点击显示 收起
1 From the Departments of Nutrition (SEC, ELG, DJH, MJS, WCW, and EBR) and Epidemiology (ELG, DJH, SEH, MJS, WCW, and EBR), Harvard School of Public Health, Boston, MA, and the Division of Preventive Medicine (SEC) and the Channing Laboratory (ELG, DJH, SEH, MJS, WCW, and EBR), Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA
2 Supported by grants no. CA87969, CA100971, AA11181, and HL07575 from the National Institutes of Health and by supplemental funding from Merck Research Laboratories to help defray the cost of the plasma assays.
3 Reprints not available. Address correspondence to SE Chiuve, Harvard School of Public Health, 665 Huntington Avenue, Boston, MA 02115. E-mail: schiuve{at}hsph.harvard.edu.
See corresponding editorial on page 3.
ABSTRACT
Background: Folate intake increases plasma folate and reduces total homocysteine (tHcy) concentrations, which may lower coronary artery disease (CAD) and cancer risks. Folate metabolism may be altered by alcohol intake and 2 common polymorphisms in the methylenetetrahydrofolate reductase (MTHFR) gene, 677CT and 1298AC.
Objective: We examined whether the associations between folate intake and plasma folate and tHcy concentrations were modified by alcohol intake or variations in the MTHFR gene.
Design: We conducted a cross-sectional analysis among 988 women by using multivariate linear regression models to estimate mean plasma tHcy and folate concentrations. Folate intake was the sum of food and supplemental sources.
Results: We observed an inverse association between folate intake and tHcy, which was modified by alcohol intake (P for interaction = 0.04) and MTHFR677 genotype (P for interaction = 0.05) but not by MTHFR1298 genotype (P for interaction = 0.97). In the lowest quintile of folate intake, moderate drinkers (15 g alcohol/d) had significantly higher tHcy concentrations (15.2 ± 2.9 nmol/mL) than did light drinkers (11.3 ± 0.7 nmol/mL) and nondrinkers (11.0 ± 0.8 nmol/mL). However, the reduction in tHcy between the highest and lowest quintiles of folate intake was significantly greater in moderate drinkers (–6.6 nmol/mL) than in light drinkers (–2.3 nmol/mL) and nondrinkers (–2.1 nmol/mL). The elevated tHcy in women with low folate intake who also consumed moderate amounts of alcohol was even higher (22.4 ± 4.8 nmol/mL) in the presence of the variant MTHFR677 allele. The positive association between folate intake and plasma folate was somewhat modified by alcohol intake (P for interaction = 0.08) but not by either MTHFR genotype.
Conclusions: Moderate alcohol intake and low MTHFR activity have adverse effects on tHcy, but those effects may be overcome by sufficient folate intake.
Key Words: Folate intake • homocysteine • plasma folate • methylenetetrahydrofolate reductase • MTHFR • polymorphism • alcohol
INTRODUCTION
High concentrations of circulating total homocysteine (tHcy), a sulfur-containing amino acid, are associated with increased risk of coronary artery disease (CAD) (1). High tHcy concentrations are also an indicator of reduced DNA methylation, which may be a risk factor for some cancers (2). Homocysteine can be catabolized by a vitamin B-6–dependent pathway or remethylated to methionine by a pathway that requires folate, vitamin B-12, and riboflavin (Figure 1). Folic acid supplementation lowers tHcy in a linear fashion up to 400 µg/d, but tHcy concentrations plateau at higher doses (3, 4). The positive linear association between folate intake and plasma folate extends to folic acid supplementation of >400 µg/d (3, 5).
FIGURE 1.. Homocysteine metabolism by transulfuration and remethylation. Homocysteine can be converted to cysteine by a vitamin B-6–dependent transulfuration. Homocysteine can also be remethylated to methionine through a vitamin B-12–and folate-dependent pathway. tHcy, total homocysteine; THF, tetrahydrofolate; DHF, dihydrofolate; 5,10-MTHF, 5,10-methylene THF; 5-MTHF, 5-methyl THF; MTHFR, methylenetetrahydrofolate reductase gene; MS, methionine synthase.
Polymorphisms in the gene that encodes the methylenetetrahydrofolate reductase (MTHFR) enzyme may impede homocysteine remethylation. MTHFR irreversibly converts 5,10-methylenetetrahydrofolate (5,10-MTHF) to 5-methyltetra-hydrofolate (5-MTHF), the most prevalent form of folate in circulation (Figure 1). A CT transition at base pair 677, which causes an alanine-to-valine substitution, is associated with low enzyme activity (6). Heterozygotes (CT) and homozygotes (TT) have 71% and 33%, respectively, the activity of persons with the wild-type (CC) genotype (7). Persons with the TT genotype have significantly lower plasma folate (8) and higher tHcy concentrations (9, 10) than do those with the CC genotype. Although folate effectively decreases tHcy among all genotypes, persons with the TT genotype may require higher intakes than do those with other genotypes to optimally reduce tHcy (11). A second variation at base pair 1298 (AC) also results in decreased enzyme activity, and, although this polymorphism alone may not alter plasma folate or tHcy, combined heterozygosity for MTHFR677 and MTHFR1298 may significantly influence these plasma markers (12).
Alcohol (ethanol) can interfere with folate metabolism, either directly (13) or through its metabolite acetaldehyde (14, 15). The tHcy concentration is twice as high, and plasma B vitamins are lower, among chronic alcoholics than among healthy controls, most likely as a result of a combination of malnourishment and the direct effects of heavy alcohol intake on folate status (16). The effect of moderate alcohol intake on homocysteine metabolism is unclear. Previous studies have reported positive (17-20) and inverse (21-24) associations between alcohol intake and tHcy. Our objective in this study was to assess whether the associations of folate intake with fasting plasma folate and tHcy concentrations were modified by MTHFR genotype or alcohol intake among women.
SUBJECTS AND METHODS
Study population
The Nurses' Health Study
The Nurses' Health Study (NHS) is a prospective cohort consisting of 121 700 female nurses aged 30–55 y at baseline in 1976. Participants provided medical history and lifestyle information on mailed questionnaires at the study's inception and biennially thereafter. Blood samples were collected from 32 826 women in 1989–1990. As described previously (25), the women received a kit that contained necessary supplies for venipuncture. The women returned the whole-blood samples in an enclosed ice pack via overnight mail, along with a completed short questionnaire that obtained information on menopausal status, recent postmenopausal hormone use, the time since the last meal, and the time of day of the blood drawing. On arrival, samples were centrifuged, separated, and stored in the gas phase of a liquid nitrogen freezer at –130 °C. The women in this analysis were healthy control subjects, not currently using exogenous hormones, from separate nested case-control studies of cardiovascular disease and colon neoplasia (26). Participants had no history of cancer (except nonmelanoma skin cancer), stroke, myocardial infarction, angina, or revascularization surgery before they returned their blood samples.
The Nurses' Health Study II
The NHS II (NHS2) is a prospective cohort of 116 671 female nurses aged 25–41 y at baseline in 1989. Blood samples were obtained in 1997 and 1998 from 29 613 women. The methods used to obtain lifestyle and medical information and blood samples were similar to those described above.
We analyzed blood from a subset of premenopausal women who were not users of exogenous hormones. For this sample collection, women were in the luteal phase of their menstrual cycle. Women with a history of cardiovascular disease, diabetes mellitus, gastric or duodenal ulcers, liver or gallbladder disease, or cancer (excluding nonmelanoma skin cancer) before date of blood draw were excluded. From the remaining women, we randomly selected 473 on the basis of self-reported alcohol use to study the effects of alcohol on biological markers of CAD. Specific drinking patterns determined on the basis of frequency, amount, and use with meals were oversampled for adequate variation. The characteristics of this population did not differ significantly from those of the larger cohort. Further details on the selection process were published elsewhere (27).
Written informed consent was obtained from all participants. The Institutional Review Board of the Harvard School of Public Health approved the study protocol.
Assessment of dietary and nondietary factors
We assessed dietary information with a validated semi-quantitative food-frequency questionnaire (FFQ). Average nutrient intake over the previous year was calculated from the FFQ by using nutrient values obtained from the Harvard University Food Composition Database, which was derived from the US Department of Agriculture and other sources. The reproducibility and validity of the FFQ have been documented elsewhere (28, 29). The correlations between the FFQ and multiple 1-wk diet records were 0.88 for riboflavin, 0.85 for vitamin B-6, 0.56 for vitamin B-12, 0.77 for folate, and 0.84 for alcohol intake. All nutrient intakes were adjusted for total energy by using the residual method (30). Total intake for all vitamins, including folate, was the sum of food and supplemental sources.
We used the 1990 FFQ for NHS and the 1999 FFQ for NHS2 to correspond to blood samples drawn in 1989–1990 and 1997–1998, respectively, because the FFQs assess diet over the previous year. The food-composition database was updated to reflect the fortification of grain products with folate, which began in 1996 and became mandatory in 1998 (31). Information on systolic blood pressure, use of aspirin and hypertensive medication, smoking status, physical activity, height, and weight was obtained through self-administered questionnaires, and we used the questionnaire that was completed the closest in time to the blood drawing.
Assays for plasma markers
Because this analysis was based on samples from several different datasets, some analytes were measured with different methods. We controlled for laboratory batch in our analysis. For samples from the NHS, plasma folate was measured by using a radioimmunoassay kit (Bio-Rad, Richmond, CA), and tHcy was measured by using HPLC at the Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging (Tufts University, Boston, MA). For the NHS2 samples, both plasma folate and tHcy concentrations were measured by using an immunoassay on an IMx analyzer (Abbott Laboratories, Abbott Park, IL) at the laboratory of Nadar Rifai (Children's Hospital, Boston MA). Quality-control samples (5% of all samples), obtained from a plasma pool from healthy volunteers, were given indicator identification numbers and interspersed randomly among the specimens. The CVs were <10% for plasma folate and tHcy.
DNA was extracted from buffy coat fractions, and MTHFR677 and MTHFR1298 genotypes were assessed by using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). Primers and probes are available from the authors on request. Polymerase chain reaction amplifications were carried out on 5–20 ng DNA using 1x TaqMan universal PCR master mix (Applied Biosystems). Amplification conditions were 1 cycle of 95 °C for 10 min and then 50 cycles of 92 °C for 15 s and 58 °C for 1 min. The frequencies of both genotypes were in Hardy-Weinberg equilibrium.
Exclusions
Women whose information on folate intake or smoking status was missing and those who did not fast 6 h before the blood drawing were excluded from all analyses. Women whose concentrations of plasma folate or tHcy or information on alcohol intake or MTHFR genotype was missing were excluded only from analyses that required these specific variables. The total population consisted of 1146 women. Of these women, 988 were included in the homocysteine analysis and 966 in the plasma folate analysis.
Statistical analysis
We used multivariate linear regression to calculate mean concentrations of plasma tHcy and folate in each quintile of folate intake. Robust variance estimates were used for valid statistical inference of linear regression models, despite a lack of normality in the outcome variable (32). In multivariate models, we adjusted for age (centered at 50 y); smoking status (never or past, current <15 cigarettes/d, or current 15 cigarettes/d); regular aspirin use (yes or no); hypertensive medication use (yes or no); intakes of riboflavin, vitamins B-6 and B-12 (centered at median), methionine (above or below median of 1.9 g/d), and coffee (cups/d); laboratory batch; MTHFR genotype (CC, CT, or TT); and alcohol intake (0 g/d, 0.01–14.9 g/d, or 15.0 g/d). Further adjustment for tea intake, physical activity, body mass index, total dietary protein, and dietary fat intake had no appreciable effect on results (data not shown). To conduct a test for trend, we created a continuous score variable for folate intake by using the median value from each quintile.
We stratified the predictive models for plasma tHcy and folate by MTHFR677 genotype, assuming a codominant mode of inheritance (CC, CT, and TT). To test formally for interaction, we included in our models the multiplicative interaction term of folate intake (as a continuous variable, for which the median value for each category of folate intake was used) and MTHFR677 genotype (as a continuous variable with 3 categories). We performed similar analyses that were stratified by MTHFR1298 genotype (AA, AC, and CC) and alcohol intake (0 g/d, 0.01–14.9 g/d, and 15 g/d). To evaluate the interaction between alcohol intake and MTHFR677 genotype, we stratified models of plasma tHcy and folate by alcohol, separately in women with the CC genotype and in women with at least one copy of the variant allele. We had limited power to examine this interaction among women with the TT genotype alone because of the limited number of participants with this genotype (n = 34, 60, and 15 for nondrinkers, light drinkers, and moderate drinkers, respectively). Therefore, we combined women with CT and TT genotypes. We included the multiplicative interaction term of alcohol and MTHFR677 genotype (CC or CT/TT) in our models. We performed a similar analysis with alcohol and the MTHFR1298 genotype. All P values are two-tailed. Statistical analyses were conducted with SAS software (version 8; SAS Institute, Cary, NC).
RESULTS
The age-adjusted characteristics of the population used in the homocysteine analysis are shown by quintile of folate intake in Table 1. The distribution of these characteristics did not differ significantly within the population for the plasma folate analysis (data not shown). The median folate intake of women in the lowest quintile was 241 µg/d, and that of women in the highest quintile was 856 µg/d. Women with high folate intake tended to be younger than those with low folate intake. More women from the NHS2 than from the NHS contributed to the higher folate quintiles because blood and dietary information were collected from the former group after the fortification of grain products. As expected, the intake of other B vitamins increased as folate intake increased. The average alcohol intake did not differ significantly across quintiles of folate intake. Except for the CT variant of MTHFR677, frequencies of the MTHFR677 and MTHFR1298 genotypes did not differ across quintiles of folate intake. The 2 polymorphisms were in linkage disequilibrium; among the women with the TT genotype for MTHFR677, only 1 carried the variant allele for the MTHFR1298 polymorphism, and none of the women who were homozygous for a variant of MTHFR1298 did so. Folate intake was inversely associated with plasma tHcy and positively associated with plasma folate in both age-adjusted and multivariate models (Table 2).
View this table:
TABLE 1. Selected characteristics of subjects across quintiles of energy-adjusted folate intake1
View this table:
TABLE 2. Adjusted mean values of plasma total homocysteine (tHcy) and folate across quintiles of energy-adjusted folate intake
Modification of the folate intake–tHcy relation by MTHFR genotypes and alcohol intake
The MTHFR677 polymorphism modified the dose-response association between folate intake and tHcy (P for interaction = 0.05) (Figure 2). The inverse association between folate intake and tHcy was strongest among the women with the variant allele. We observed a reduction in mean tHcy from 13.2 nmol/mL in the first quintile to 8.8 nmol/mL in the fifth quintile among women who were heterozygous for the CT variant and from 11.6 to 8.3 nmol/mL among those who were homozygous for the TT variant. We observed the smallest reduction in tHcy (from 10.9 to 9.3 nmol/mL) between the lowest and highest quintiles of folate intake among women who had the wild-type MTHFR677 genotype (CC). We did not observe an interaction between folate intake and the MTHFR1298 genotype for tHcy (P for interaction = 0.97).
FIGURE 2.. Adjusted mean plasma total homocysteine (tHcy) according to quintiles of energy-adjusted folate intake by methylenetetrahydrofolate reductase (MTHFR) 677 genotype (CC: n = 421, ; CT: n = 407, ; TT: n = 109, ). Mean tHcy values were calculated by linear regression models adjusted for age; intakes of riboflavin, vitamins B-6 and B-12, methionine, coffee, and alcohol; smoking status; regular aspirin use; hypertension medication use; and laboratory batch. The SEs of mean tHcy within quintiles of folate intake ranged from 0.6 to 0.7 for CC, from 0.9 to 1.2 for CT, and from 1.1 to1.3 for TT genotype. P for trend < 0.01 for CC, < 0.0001 for CT, and = 0.004 for TT. Folate (continuous variable based on median value of quintiles) x MTHFR (in 3 categories) interaction, P = 0.05. *,**Significantly different from the first quintile of folate intake within the same genotype: *P < 0.01, **P < 0.05.
Compound heterozygosity (CT677/AC1298) was not associated with a greater response to folate intake. Among women who were heterozygous for the CT variant of MTHFR677, the trend in mean tHcy across quintiles of folate intake did not differ significantly between women who had the wild-type AA variant and those who were heterozygous for the AC variant of MTHFR1298 (P for interaction = 0.38).
Alcohol intake significantly modified the association between folate intake and tHcy (P for interaction = 0.04) (Figure 3). Higher folate intake was only modestly associated with lower tHcy among light drinkers and nondrinkers, whereas, among moderate drinkers, this inverse relation was much stronger. Among moderate drinkers, we observed a reduction in mean tHcy from 15.2 nmol/mL in the first quintile to 8.9 nmol/mL in the fifth quintile; the decrease in tHcy from the first quintile to the fifth quintile of folate intake was from 11.3 to 9.0 nmol/mL among light drinkers and from 11.0 to 8.9 nmol/mL among nondrinkers. The steepest decline in tHcy among the moderate drinkers was seen with low folate intake (quintiles 1 and 2), whereas the dose-response curve of moderate drinkers began to approximate the curves of light drinkers and nondrinkers in the third quintile of folate intake (400 µg/d).
FIGURE 3.. Adjusted mean plasma total homocysteine (tHcy) according to quintiles of energy-adjusted folate intake by alcohol intake categories (0 g/d, , n = 300; <15 g/d, , n = 522; and 15 g/d, , n = 163). Mean tHcy values were calculated by linear regression models adjusted for age; intakes of riboflavin, vitamins B-6 and B-12, methionine, and coffee; smoking status; regular aspirin use; hypertension medication use; methylenetetrahydrolate reductase 677 genotype; and laboratory batch. The SEs of mean tHcy within quintiles of folate intake ranged from 0.7–0.9 to 0.6–0.7 and to 1.7–3.1 for intakes of 0, <15, and 15 g alcohol/d, respectively. P for trend = 0.0002, < 0.001, and < 0.0004 for intakes of 0, 0.01–14.9, and 15 g alcohol/d, respectively. Folate (continuous variable based on median value of quintiles) x alcohol intake (in 3 categories) interaction, P = 0.04. *,**Significantly different from the first quintile of folate intake within the same alcohol intake category: *P < 0.01, **P < 0.05.
Women with heavy alcohol intake may have a different dose response between folate intake and tHcy, as well as a different set of potential confounders of the dose response, than do moderate drinkers. To remove this extraneous variation, in a secondary analysis, we excluded women who drank >50 g alcohol/d (n = 11). However, results from this analysis were not appreciably different (data not shown).
The inverse association between folate intake and tHcy among the moderate alcohol drinkers was primarily limited to the women with 1 variant allele for MTHFR677 (alcohol x genotype interaction, P = 0.01) (Table 3). Although the moderate drinkers with the variant allele had elevated tHcy (22.4 nmol/mL) at low folate intake, tHcy was no longer elevated with high folate intake, especially when the heaviest drinkers (50 g alcohol/d; n = 11) were excluded from the analysis. The MTHFR1298 x alcohol interaction for tHcy was not significant (P = 0.72).
View this table:
TABLE 3. Adjusted mean values of total homocysteine (tHcy) by alcohol consumption and MTHFR677 genotype across quintiles of energy-adjusted folate intake1
Modification of the folate intake–plasma folate relation by MTHFR genotypes and alcohol intake
The relation between folate intake and plasma folate was not modified by the MTHFR677 (P for interaction = 0.97) or MTHFR1298 (P for interaction = 0.13) genotype. This relation was influenced by alcohol, although the interaction was not significant (P = 0.08) (Figure 4); the overall positive association was strongest among nondrinkers. We did not see further modification by variations in either the MTHFR677 (alcohol x genotype interaction, P = 0.49) or MTHFR1298 (alcohol x genotype interaction, P = 0.97) genotype.
FIGURE 4.. Adjusted mean plasma folate concentrations according to quintiles of energy-adjusted folate intake by categories of alcohol intake (0 g/d, , n = 288; <15 g/d, , n = 514; 15 g/d, , n = 163). Mean plasma folate values were calculated by linear regression models adjusted for age; intakes of riboflavin, vitamins B-6 and B-12, methionine, and coffee; smoking status; regular aspirin use; hypertension medication use; methylenetetrahydrofolate reductase 677 genotype; and laboratory batch. The SEs of mean plasma folate within quintiles of folate intake ranged from 1.3–2.8 to 0.7–1.3 and to 1.3–2.8 for intakes of 0, <15, and 15 g alcohol/d, respectively. P for trend < 0.0001 for all alcohol intake categories. Folate (continuous variable based on median value of quintiles) x alcohol intake (in 3 categories) interaction, P = 0.08. *,**Significantly different from the first quintile of folate intake within the same alcohol intake category: *P < 0.0001, **P < 0.01.
DISCUSSION
Folate intake was inversely associated with fasting plasma tHcy, and this relation was modified by MTHFR677 genotype and alcohol intake. Women with the thermolabile variant of MTHFR677 or moderate alcohol intake had significantly higher tHcy at low intakes of folate. However, adequate folate intake minimized these differences. In addition, the elevation in tHcy among women who consumed low folate and drank moderate amounts of alcohol was greater in the presence of the variant MTHFR677 allele.
To our knowledge, this is the first study to examine the associations of the combination of folate intake, alcohol intake, and the MTHFR polymorphisms with plasma tHcy and folate. In previous studies, the association between alcohol intake and tHcy has been unclear (17-24), possibly because of differences in folate intake or a lack of stratification by folate and MTHFR genotype. The increase in tHcy among moderate drinkers with low intake of folate was restricted to women with the thermolabile variant of MTHFR. In women who had the wild-type MTHFR677 variant, tHcy concentrations were not significantly elevated among moderate alcohol drinkers.
Ideally, we would have examined separately the modification by alcohol intake of the relation between folate intake and both tHcy and plasma folate in the 3 MTHFR677 genotypes. Conventionally, persons with the CT genotype are combined with those who have the CC rather than the TT genotype (10, 33). However, because of the limited number of women who were homozygous for the variant allele, we were unable to examine any associations in these women separately. By combining subjects with the CT and TT genotypes, we could assess the combined effect of the variant allele and alcohol intake on these relations.
Significant interactions between alcohol intake, MTHFR677 genotype, and folate status have also been seen in studies of CAD. The MTHFR677 polymorphism is an independent risk factor for CAD only among those with low folate status (34). Women with both high folate intake and moderate alcohol intake had a significantly lower risk of CAD than did nondrinkers with low folate intake (35). The effect of these interactions between folate intake, alcohol intake, and MTHFR genotype on the risk of CAD may be partially mediated through tHcy.
The evidence of an interaction between alcohol intake, MTHFR677 genotype, and folate status is equally strong with respect to cancer. Alcohol intake significantly increases the risk of breast (36, 37) and colon (38) cancer, although the association appears limited primarily to persons with low folate intake. The MTHFR677 polymorphism has been associated with reduced risk of colon cancer, but only in combination with a methyl-rich diet (low alcohol or high folate intake or both) (39). Persons with the variant MTHFR677 genotype may be more susceptible to colon cancer if they have a methyl-poor diet (ie, high alcohol or low folate intake or both). These interactions in relation to cancer are not likely to be due to a biological effect of homocysteine but rather to the importance of folate on gene methylation. A methyl-poor diet or reduced MTHFR activity may enhance carcinogenesis through a reduction in the 5-MTHF form of folate and a subsequent inhibition of homocysteine remethylation (Figure 1). Reduced remethylation could result, in addition to an accumulation of tHcy, in a decrease in methionine, which is used in DNA methylation. The markedly higher tHcy concentrations we observed in the methyl-poor subgroup in our study may reflect aberrations in DNA methylation, a potential risk factor for cancer (2).
Alcohol may interfere with folate metabolism through a reduction in folate absorption at the brush border (40) or an inhibition of the methionine synthase enzyme, which is needed to transfer a methyl group from 5-MTHF to homocysteine (41). The inhibition of this enzyme by alcohol traps folate in its 5-MTHF form (13) and may result in a 5-MTHF pool that cannot remethylate homocysteine (Figure 1). Inhibition of methionine synthase may be overcome by adequate plasma folate, achieved through sufficient folate intake or with efficient MTHFR activity, such as that among those who are wild-type.
As had earlier investigators (42), we found the MTHFR677 polymorphism to be a stronger determinant of tHcy than was the MTHFR1298 polymorphism. Combined heterozygosity of MTHFR677 and MTHFR1298 has been associated with higher tHcy concentrations than were seen with the MTHFR677 variant alone (43, 44). However, we did not observe a stronger association between folate intake and tHcy in those subjects with combined heterozygosity than in those with only the MTHFR677 variation.
Betaine, derived from choline, may lower tHcy independent of the folate remethylation pathway (45), especially in the presence of ethanol (46). Moreover, its metabolism may be influenced by MTHFR genotype (47). We were unable to assess betaine or other factors that may influence tHcy and contribute to its variability. Unless correlated with folate intake, these factors would not affect the magnitude of the association between folate intake and tHcy.
Others have reported an interaction between folate status and alcohol intake in relation to tHcy, but only among men (48), which may be due to the larger variation in alcohol intake among men than among women. To increase variation in alcohol intake, we oversampled the drinkers in the NHS2 population. Because blood was drawn both before and after folate fortification, we also have a wider range of folate intakes than had been previously studied. The median of our lowest quintile of folate intake (241 µg/d) may still be high with respect to that in other populations, especially those without fortification programs. The elevation of tHcy at low folate intake among alcohol drinkers or among persons with the MTHFR variant may be even greater in other populations.
Although everyone should consume sufficient folate, these findings specifically highlight the necessity of adequate folate intake among moderate drinkers. In our study, when all groups had a folate intake of 400 µg/d, tHcy concentrations among moderate drinkers began to approach those among light drinkers and nondrinkers. Alcohol drinkers who have the variant MTHFR677 allele may require even higher folate intakes.
We observed that, among moderate drinkers, tHcy was no longer elevated with high folate intake. We observed this relation only after excluding heavy alcohol consumers (50 g/d) from the analysis. This finding suggests that, whereas the adverse effects of moderate alcohol may be overcome through adequate folate intake, the same may not be true with excessive intake of alcohol.
In conclusion, concentrations of plasma folate and tHcy are primarily determined by the intake of folate. However, alcohol intake and genetically determined MTHFR activity can modify these dose-response associations. Future research on the health effects of folate intake on tHcy, CAD, or cancer should incorporate these potential risk modifiers.
ACKNOWLEDGMENTS
We thank the participants of the NHS and NHS2 for their cooperation and participation.
SEC was responsible for the design of the study, analysis of the data, and writing the manuscript; ELG, SEH, DJH, MJS, and WCW were responsible for critical review of the manuscript. EBR was responsible for securing funding, design of the study, analysis of the data, and writing the manuscript. No authors had any financial or personal interest in any organizations sponsoring the research reported in this article.
REFERENCES
Related articles in AJCN: