点击显示 收起
1 From INSERM, U476, and INRA 1260, Université de la Méditerranée, Marseille, France (DL and RP), and INSERM, U557, and CNAM, ISTNA, Paris, France (NA, SB, EC, SH, and M-CB-R)
2 Supported by Fruit d'Or Recherche, Lipton, Cereal, Candia, Kellogg's, CERIN, LU/Danone, Sodexho, L'Oréal, Estée Lauder, Peugeot, Jet Service, RP Scherer, France Telecom, Becton Dickinson, Fould Springer, Boehringer Diagnostic, Seppic Givaudan Lavirotte, Le Grand Canal, Air Liquide, Carboxyque, Klocke, Trophy Radio, Jouan, and Perkin-Elmer during the establishment and follow-up of the SU.VI.M.AX cohort. Bio-Rad (Vitry-sur-Seine, France) and Carlo Erba France (Val-de-Reuil, France) provided the materials needed to analyze the blood samples. 3 Reprints not available Address correspondence to D Lairon, UMR 476 INSERM/1260 INRA, Faculté de Médecine, 27 Bd Jean Moulin, 13385 Marseille Cedex 5, France. E-mail: denis.lairon{at}medecine.univ-mrs.fr.
ABSTRACT
Background: Increased consumption of dietary fiber is widely recommended to maintain or improve health, but knowledge of the relation between dietary fiber sources and cardiovascular disease risk factors is limited.
Objective: We examined the relation between the source or type of dietary fiber intake and cardiovascular disease risk factors in a cohort of adult men and women.
Design: In a cross-sectional study, quintiles of fiber intake were determined from dietary records, separately for 2532 men and 3429 women. Age- and multivariate-controlled logistic models investigated the odds ratios of abnormal markers for quintiles 2-5 of fiber intake compared with the lowest quintile.
Results: The highest total dietary fiber and nonsoluble dietary fiber intakes were associated with a significantly (P < 0.05) lower risk of overweight and elevated waist-to-hip ratio, blood pressure, plasma apolipoprotein (apo) B, apo B:apo A-I, cholesterol, triacylglycerols, and homocysteine. Soluble dietary fiber was less effective. Fiber from cereals was associated with a lower body mass index, blood pressure, and homocysteine concentration; fiber from vegetables with a lower blood pressure and homocysteine concentration; and fiber from fruit with a lower waist-to-hip ratio and blood pressure. Fiber from dried fruit or nuts and seeds was associated with a lower body mass index, waist-to-hip ratio, and fasting apo B and glucose concentrations. Fiber from pulses had no specific effect.
Conclusion: Dietary fiber intake is inversely correlated with several cardiovascular disease risk factors in both sexes, which supports its protective role against cardiovascular disease and recommendations for its increased consumption.
Key Words: Dietary fiber • cardiovascular disease • diabetes • lipids • obesity
INTRODUCTION
The recent and drastic changes that have occurred in industrialized countries have led to dietary patterns that are based mainly on animal foodstuffs and refined-cereal products and are low in dietary fiber. Nevertheless, as early as the 1970s, Burkitt and Trowell (1) observed major differences in the incidence of several diseases between native and white populations in Africa. Accordingly, the hypothesis was made that a lack of dietary fiber could be involved in the etiology of some Western diseases such as constipation, diverticulosis, cardiovascular disease, and cancers.
During the past decade, numerous key epidemiologic studies relating dietary fiber intake and cardiovascular disease have been published. One of the pioneer studies showed a 4-fold lower death rate from cardiovascular disease in men who ingested high amounts of fiber (37 g/d) than in those who ingested the lowest amounts (<20 g/d) (2). More recently, a 6-y prospective study conducted in a large cohort of adult men aged 40–75 y in the United States showed an inverse relation between fiber intake and cardiovascular death rate, with an odds ratio (OR) of 0.6 for a fiber intake of 30 g/d compared with the lowest quintile of intake (10–17 g/d) (3). Similar data have been obtained in other prospective cohort studies conducted in young men in the United States (4) or in European adults (5). In the latter study, a 3-g increase in the daily intake of soluble fiber (SF) was associated with a 27% reduction in cardiovascular disease mortality. A 10-y prospective cohort study performed in a large cohort of women aged 37–64 y in the United States (6) showed that women in the highest fiber intake quintile (22.9 g/d) had a 34% lower risk of coronary heart disease than did those in the lowest quintile (11.5 g/d). This observation was confirmed by other prospective cohort studies in US adult women (7) or men and women (8). A recent meta-analysis of some recent prospective cohort studies confirmed the protective effect of dietary fiber against cardiovascular disease (9).
In most prospective cohort studies, cereal fiber was strongly associated with the observed reductions in cardiovascular events and deaths. This finding is supported by recent data, which showed a protective effect of whole grains (10, 11). However, experimental and clinical studies showed that SF from various sources has mostly metabolic effects (12-15).
In fact, limited epidemiologic data regarding dietary fiber and cardiovascular disease risk factors are available. The aim of our study was thus to investigate the association between risk factors for cardiovascular disease and the different types [total fiber, SF, or nonsoluble fiber (NSF)] and sources (cereal products, legumes, vegetables, fresh fruit, dried fruit or nuts, and seeds) of dietary fiber in a cross sectional study of 5961 subjects included in the French SU.VI.MAX (SUpplementation en VItamines et Mineraux AntioXydants) cohort.
SUBJECTS AND METHODS
Subjects
The subjects were participants in the SU.VI.MAX Study, an ongoing randomized, double-blind, placebo-controlled, primary-prevention trial that was designed to evaluate the effect of daily antioxidant supplementation (vitamin C, vitamin E, ß-carotene, selenium, and zinc) at near-nutritional doses on the incidence of cancer and ischemic heart disease. The design of the study was described elsewhere (16). Briefly, a total of 12 741 eligible subjects (women aged 35–60 y and men aged 45–60 y) were enrolled in 1994–1995 and followed up for 8 y. All subjects were encouraged to provide dietary data in the form of 24-h dietary recalls every 2 mo. For the present study, we elected to include only subjects who had provided at least six 24-h dietary recalls over the first 2 y of the study; we excluded dietary recalls that reported <418 kJ/d (100 kcal/d) or >25 080 kJ/d (6000 kcal/d) and further excluded subjects who had 2 of 6 recalls that reported <3344 kJ/d (800 kcal/d) in men and 2090 kJ/d (500 kcal/d) in women. The initial sample included 6634 subjects. We further excluded subjects for whom either anthropometric data or data on physical activity or tobacco intake were missing. The final study sample included 5961 subjects, 2532 men and 3429 women.
The SU.VI.MAX study was approved by the Ethical Committee for Studies with Human Subjects of Paris-Cochin (CCPPRB no. 706) and by the "Commission Nationale Informatique et Liberté" (CNIL no. 334641).
Dietary assessment
The 24-h dietary recalls were provided in a random assessment of 4 weekdays and 2 weekend days per year (17). The subjects transmitted the corresponding data via the Minitel Telematic Network (France Telecom, Paris, France) or Internet, which connected them to the main SU.VI.MAX computer server. They were helped by conversational facilities of the software and by an instruction manual for the codification of food, which included photographs for estimating portion sizes (17). Daily nutrient intakes were calculated from food consumption by using the French computerized CIQUAL food database (18), which was fully updated with data on total dietary fiber and SF (19) based on the measurement method of the Association of Official Analytical Chemists (20). We thus examined the relation between cardiovascular disease risk factors and the following types of dietary fiber: total dietary fiber, SF, NSF (calculated as the difference between total fiber and SF), and total dietary fiber from cereal products, vegetables, pulses, fruit (fresh, canned, or juice), dried fruit, and nuts and seeds.
Anthropometric, clinical, and biological data
Anthropometric and clinical data were recorded during the second year of follow-up, when about one-half of the dietary recalls were obtained. Height, weight, and waist and hip circumferences were measured on subjects in their underwear with the use of standardized procedures. Overweight corresponded to a body mass index (BMI; in kg/m2) of 25. The waist-to-hip ratio was considered to be abnormal if 0.95 in men or 0.80 in women (21). Diastolic and systolic blood pressures were measured once on each arm with the use of a standard mercury sphygmomanometer in subjects who had been lying down for 10 min, and the mean value was recorded. High blood pressure was defined as a systolic blood pressure >140 mm Hg, a diastolic blood pressure >90 mm Hg, or use of antihypertensive medication (n = 350).
Blood samples were collected in Vacutainer tubes (Becton Dickinson, Le Pont de Claix, France) during the first year follow-up (except at year 3 for homocysteine) after the subjects had fasted for 12 h. Fasting plasma glucose, total cholesterol, and triacylglycerols were measured by enzymatic methods (Technicon DAX, Tarrytown, NY). Plasma concentrations of apolipoprotein (apo) A-I and apo B were measured by immunonephelometry with the use of specific anti-sera on a BNA autoanalyzer (Behring, Deerfield, IL). Serum concentrations of lipoprotein(a) were measured with an enzyme-linked immunosorbent assay. Plasma homocysteine was measured in a randomly selected subsample of 851 men (plasma samples at year 3) with the use of HPLC and fluorometric detection with a BioRad kit (Hercule SA, Vitry sur Seine, France). Laboratory quality assurance included analysis of serum from standard pools within each run and, if available, international standards.
The biochemical variables were studied in terms of normal compared with abnormal. Acknowledged limits were used for apo A-I, apo B, cholesterol, triacylglycerols, and glucose [1.30 g/L, 1.35 g/L, 6.2 mmol/L (2.40 g/L), 1.7 mmol/L (1.50 g/L), and 7 mmol/L, respectively]. The ratio of apo B to apo A-I was considered to be abnormal if >1.0 (22). The subjects treated for diabetes or high cholesterol were classified as abnormal for fasting glucose or total cholesterol, respectively. Mixed hyperlipidemia was defined as a total cholesterol concentration >2.40 g/L or as the use of medication for the treatment of high cholesterol in combination with a triacylglycerol concentration >1.50 g/L. For total plasma homocysteine (tHcy), we considered a 10 µmol/L limit (23), which provided a sufficient proportion of abnormal values for a high study power.
Statistical analysis
Statistical analyses were carried out by using the SAS (version 8, 1999; SAS Institute Inc, Cary, NC) statistical software package. Quintiles of dietary fiber intake were established separately for each sex. Unconditional logistic regression estimated the OR and 95% CIs of having abnormal values of the studied factor, for each quintile relative to the first quintile of dietary fiber intake. Models were adjusted for sex, age, alcohol-free energy intake, saturated fatty acid intake, carbohydrate intake, tobacco use, alcohol use, and leisure-time exercise. For homocysteine, models also included folate, vitamins B-6 and B12, and coffee consumption. Tests were based on the likelihood ratio test statistic; the importance of main effects was assessed by using the Wald chi-square test statistic, and 2-sided tests and P < 0.05 were used for statistical significance.
RESULTS
In Table 1, the study population is described in terms of biological (obtained during the first year of follow-up, except for homocysteine at the third year) and clinical (obtained during the second year of follow-up) variables and prevalence of risk factors for cardiovascular disease. The prevalence of abnormal markers was much higher in men than in women. In men, the most common findings were high homocysteine (57.4%), cholesterol (50.5%), BMI (51.3%) and blood pressure (42.6%). In women, the most common findings were high cholesterol (37.4%), waist-to-hip ratio (25.6%), and homocysteine (25.3%).
View this table:
TABLE 1. Biological and anthropometric characteristics of French adults from the SU.VI.MAX (SUpplementation in VItamines et Mineraux AntioXdants) cohort1
Mean dietary fiber intake by type of fiber (total dietary fiber, SF, or NSF) or dietary source are reported in Table 2. Total dietary fiber intake ranged from 4.5 to 64.9 g/d in men with a mean (±SD) intake of 21.9 ± 7.2 g/d and from 2.4 to 65.0 g/d in women with a mean (±SD) intake of 17.9 ± 5.6 g/d. Total dietary fiber intake was mostly composed of NSF: a mean 80.1% in men and 79.9% in women. Most total dietary fiber came from cereal products (37% and 33% in men and women, respectively), vegetables (30.6% and 32.4% in men and women, respectively), and fruit (19.6% and 21.2% in men and women, respectively). The amounts of dietary fiber usually provided by legumes, dried fruit, or nuts and seeds were low.
View this table:
TABLE 2. Fiber intake in French adults from the SU.VI.MAX (SUpplementation en VItamines et Mineraux AntioXidants) cohort1
Multivariate adjusted ORs and 95% CIs for abnormal values of the studied markers of cardiovascular disease are shown in Tables 3 and 4, by quintile of total dietary fiber, SF, and NSF intakes, respectively; these values were adjusted for sex, age, alcohol-free energy intake, saturated fatty acid intake, carbohydrate intake, tobacco use, alcohol use, leisure-time exercise, and intervention supplement intake.
View this table:
TABLE 3. Odds ratios (and 95% CIs) for the relative risk of abnormal clinical markers of cardiovascular disease by quintile of dietary fiber intake1
View this table:
TABLE 4. Odds ratios (and 95% CIs) for the relative risk of abnormal biological markers of cardiovascular disease by quintile of dietary fiber intake1
BMI and waist-to-hip ratio were inversely and significantly associated with total dietary fiber or NSF intakes (Table 3). A 5-g increase in total dietary fiber resulted in a 10.6% decrease in overweight risk and a 14.7% decrease in risk of high waist-to-hip ratio. Overall, the effects were not significantly different between men and women (data not shown). The only significant interaction was between sex and NSF intake regarding risk of overweight (P for interaction < 0.01). The effect was somewhat stronger in men [OR for quintile 5 versus quintile 1 (ORQ5 versus Q1): 0.60; 95% CI: 0.41, 0.88; P for linear trend = 0.009] than in women (ORQ5 versus Q1: 0.73; 95% CI: 0.50, 1.08; P for trend = 0.18). SF intake was not associated with BMI, whereas it was negatively associated with waist-to-hip ratio (ORQ5 versus Q1: 0.80; 95% CI: 0.63, 1.02; P for trend = 0.03). The risk of high blood pressure was inversely and significantly associated with total dietary fiber intake (ORQ5 versus Q1: 0.71; 95% CI: 0.54, 0.93; P for trend = 0.02) and NSF intake (ORQ5 versus Q1: 0.68; 95% CI: 0.52, 0.89; P for trend = 0.01) but not with intake of SF (P for trend = 0.10). A 5-g increase in total dietary fiber resulted in an 11.6% decrease in high blood pressure risk.
No type of dietary fiber had an effect on apo A-I concentrations (Table 4). The risk of a high apo B concentration was significantly and negatively associated with total dietary fiber and NSF intake and borderline significantly associated with SF intake. The ORQ5 versus Q1 values were 0.68 (95% CI: 0.49, 0.95), 0.69 (95% CI: 0.49, 0.95), and 0.79 (95% CI: 0.59, 1.05), respectively. The ratio of apo A-I to apo B decreased mostly by NSF intake. A reduced risk of high cholesterol was associated with high total dietary fiber intake (ORQ5 versus Q1: 0.68; 0.53, 0.87; P for trend = 0.01) and borderline associated with NSF intake (P for trend = 0.06), whereas SF intake had no significant effect. The risk of hypertriacylglycerolemia was negatively associated with total dietary fiber intake (ORQ5 versus Q1: 0.68; 95% CI: 0.48, 0.96; P for trend = 0.01) and NSF intake (ORQ5 versus Q1: 0.65; 95% CI: 0.46, 0.92; P for trend = 0.01), but there was no effect of SF. In contrast, total dietary fiber, SF, and NSF intakes did not significantly affect lipoprotein(a) and fasting glucose concentrations.
A reduced risk of hyperhomocysteinemia was markedly associated with total dietary fiber intake (ORQ5 versus Q1: 0.50; 95% CI: 0.31, 0.82; P for trend = 0.01), NSF intake (ORQ5 versus Q1: 0.52; 95% CI: 0.32, 0.85; P for trend = 0.01), and SF intake (ORQ5 versus Q1: 0.52; 0.34, 0.80; P for trend = 0.005).
The results were not substantially modified when additionally adjusted for BMI; the only significant changes were a slightly decreased effect of total dietary fiber on the ratio of apo A-I to apo B (P for trend = 0.07) and on hypertension (P for trend = 0.06).
Considering that the recommended daily intake of dietary fiber is 25 g (24), the subjects who consumed 25 g/d fiber had a lower risk of elevated plasma cholesterol (–7%; P > 0.10), BMI (–19%, P = 0.03), waist-to-hip ratio (–17%; P = 0.06), and hypertension (–26%; P = 0.002) than did consumers of lower amounts of fiber. Overall, an increase of 5 g total dietary fiber/d decreased the risk of high apo B by 9.2%, high homocysteine by 15.4%, overweight by 10.6%, elevated waist-to-hip ratio by 14.7%, and hypertension by 11.6%.
We analyzed correlations between total dietary fiber, SF, or NSF and the whole range of BMI, waist-to-hip ratio, apo A-I, apo B, the apo B:apo A-I, lipoprotein(a), and homocysteine. All correlations were statistically significant (P < 0.001). Correlations were high for waist-to-hip ratio: R2 = 0.60 for all 3 types of dietary fiber. For BMI, the R2 values were 0.18, 0.17, and 0.18 for total dietary fiber, SF, and NSF, respectively. Values were identical for all 3 types of dietary fiber: R2 = 0.14 for apo A-I, 0.15 for apo B, 0.20 for apo B:A-I, 0.10 for homocysteine, and 0.005 for lipoprotein(a).
A significant interaction between the vitamin-mineral intervention supplement and BMI (P for interaction = 0.02) and apo B (P for interaction = 0.01) was observed only for SF. In models stratified on the supplement intake versus no supplement intake, an inverse association was observed between BMI and SF only in the supplemented group (ORQ5 versus Q1: 0.80; 95% CI: 0.57, 1.12; P for trend = 0.03), whereas there was no effect in the placebo group (ORQ5 versus Q1: 1.16; 95% CI: 0.82, 1.63; P for trend = 0.54). In contrast, the inverse association between apo B and SF was only observed in the placebo group (ORQ5 versus Q1: 0.52; 95% CI: 0.34, 0.81; P for trend = 0.003), whereas there was no association in the intervention group (ORQ5 versus Q1: 1.11; 95% CI: 0.75, 1.64; P for trend = 0.93).
Effects of dietary fiber from different food sources on anthropometric data, blood pressure, and biological markers in models adjusted on sex, age, alcohol-free energy intake, saturated fatty acids, carbohydrates, tobacco, alcohol, leisure-time exercise, and intervention supplement intake are described in Tables 5 and 6. Fiber intakes from vegetables or legumes had no effect on BMI or the waist-to-hip ratio. Cereal fiber intake decreased the overweight risk but had no effect on waist-to-hip ratio (Table 5). High blood pressure risk was decreased with high intakes of fiber from cereal products and vegetables. Fiber from cereal products, vegetable, and pulses had no significant effect on any biological marker of cardiovascular disease, except for a reduced risk of high homocysteine with high fiber intakes from cereal products or vegetables.
View this table:
TABLE 5. Odds ratios (and 95% CIs) for the relative risk of abnormal markers of cardiovascular disease for the fifth versus the first quintile of dietary fiber intake (g/d) from cereal products, vegetables, or legumes1
View this table:
TABLE 6. Odds ratios (and 95% CIs) for the relative risk of abnormal markers of cardiovascular disease for the fifth versus the first quintile of dietary fiber intake (g/d) from fruit (fresh, canned, or juice), dried fruit, or nuts and seeds1
Fiber from fruit decreased the risk of a high waist-to-hip ratio but had no effect on BMI. We observed a marked reduction in risk associated with both a high BMI and a high waist-to-hip ratio with increasing fiber intake from dried fruit or from nuts and seeds. High blood pressure risk was decreased with high intakes of fiber from fresh and dried fruit. Fiber from fruit was not specifically associated with abnormal biological markers of cardiovascular disease risk. In contrast, fiber from dried fruit was inversely associated with high apo B, triacylglycerol, and fasting glucose concentrations (P for trend = 0.02, 0.01, and < 0.0001, respectively) and positively associated with high lipoprotein(a) (P for trend = 0.02). Fiber from nuts and seeds was associated with a reduced risk of elevated fasting apo B and glucose concentrations (P for trend = 0.03 and 0.002, respectively).
DISCUSSION
In this large cohort of French middle-aged men and women, we found consistent evidence of an inverse association between dietary fiber intake and several risk factors for cardiovascular disease, despite a moderate intake in line with preliminary data (19). Most of the total dietary fiber was NSF, whereas SF accounted for only a minor amount. These values are somewhat different from some European values (5, 25) but are in line with others in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort (26) and in the United States (3, 4, 27). The prevalence of abnormal risk markers was generally higher in men than in women and varied depending on the item concerned. Our data indicate that an important part of the French middle-aged population, especially men, have an elevated risk of cardiovascular disease, as in other industrialized countries.
Overweight and abdominal adiposity are known risk factors for cardiovascular disease and type 2 diabetes. Total dietary fiber intake was clearly inversely associated with elevated BMI and waist-to-hip ratio in the total cohort and by sex, which strongly suggested that a high-fiber diet helps to maintain a lower body fat mass. Two other epidemiologic studies in the United States have already provided evidence that body weight (4, 27) and waist-to-hip ratio (4) are inversely associated with total dietary fiber intake, whereas another study found no such relation (28). NSF intake had the same inverse association with these 2 variables, whereas SF intake was associated with a lower risk of high waist-to-hip ratio only. Finally, higher dietary fiber intakes from fruit, dried fruit, or nuts and seeds were associated with lower BMI and waist-to-hip ratio, whereas cereal fiber intake was associated with low BMI only. This finding is in line with recent observations that whole-grain intake was inversely associated with BMI and waist-to-hip ratio (27, 29). The mechanisms underlying such an effect of dietary fiber on body fat stores might be reduced energy intake, increased fat excretion, and reduced chylomicron concentrations after high-fiber meals (30, 31); blunted postprandial increases in insulinemia after high-fiber meals; improved insulin sensitivity; and reduced insulinemia (4).
Elevated blood pressure is another important risk factor for cardiovascular disease. We provide evidence of an inverse association of total dietary fiber intake with hypertension, an effect already reported in one cohort study on young adults in the United States (4). Such an inverse association was also found with NSF, whereas a trend was found with SF. This association of fiber with blood pressure was independent of the observed inverse association of fiber with BMI. Fiber from cereals, vegetables, and fresh or dried fruit independently showed an inverse association with elevated blood pressure. Such an association was not found in a cohort study in the United States (28) but was observed in a meta-analysis of clinical trials (32) and in a study with whole grain (29). As reviewed previously (33), the possible mechanisms involve reduction in abdominal obesity as discussed above, improvement in vascular reactivity as modulated by cereals (34) or plant phytoestrogens, and other possible mechanisms not yet elucidated.
Fasting total cholesterol is a well-known risk factor for cardiovascular disease as well as are high apo A-I (HDL main apoprotein) and apo B (LDL and VLDL main apoprotein) concentrations (35). In this study, total dietary fiber and NSF intakes were found to be inversely associated with hypercholesterolemia, whereas the highest versus the lowest SF intake was associated with a 19% reduction in hypercholesterolemia risk. Moreover, we found an inverse association of total dietary fiber, SF, and NSF intakes with elevated apo B concentrations and of total dietary fiber and NSF intakes with elevated apo B:A-I ratios, a relevant marker of cardiovascular disease risk (35). Another cohort study in young adults (4) reported a negative or positive association of dietary fiber intake with LDL cholesterol or HDL cholesterol, respectively, whereas another study in middle-aged adults reported a negative association of dietary fiber intake with the total/HDL cholesterol ratio (28). Our data are also in line with numerous clinical trials with fiber-enriched diets, as reviewed previously (13, 14). No specific fiber source was found to be inversely associated with elevated cholesterolemia, whereas fiber intakes from dried fruit or seeds and nuts were associated with a reduced risk of high apo B concentrations. Whole-grain intake was also found to be negatively associated with fasting total and LDL-cholesterol concentrations another study (29).
Elevated fasting triacylglycerols is another independent risk factor for cardiovascular disease (36). Total dietary fiber and NSF intakes were inversely associated with fasting triacylglycerols. This relation was independent of the association of fiber with BMI, which suggests some specific mechanisms. This was also observed in a cohort of young men (4) and is supported by data obtained during clinical trials (9). Dried fruit fiber intakes were specifically inversely associated with triacylglycerols. Whole-grain intake was also recently found to be negatively associated with fasting triacylglycerol concentrations in one study (29). As discussed above and reviewed previously (9), the consumption of dietary fiber or foods with a low glycemic index food or both zcan alter several steps in fat digestion and lipoprotein metabolism (30, 31, 37).
Mixed hyperlipidemia was inversely associated with total dietary fiber and NSF intakes only, eg, a 27% and 26% reduction in this syndrome was observed in the highest and the lowest quintiles of fiber intake, respectively.
Elevated fasting blood glucose is a marker of reduced insulin sensitivity, which predisposes to type 2 diabetes and, subsequently, to an increased risk of cardiovascular disease. In this study, only fiber from dried fruit or nuts and seeds was associated with a significantly reduced risk of high glycemia. McKeown et al (29) reported that whole-grain intake was inversely associated with fasting glucose concentrations. A few epidemiologic studies (4, 29) or clinical trials (38) reported an association between higher fiber intakes and improved insulin sensitivity.
Finally, homocysteine is an acknowledged risk factor for cardiovascular disease, the blood concentration of which is noticeably influenced by dietary habits (39). In a previous study (40), some of the authors of the present study reported that total dietary fiber, SF, and NSF intakes are inversely associated with fasting hyperhomocysteinemia; the relative risk was 0.5 for the highest versus the lowest fiber intakes. This effect was specifically associated with cereal and vegetable fiber intakes. The association might be partly explained by the high vitamin content of plant foods, especially folate—a known down-regulator of homocysteine concentrations. However, the effect remained after adjustment for intakes of folate, coffee, and vitamins B-6 and B12, which suggests a role for other mechanisms (9). An intervention trial in patients with coronary artery disease reported that a whole-grain diet decreased homocysteinemia by 28% and increased plasma antioxidant concentrations (41).
To our knowledge, this report is the first to show that nut and seed fiber intakes, even when low, could have beneficial effects on body weight, abdominal adiposity, apo B concentrations, and glycemia. Further studies are needed to evaluate these possible specific effects.
In conclusion, we obtained evidence of a consistent inverse association between dietary fiber intake and several risk factors for cardiovascular disease. These data thus provide a sound metabolic basis for an association between dietary fiber intake and reduced cardiovascular disease, as already shown in several epidemiologic studies (2, 3, 5-8, 25, 28, 42, 43). The ORs obtained herein for risk of cardiovascular disease, which ranged from 0.5 to 0.25, are comparable with those observed for cardiovascular events or deaths in the above-cited studies. Our observations agree well with the reported association between cardiovascular disease and a prudent (44), Mediterranean (45), or healthy pattern (46), which are characterized by a higher intake of plant foods containing dietary fiber, vitamins, antioxidants, and other potentially active phytochemicals.
An increase in the consumption of dietary fiber-rich foods, of up to 25 g total dietary fiber/d, is widely recommended to prevent pathologic conditions associated with low intakes (24, 47). Scientific evidence is needed to confirm that this intake is optimal. Our data show that total dietary fiber intakes 25 g provide a significantly reduced risk of elevated BMI, waist-to-hip ratio, plasma cholesterol, and hypertension—4 key cardiovascular disease risk factors. An increase in the total dietary fiber intake of 5 g/d was found to be associated with a further beneficial effect on these risk factors. These findings suggest that 25 g total dietary fiber/d is the minimum intake required to achieve a significant protective effect against cardiovascular disease; however, total dietary fiber intakes of 30-35 g/d will provide an even greater protective effect.
ACKNOWLEDGMENTS
We thank Grand Moulins de Paris for helping to create a table of the dietary fiber content of foods.
DL and M-CB-R coordinated the study and wrote the manuscript. SH managed the overall SU.VI.MAX cohort study. NA, SB, and EC managed the SU.VI.MAX database and performed the statistical analyses. RP was involved in risk factor evaluation and corrected the manuscript. All authors contributed to the final report. None of the authors declared a conflict of interest.
REFERENCES