RDA reported in the 9th edition, and evidence of widespread folate inadequacy was lacking (Senti and Pilch 1984). Potential limitations to this approach have been reported (Bailey 1992 and 1995) and include weaknesses in analytical methodologies used to establish food composition tables (Gregory et al. 1990) and underreporting of dietary intake. Controlled metabolic studies addressing the issue of folate requirements in nonpregnant women (O'Keefe et al. 1995, Sauberlich et al. 1987) clearly illustrated that 180 µg/d was not sufficient to maintain normal folate status and suggested that 300-400 µg/d more adequately met the definition of an RDA. The potential increase in folate requirements associated with rapid tissue growth during pregnancy is widely accepted; however, the increment above that required by nonpregnant women has not been ascertained.

The reduction in the folate RDA for pregnant women was based largely on the findings of two studies (NRC 1989). Chanarin et al. (1968b) reported that 100 µg/d synthetic folic acid, in addition to dietary folate, maintained red cell folate (RCF) concentration in pregnant women throughout gestation. Although dietary folate intake was estimated as 676 µg/d after analyzing 111 24-h duplicate meals from 16 subjects (Chanarin et al. 1968a), the 10th edition RDA committee used an estimated dietary folate intake of 190 µg/d, which was subsequently reported by Bates et al. (1982) for this population. This estimate (Bates et al. 1982) in conjunction with Chanarin's report that 100 µg/d plus diet maintained RCF was cited by the 10th RDA committee in support of 400 µg/d RDA (NRC 1989). Colman et al. (1975) found that maize fortified with 300 µg folic acid consumed with a constant diet maintained normal serum and RCF concentrations during the last 30 d of pregnancy in a nutritionally compromised, rural, African population. Interpretation of these studies is complex, and illustrates the importance of investigating folate requirements during pregnancy under controlled metabolic conditions.

McPartlin et al. (1993) estimated folate requirements in pregnant women at three different stages of gestation by quantifying the urinary excretion of the folate catabolites, p-aminobenzoylglutamate (pABG) and its acetylated derivative, acetamidobenzoylglutamate (apABG) in 24-h urine collections. This approach is based on the assumption that folate catabolites represent folate turnover and thus requirements. Pregnant women in their second trimester excreted twice as much apABG as pregnant women in their first trimester or nonpregnant controls. Folate requirement estimates were made by converting the quantities of pABG and apABG to "folate equivalents" on the basis of their molecular weights. McPartlin et al. (1993) estimated an RDA of 660 µg/d during the second trimester following adjustment for bioavailability and population variance.

The decline in folate status (both serum and red cell) during pregnancy has been well documented in both developed (Bailey et al. 1980, Ek and Magnus 1981, Lowenstein et al. 1966, Qvist et al. 1986) and underdeveloped countries (Colman et al. 1975). Although overt megaloblastic anemia is infrequent in the United States, it seems desirable to consume enough folate to maintain maternal stores to keep pace with the increased demand resulting from marked cellular proliferation of maternal, placental and fetal tissues (Herbert 1987a). Moreover, compromised folate status has been linked to poor pregnancy outcomes (Goldenberg et al. 1992, Hibbard 1964, O'Scholl et al. 1996), although not in all investigations (Blot et al. 1981, Pritchard et al. 1970).

Supplemental folic acid, in addition to dietary folate, has been shown to prevent the negative folate status associated with pregnancy. Willoughby (1967) and Willoughby and Jewel (1968) illustrated that 300-350 µg/d synthetic folic in addition to diet (low folate content) was able to maintain normal folate status in pregnant women throughout gestation, whereas supplemental amounts of either 100 or 200 µg/d were inadequate. The discrepancy between these data and the findings of Chanarin's group (1968b), in which the addition of 100 µg/d supplemental folic acid was adequate, may be explained by differences in dietary folate intake.

The role of folate in reducing the risk of neural tube defects (NTD) has been firmly established (Scott et al. 1995) and resulted in the U.S. Public Health Service recommendation that all women of childbearing age consume 400 µg/d of folate to reduce the risk of NTD (CDC 1992). Daly et al. (1995) subsequently demonstrated that RCF concentrations of 906 nmol/L (400 ng/mL) or higher were associated with a low risk of folate-responsive NTD. A study recently conducted by Brown et al. (1997) indicated that RCF concentrations 906 nmol/L (400 ng/mL) could be achieved by folate intakes of at least 450 µg/d (supplement users) or 500 µg/d (food and folic acid fortified cereals, only) provided that dietary estimates of folate consumption during the study period reflected folate consumption 2-3 mo earlier (during folate incorporation into reticulocyte).

This study was conducted to assess folate status in relation to highly controlled folate intake in pregnant adult women compared with that of nonpregnant controls. The study was designed to evaluate the adequacy of the current RDA for pregnant women (400 µg/d) (NRC 1989). Folate intakes were chosen to approximate the current and former RDA, 400 µg/d (NRC 1989) and 800 µg/d (NRC 1980), respectively. The folate status response of pregnant women was investigated throughout the second trimester of pregnancy as defined by Cunningham et al. (1993) because data suggest that this period of gestation may be associated with marked changes in folate utilization (McPartlin et al. 1993).

SUBJECTS AND METHODS

Table 1. Five-day cycle menus consumed by pregnant and nonpregnant subjects throughout 12-wk study1,2,3 [View Table]

Table 2. Nutrient composition of 5-d cycle menus1,2,3 [View Table]

Energy intake was modified to maintain weight or ensure weight gain in nonpregnant and pregnant subjects, respectively. Examples of nonnutritive food sources included margarine, candy, Jello®, Cool-Whip® and sweetened or unsweetened beverages.

Commercially available folic acid (Sigma Chemical, St. Louis, MO) was used to prepare the folate supplements. Deuterated [3,5-²H₂]folic acid was synthesized by the method of Gregory (1990), and proton nuclear magnetic resonance and HPLC were used to verify the purity and identity of these compounds before use. To prepare the supplement for consumption by subjects, the amount of deuterium labeled folic acid required to make the specific dose was weighed, dissolved in a small amount of 0.1 mol/L NaOH and brought to volume with food-grade sodium phosphate (dibasic) buffer adjusted to pH 7.0 with food-grade phosphoric acid. Unlabeled folic acid solutions in the desired quantities were similarly made. These stock solutions were divided into portions and stored at 70°C for use throughout the study. The volume of labeled and unlabeled folic acid solutions to be dispensed into each tube of apple juice was calculated after determining the folic acid concentration spectrophotometrically at 282 nm by using a molar absorptivity coefficient at pH 7.0 of 27,600 L/(mol·cm) (Blakley 1969). Labeled and unlabeled folic acid were dispensed into 50-mL conical tubes to which 45 mL apple juice was added. The tubes were capped tightly, mixed and stored at 30°C until ready for use (within 3 mo). Supplemental folic acid content was analyzed periodically by HPLC and found to be fully stable.

Complete duplicate composites of each menu were collected on three separate occasions during the study. Homogenized samples were subjected to trienzyme treatment (Martin et al. 1990) and double extraction (Gregory et al. 1990, Wilson and Horne 1984) to enhance the yield of extracted folate. Total folate content was measured microbiologically, as described above, and folate derived from enzymes was subtracted. The interassay and intraassay CV was 15 and 10%, respectively, including both the extraction process and microbiological quantification.

RESULTS

DISCUSSION

Serum folate concentration is considered a sensitive index of recent folate status (Herbert 1987b) because it is highly influenced by current dietary intake. However, under metabolic conditions, in which dietary intake is constant, SF concentration should reflect the overall folate status of the individual. The rapid decline in SF concentration in the pregnant women consuming 450 µg/d illustrates the SF response to a lower folate intake. Once acclimated to this lower intake, SF concentration was maintained within normal limits in the pregnant group consuming 450 µg/d and was equivalent to the nonpregnant controls at the same supplementation level. Hemodilution did not appear to be a factor in the initial decline in SF concentration because the 850 µg/d pregnant group did not experience any decline throughout the 12-wk period and steady-state SF concentrations were equivalent to those of the 850 µg/d controls. Hemodilution or expansion of blood volume is a known physiologic consequence of pregnancy (Cunningham et al. 1993), and a significant decline in hematocrit values from baseline was observed in the pregnant women participating in this study. It can be assumed, therefore, that the pregnant women had higher total amounts of SF at steady state than the nonpregnant controls.

Red cell folate concentration reflects liver folate concentration and is considered to be an indicator of long-term folate status (Herbert 1987b). Folate is accumulated only by developing reticulocytes (Shane 1995); because the life span of red cells is about 120 d, RCF concentration more accurately reflects folate status 2-3 mo before the time of analysis. However, because red cells are being synthesized daily over a 12-wk period, one should be able to detect some change in concentration. This is especially true during pregnancy when red cell production increases by ~33% (Blackburn and Loper 1992), resulting in greater changes in RCF concentration if inadequate amounts of folate are available at the time of incorporation. In our pregnant group consuming 450 µg/d, RCF was maintained throughout the study period and was equivalent to the 450 µg/d control group at steady state. Perhaps more importantly, normal RCF concentrations were also maintained throughout the 3-mo follow-up period in the subsample of subjects who returned for subsequent blood draws. In addition, 450 µg/d (food folate + supplemental folic acid) was sufficient to maintain mean RCF concentrations above 906 nmol/L (400 ng/mL) throughout the study and follow-up period. thus supporting the findings of Brown et al. (1997). Red cell folate concentrations were also maintained above 906 nmol/L in both the pregnant and nonpregnant women consuming 850 µg/d throughout the study and follow-up period.

Although urinary excretion of folate is not considered to be a reliable index of folate status (Sauberlich et al. 1987), one might expect to observe differences between pregnant and nonpregnant women and between supplementation groups. These differences may reflect metabolic differences, thereby enhancing the information gained from blood indices. Pregnant women have been found to excrete significantly more folate than nonpregnant women, which is hypothesized by some investigators to contribute to the increase in requirements during pregnancy (Fleming 1972, Landon and Hytten 1971). Pregnant and nonpregnant women consuming 450 µg/d were excreting similar amounts of 5-methyl-THF (~11.3 nmol/d) at steady state and no detectable folic acid. Pregnant and nonpregnant women consuming 850 µg/d were also excreting similar amounts of 5-methyl-THF (~180 compared with 136 nmol/d, respectively) at steady state, ~15-fold higher than the 450 µg/d groups. Only the 850 µg/d nonpregnant group excreted folic acid (19% of total urinary folate). Saleh et al. (1980) reported a lower folic acid to 5-methyl-THF ratio in patients with malignant disease (state of increased cellular proliferation) compared with controls. They suggested that malignant disease increased the demand for folate and led to more rapid metabolism of folic acid to the reduced folate pool as indicated by an increase in 5-methyl-THF excretion relative to folic acid. Their explanation may apply to our finding that only nonpregnant women in the 850 µg/d group excreted unmetabolized folic acid because pregnancy also represents a period of increased cellular proliferation. The higher urinary excretion of 5-methyl-THF (metabolized form) in the 850 µg/d groups may reflect the saturable process of folate reabsorption from glomerular filtrate by proximal tubular cells (Williams and Huang 1982). Overall, these data on urinary folate excretion indicate that pregnant women are not excreting more folate than nonpregnant women and support the blood data in which no differences between pregnant and nonpregnant women within the same supplementation group were detected. Homocysteine, a functional index of folate status, was also quantified but will be reported separately.

From these data and within the constraints of our study, we conclude that 450 µg/d derived from both food (120 µg) and synthetic folic acid (330 µg) is adequate to maintain normal folate status in well-nourished pregnant women during their second trimester. However, there remains the question of translating this finding into dietary equivalents because the overall goal was to provide metabolic data to assist in the interpretation of the current RDA. Recent studies (Cuskelly et al. 1996, Pfeiffer et al. 1997) indicate that folic acid in fortified food is highly available unlike naturally occurring food folate. Therefore one may hypothesize that synthetic folic acid consumed with meals is more available than endogenous food folate (~50%) (Sauberlich et al. 1987) but less available than the essentially complete absorption of synthetic folic acid (~100%) consumed under fasting conditions (Gregory 1995). A reasonable, although somewhat conservative approach is to assume that synthetic folic acid consumed with meals is ~75% available. Based on the above assumptions, our subjects in the 450 µg/d group consumed ~307 µg/d available folate which translates into 615 µg/d dietary equivalents.

Our dietary equivalent estimate resembles the estimations of Chanarin et al. (1968a and 1968b) and McPartlin et al. (1993). It also supports the recommendation of Willoughby et al. (1968) regarding the administration of 300-350 µg/d synthetic folic acid throughout gestation, assuming an average folate intake of 150 µg/d (low dietary folate). It is questionable whether the majority of pregnant women can consume ~600 µg/d of folate from diet alone (Brown et al. 1997, Huber et al. 1988, LSRO 1980), although folic acid enrichment of cereal-grain foods (140 µg/100 g product) effective January 1,1998 (FDA 1996) is estimated to increase average daily consumption by 80-100 µg (Brown et al. 1997, FDA 1996). Our findings suggest that prenatal vitamins containing more than 500 µg/d are not necessary to maintain adequate folate status in well-nourished pregnant populations and support the findings of Lowenstein et al. (1966), which suggested that provision of 500 µg/d synthetic folic acid, in addition to low dietary folate intake, may be above the minimal daily requirement because higher serum folates were observed in pregnant women compared with nonpregnant controls (24.9 vs. 15.9 nmol/L, respectively). Because the effects of oversupplementation with folate on the developing fetus are unknown (Scott et al. 1991), high supplemental doses (>1000 µg/d) should be avoided under normal circumstances. The first and third trimesters or postpartum periods were not investigated under controlled conditions in the current study; therefore, future areas of research should encompass these time frames under controlled conditions as well as the response of previously unsupplemented pregnant women to defined folate intakes.

ACKNOWLEDGMENTS