点击显示 收起
1 Cardiovascular Sciences Research Group, Wales Heart Research Institute (PALA-W, CHP, JMW, ZEC, MJL, and IFWM), and the Departments of Epidemiology and Public Health Medicine (MLB) and Medical Computing and Statistics (RGN), University of Wales College of Medicine, Cardiff, Wales, United Kingdom, and the University of Sheffield, Centre for Human Nutrition, The Northern General Hospital, Sheffield, United Kingdom (SJM and HJP).
2 Supported by the UK Ministry of Agriculture, Fisheries and Foods (UK, Food Standards Agency, project reference N05002). The Kellogg Company of Great Britain provided funding for reimbursement of subjects for the cost of fortified foods. Folic acid supplements and matched placebos were provided by Peter Black Healthcare Ltd, United Kingdom. 3 Address reprint requests to IFW McDowell, Department of Medical Biochemistry, c/o Wales Heart Research Institute, University of Wales College of Medicine, Heath Park, Cardiff, CF14 4XN, United Kingdom. E-mail: ian.mcdowell{at}uhw-tr.wales.nhs.uk.
ABSTRACT
Background: Low folate status and elevated plasma homocysteine are associated with increased risk of neural tube defects and cardiovascular disease. Homocysteine responses to folate may be influenced by genetic variants in folate metabolism.
Objective: We determined the effect of folate-enhancing dietary interventions on plasma folate and plasma total homocysteine (tHcy) with respect to the methylenetetrahydrofolate reductase 677CT genotype.
Design: A total of 126 healthy subjects (42 TT, 42 CT, and 42 CC genotypes) completed 3 dietary interventions (4 mo each) in random order: 1) exclusion diet (avoidance of folic acid–fortified foods and ingestion of a placebo daily), 2) folate-rich diet (increased intake of fortified and naturally folate-rich foods to achieve 400 µg folate/d), and 3) supplement (exclusion diet plus a folate supplement of 400 µg/d).
Results: Plasma folate was higher (P 0.001) and plasma tHcy lower (P 0.001) after the folate-rich and supplement interventions than after the exclusion diet. Plasma folate was significantly greater after supplementation than after the folate-rich diet, but there was no significant difference in tHcy concentration (P = 0.72). TT homozygotes had higher plasma tHcy (14.5 compared with 8.9 µmol/L, P 0.001) and lower plasma folate (14.8 compared with 19.0 nmol/L, P 0.01) than did subjects with the CC genotype after the exclusion diet. CT heterozygotes had intermediate concentrations. The trend toward higher tHcy in TT homozygotes persisted throughout the study but was less marked with increasing folate intake (TT compared with CC after supplementation, P = 0.097).
Conclusions: A folate-rich diet including folic acid–fortified foods or low-dose supplements effectively increases folate status. TT homozygotes require higher folate intakes than do individuals with the CT or CC genotype to achieve similar tHcy concentrations but are responsive to folate intervention.
Key Words: Folic acid • folate • homocysteine • MTHFR genotype • methylenetetrahydrofolate reductase • fortification • healthy population • dietary intake • neural tube defects
INTRODUCTION
Suboptimal folate status is a risk factor for neural tube defects and cardiovascular disease. The benefits of folate in preventing neural tube defects have been established in intervention trials of folic acid supplements (1,2), but the evidence for cardiovascular disease is equivocal (3,4). Hyperhomocysteinemia is associated with cardiovascular disease, but a causal role remains unproven. Plasma homocysteine is inversely related to blood folate concentrations and is a responsive marker of folate status. Folic acid supplements increase blood folate concentrations and reduce homocysteine concentrations (5).
How the general population can safely achieve an optimal folate intake remains a major public health issue. Natural food folates have poor bioavailability and dietary interventions to increase natural folate intake have been largely ineffective (6). Synthetic folic acid, available in supplement form, has greater bioavailability and stability. Supplements are effective only in planned pregnancy, however, and are often ignored by those most at risk. Therefore, fortification of dietary staples has been the method of choice for folic acid delivery in the United States (7) and Canada and has been recommended by the Committee on Medical Aspects of Food (COMA) in the United Kingdom (8), where voluntary fortification of cereal products is widespread. Mandatory food fortification would increase folate intake at the population level and has proved effective in terms of biochemical response in the United States (9,10). A reduction in neural tube defects was reported in Australia after the introduction of fortified foods (11). If adopted in the United Kingdom at the level recommended by COMA (240 µg/100 g cereal product), food fortification would result in most of the population attaining intakes of 400 µg/d. Epidemiologic studies suggest that this would result in maximum homocysteine lowering for the general population (12,13).
The strong inverse relation observed between plasma folate and homocysteine is a consequence of the role of folate in single-carbon metabolism. Homocysteine, which is produced during methionine metabolism, is remethylated to methionine by methionine synthase (EC 2.1.1.13) and its cosubstrate 5-methyltetrahydrofolate, the most abundant form of folate in the plasma.
Dietary folate intake is a key factor in determining folate status because folates are not synthesized de novo in humans. Folate concentrations depend on gene-nutrient interactions. The C-to-T substitution at nucleotide 677 (677CT mutation) in the gene encoding 5,10-methylenetetrahydrofolate reductase (MTHFR) produces a partially defective enzyme (14) that is present in the TT form in 12% of whites (15). MTHFR activity regulates 5-methyltetrahydrofolate, which is required for the remethylation of homocysteine. TT homozygotes have elevated homocysteine concentrations at low folate intakes and may have a greater requirement for folate than do persons with other genotypes. It has been suggested that homozygosity for the MTHFR mutation is a risk factor for both neural tube defects (16) and, more controversially, cardiovascular disease (17,18).
No information is available from randomized controlled trials concerning the influence of MTHFR genotype on the adequacy of the fortification or supplementation levels recommended. Thus, the present study compared interventions with fortified foods or low-dose (400 µg) folic acid supplements in healthy individuals with different MTHFR genotypes to assess the influence of genotype on folate requirements and responsiveness.
SUBJECTS AND METHODS
Subjects
Healthy volunteers identified from the local work force and through blood donation sessions were invited to participate in the study. Subjects were eligible for inclusion if they satisfied the following criteria: 1) were nonsmoking; 2) had no history of cardiovascular disease or epilepsy; 3) were aged between 18 and 65 y; 4) were not taking supplements containing vitamins B-6, B-12, or folic acid; 5) were not taking any drugs known to interfere with folate metabolism (eg, methotrexate or bile acid sequestrants); and 6) if female, were not pregnant or planning pregnancy during the time frame of the study. All subjects gave their written, informed consent and underwent MTHFR genotyping. Women additionally consented to undergo a pregnancy test at their first study visit if selected to take part on the basis of their MTHFR genotype. Ethical approval was obtained from Bro Taf Health Authority.
In total, 634 subjects were genotyped for MTHFR. Of these, 126 subjects were selected (42 TT, 42 CT, and 42 CC) and entered into a crossover study comprising three 4-mo dietary interventions. The study profile is illustrated in Figure 1. Within each genotype group, subjects were randomly allocated to receive the 3 interventions in 1 of the 6 possible orders to minimize the effect of any carryover (ie, 1, 2, 3; 1, 3, 2; 2, 1, 3; 2, 3, 1; 3, 1, 2; and 3, 2, 1). The randomization was therefore balanced in blocks of 18 (ie, 6 possible order combinations x 3 genotypes). Systematic bias was avoided by entering subjects of each genotype into the study at the same rate. The least common genotype (TT) was therefore rate limiting. In practice, subjects were entered into the study in batches of 3 (1 TT, 1 CT, and 1 CC), ie, when a subject with TT genotype was identified, the succeeding CT and CC subjects on the screening register were entered into the study. The clinical investigators and the subjects were blinded to the subjects' genotype throughout. To reduce the possibility of a seasonal effect, most subjects were entered into the study over a 12-mo period.
FIGURE 1. . Recruitment of subjects by methylenetetrahydrofolate reductase (MTHFR) genotype and the number of subjects returning for follow-up after each dietary intervention.
Subjects attended the Research Institute for assessment on 4 occasions. Baseline measurements were made at visit 1 and postintervention measurements at visits 2, 3, and 4. Each visit consisted of venipuncture, a dietary interview, and assessment of vascular endothelial function. Vascular measurements will be reported elsewhere.
Nutritional assessment
Baseline and follow-up dietary assessments were by semiquantitative food-frequency questionnaire. Nutritional analysis was based on McCance and Widdowson's The Composition of Foods (19) and UK food portion sizes (20). Subjects were also instructed on how to complete a chart for monitoring folate intake. This involved a folate scoring system that gave points according to the folate content of foods. Subjects were asked to complete a 2-wk chart at the beginning of each 4-mo phase and again halfway through the phase and to return these in prepaid envelopes. A target number of points per day was set depending on the dietary intervention required.
The concept of dietary folate equivalents (DFEs) (21) has been introduced to facilitate comparison of dietary folate intakes. Use of DFEs simplifies dietary folate advice by introducing a measurement of folate intake that accounts for the varying bioavailability of folates from different sources. The DFE value of natural source folates is that figure quoted in food-composition databases. For synthetic folic acid, the amount in micrograms per portion must by multiplied by 1.7 to adjust for the greater bioavailability of folic acid. The current study began before the advent of DFEs and therefore target intakes and analyses are presented in the original units (µg/d). Estimated DFEs were calculated and are presented in parentheses in the text.
Interventions
The subjects undertook 3 dietary interventions (each lasting 4 mo) in random order: exclusion diet, folate-rich diet, and supplement. During the exclusion diet and the supplement phase, the intervention was carried out double blind such that neither the nutritionist nor the subject knew which treatment (placebo or supplement) the subject was receiving. It was not possible to blind either party regarding folate intake from dietary sources.
Exclusion diet
Subjects were asked to consume their usual diet but to substitute foods fortified with folic acid with unfortified ones and to take a placebo tablet daily.
Folate-rich diet
Subjects were advised to consume their usual diet plus additional folic acid–fortified foods and naturally folate-rich foods to achieve a total folate intake of 400 µg/d (575 DFE). To achieve this, subjects were advised to eat one bowl of Kellogg's Cornflakes (Kellogg Company of Great Britain, Manchester, United Kingdom) or other cereal fortified at the same level (333 µg/100 g, or 170 DFE per bowl) plus 3 slices of fortified bread [Champion bread (Holgran Ltd, Burton-on-Trent, Staffordshire, United Kingdom), Mighty White bread (Allied Bakeries, Staines, Surrey, United Kingdom), or similar bread containing 118 µg/100 g, or 255 DFE] and to increase their intake of folate-rich foods such as citrus fruit, green leafy vegetables, and liver. Fortified foods were purchased by subjects from their usual retail outlets and costs were reimbursed.
Supplement
Subjects were asked to consume the exclusion diet but to take a 400-µg folic acid supplement daily (680 DFE).
Genotyping
MTHFR genotyping was performed by using heteroduplex analysis involving extraction of DNA from buccal cells, polymerase chain reaction, and electrophoresis (15). The buccal cells were obtained from the inner lining of the cheek by using a "cyto-brush."
Blood analysis
Venous blood was collected into evacuated tubes containing EDTA for the determination of plasma folate and plasma total homocysteine (tHcy). Blood samples were centrifuged at 3500 x g for 10 min at 4 °C within 10 min, and the plasma was stored at -70°C until assayed. Samples were thawed and analyzed in batches to ensure that all 4 samples from each individual were assayed together to reduce assay variation. Homocysteine was measured by enzymatic immunoassay and plasma folate by a competitive protein binding method with an Abbott IMX instrument (Abbott Diagnostics Ltd, Maidenhead, Berkshire, United Kingdom). The between-batch CVs for these measurements were 5.3% and 9.3%, respectively. Plasma folate results were converted from µg/L to nmol/L by using a conversion factor of 2.265 as recommended in the IMX folate method sheet.
Statistical analysis
The primary comparison of the effect of the 3 folate intake regimens was by three-way analysis of variance (ANOVA) appropriate to the crossover design of the study, with modeling on subject, period, and treatment and with subjects of all genotypes analyzed together. Subsequently, baseline and postintervention folate intakes were compared by two-way ANOVA. Pairs of interventions were then compared by using Wilcoxon's paired nonparametric tests with Bonferroni adjustment for multiple comparisons. Intervention-by-genotype interactions were investigated by repeated-measures ANOVA. Interaction effects were further elucidated by comparing differences between the positive (folate-rich phase and supplement phase) and negative (baseline and exclusion phase) interventions between genotypes. Associations between different variables were assessed by Spearman's rank correlation. Summary statistics presented in the text are means ± SDs. Statistical analyses were performed by using SPSS 7.5 (SPSS Inc, Chicago).
RESULTS
The study group comprised 53 men and 73 women aged 20–63 y (
Effect of interventions
Primary analyses (ANOVA) indicated a significant effect of intervention and subject on total dietary folate intake (subject, P 0.001; intervention, P 0.001), plasma folate (subject, P 0.001; intervention, P 0.001), and tHcy (subject, P 0.001; treatment, P 0.001) but no significant effect of treatment order.
Effect of interventions: dietary and supplement intake
Dietary intakes of folate and folic acid throughout the study are presented in Table 1.
View this table:
TABLE 1 . Dietary folate and folic acid intake by intervention1
During the folate-rich diet, 82% of subjects reached the target intake of 400 µg/d (575 DFE) by using combinations of fortified breads, cereals, and a small number of other fortified products together with naturally folate-rich foods. The intake of folate from fortified sources was greater (by 246 µg/d, or 417 DFE) during the folate-rich diet than during the exclusion diet (P 0.001). Folate intake from natural sources was also higher (by 25 µg/d) during the folate-rich diet than during the other 2 phases (P 0.001) but was not significantly different from baseline.
Changes (post folate-rich diet - baseline) in estimated total dietary folate intake and in fortified (folic acid) intake were correlated with change in plasma folate (r = 0.20, P < 0.05, n = 108, and r = 0.31, P = 0.001, n = 108, respectively). Intake of folate-fortified foods dropped during both the supplement and exclusion diet phases to <20% of baseline (P 0.001), as expected, with a corresponding reduction in total dietary folate intake.
Compliance with the folic acid supplements or placebo tablets during the relevant phases was high, with subjects taking 88% of the tablets dispensed as determined by returned tablet counts. This corresponds to an actual folic acid intake from supplements of 352 µg/d (599 DFE), contributing to an estimated total folate intake of 561 µg/d (814 DFE).
Effect of interventions: biochemical measurements
The biochemical measurements after the interventions are given in Table 2. Plasma folate was significantly higher (P 0.001) and tHcy significantly lower (P 0.001) after the folate-rich diet than after the exclusion diet. After the supplement phase, plasma folate concentrations significantly exceeded those during the folate-rich diet phase (P 0.001). However, tHcy was only slightly lower (0.2 µmol/L) after the supplement phase than after the folate-rich diet phase, and the difference was not significant (P = 0.72).
View this table:
TABLE 2 . Effect of folate intervention on plasma folate and plasma homocysteine in the whole group and in subgroups stratified by methylenetetrahydrofolate reductase (MTHFR) genotype1
Influence of MTHFR genotype on response to intervention
For all treatments, tHcy was negatively correlated with plasma folate for the whole group (baseline: r = -0.50, P = 0.001; exclusion diet: r = -0.50, P = 0.001; folate-rich diet: r = -0.46, P = 0.001; supplement r = -0.36, P = 0.001). TT homozygotes exhibited the strongest correlations between tHcy and plasma folate for all treatments (baseline: r = -0.6, P = 0.001; exclusion diet: r = -0.80, P = 0.001; folate-rich diet: r = -0.72, P = 0.001; supplement: r = -0.53, P = 0.001). Weaker correlations were observed in the subjects with the CT and CC genotypes.
The gradient in plasma folate observed between genotypes at baseline (CC > CT > TT) was maintained after each intervention (Figure 2). There was no obvious difference in response between genotypes.
FIGURE 2. . Mean (±SD) plasma folate concentrations after the dietary interventions (in subgroups divided by methylenetetrahydrofolate reductase genotype) in 126 healthy men and women (
The subjects with the TT genotype had the highest tHcy concentrations at baseline (P 0.001 compared with the CT or CC subjects). This difference between genotypes persisted throughout each intervention, but was less marked with increasing folate intake and was not significant after the supplement phase (TT compared with CC after supplementation, P = 0.097; Figure 3 and Table 2).
FIGURE 3. . Mean (±SD) total plasma homocysteine concentrations after the dietary interventions (in subgroups divided by methylenetetrahydrofolate reductase genotype) in 126 healthy men and women (
To explore the interaction between intervention and genotype for plasma folate and tHcy, two-way repeated-measures ANOVA models including interaction were fitted to the plasma folate and tHcy data. These showed a highly significant difference between the 3 genotypes for folate (P = 0.004) and tHcy (P < 0.001) and between each of the 4 measurement points (baseline and the 3 interventions; P < 0.001 for both). The interaction effect was significant for tHcy (P = 0.02) but not for folate (P = 0.32). Similar results were obtained in confirmatory analysis of log-transformed folate and tHcy concentrations.
To further elucidate the interaction effect for tHcy, unpaired t tests were performed comparing postintervention differences between the positive interventions (folate-rich diet and supplement) and the negative ones (baseline and exclusion diet) between the TT homozygotes and the remaining subjects. Differences were significant for supplement compared with baseline (P = 0.045) and for supplement compared with exclusion diet (P = 0.029), but not for folate-rich diet compared with baseline or exclusion diet (P = 0.141 and P = 0.086, respectively), with similar results for log tHcy and with the use of nonparametric (Mann-Whitney) tests.
After both positive interventions, TT homozygotes still had slightly higher plasma tHcy concentrations than CC subjects had at baseline (folate-rich diet, P = 0.034; supplement, P = 0.049). Conversely, the TT subjects had plasma folate concentrations significantly higher than the baseline concentrations of CC subjects after both active interventions (folate-rich diet, P < 0.05; supplement, P 0.001).
DISCUSSION
Advice to increase dietary intake of folate (mainly from foods fortified with folic acid) and low-dose folic acid supplementation significantly increased plasma folate and significantly decreased tHcy concentrations in healthy, free-living subjects. This is the first intervention study to systematically investigate the effect of the MTHFR 677CT genotype on the response to such interventions in a study population selected on the basis of and balanced for MTHFR genotype. This study, therefore, provides useful data to inform the debate on folic acid fortification for public health reasons and of supplementation for specific indications, which may be relevant to the prevention of neural tube defects or cardiovascular disease. Our data support suggestions from observational studies that TT homozygotes have a greater folate requirement than do their CC and CT counterparts. This is especially evident when folate-fortified foods are excluded from the diet. However, active folate interventions, either foods fortified with folic acid or low-dose folic acid supplements, produce tHcy concentrations in TT homozygotes that approach the prevalent concentrations in CC homozygotes. Therefore, despite the relative disadvantage of TT homozygotes at baseline and after the exclusion diet, we conclude that these individuals respond well to folate-enhancing interventions but require a greater total folate intake.
Both the folate-rich diet and the supplement were acceptable to the subjects and compliance was good as estimated by dietary questionnaire and capsule counting, respectively. This was corroborated by marked increases in plasma folate in these intervention groups. The targets for total folate intake of 400 µg/d (575 DFE) and 550 µg/d (830 DFE) for these 2 interventions were generally achieved.
The changes in blood measures of folate status after the folate-rich diet (compared with baseline) were related to the estimated change in dietary intake, but the lack of correlation between estimated dietary intake and plasma measurements at baseline was not expected. This finding is in contrast with previous observations made by using the same methods in a group of patients being treated for hyperlipidemia (22). The explanation for this difference is unclear but may reflect in part the discrepancies between the true and quoted bioavailability estimates for natural folate in a range of different foods, in contrast with fortified foods in which the folic acid content is relatively accurately known and the bioavailability is high. Alternatively, the higher frequency of TT homozygotes in the present cohort compared with the proportion in the general population (33.3% compared with 12%) may have influenced the relation between plasma folate and dietary folate intake at baseline in this particular study. TT homozygosity is associated with reduced plasma folate compared with the CC and CT genotypes despite similar dietary intakes (23,24). The reduced conversion of food folates to bioactive folate in the TT homozygotes may therefore have lowered the correlation between these 2 variables, particularly at the lower dietary folate intakes observed at baseline.
During the folate-rich diet, subjects relied mainly on fortified ready-to-eat breakfast cereals to increase their dietary folate intake. These were generally preferred and were more widely available in retail outlets in the United Kingdom than were the fortified breads. During the exclusion diet, a small number of subjects continued to consume some fortified cereals either in error or by preference, but the overall intake of fortified foods was <20% of that at baseline.
The supplement had a significantly greater effect on plasma folate than did the folate-rich diet, although the supplement seemed to have minimal additional tHcy-lowering capacity compared with diet (folate rich diet, -14%; supplement, -16%). Several studies have suggested that the tHcy-lowering effect of folate reaches a plateau at intakes of 400 µg/d, with higher doses of folate producing little additional benefit (5,12,13). These results are consistent with such a plateau effect, but this was not a dose-finding study and our results could also be explained by other factors such as differences in bioavailability between the folic acid incorporated in a food matrix and that in supplements (tablets). Furthermore, the actual amount of folic acid added to fortified foods may vary depending on the manufacturer and the overage applied to the product to maintain minimum recoveries during processing and shelf life (25,26).
Several previous intervention studies showed that pretreatment tHcy and plasma folate concentrations influence the tHcy-lowering potential of folate interventions but did not report the genotype of the subjects involved. In these studies, subjects with the highest tHcy and lowest plasma folate concentrations at baseline had the greatest reduction in tHcy after the intervention (5,27). Our baseline data show that the TT homozygotes tended to have low plasma folate and high plasma tHcy. Hence, the greater response to folate in these subjects is in keeping with previous observations and indeed provides a biochemical explanation for such a response.
We observed an additive effect of folate intervention on plasma folate across genotypes. The folate-rich and supplement interventions significantly increased plasma folate in the TT homozygotes above the baseline concentrations observed in those with the CC genotype. However, we did not observe a greater response in the TT subjects than in the subjects with the other genotypes. In contrast, there was an interaction effect of intervention and genotype for tHcy. The TT homozygotes exhibited a greater tHcy-lowering response to active folate interventions than did the subjects with the other genotypes. In terms of tHcy concentrations, the baseline disadvantage of TT homozygotes was significantly reduced after supplementation, with the folate-rich diet producing a broadly similar effect that was nearly significant. Overall, the active dietary interventions reduced tHcy concentrations in TT homozygotes almost to the baseline values of those subjects with the CC genotype. These findings reflect a greater requirement of folate by TT homozygotes to achieve tHcy concentrations comparable with the prevalent concentrations observed in persons with the CT and CC genotypes.
The apparent increased folate requirement of individuals with the MTHFR TT genotype is an important factor that must be considered in food fortification and supplementation policy. Any policy should be appropriate to treat such subjects because persons with this genotype constitute 12% of the population, and affected individuals are expected to be at greatest risk of the consequences of low folate status. Assuming that the tHcy concentration is a responsive marker of overall folate status, the near-normal tHcy values reached by TT subjects in this study suggest that a folate-enrichment policy that can help persons achieve intakes between 400 and 600 µg/d (575–830 DFE) would be appropriate for individuals with the TT genotype.
ACKNOWLEDGMENTS
We thank Jan Chapman for laboratory assistance.
REFERENCES