点击显示 收起
1 From the Section of Preventive Medicine & Epidemiology, Evans Department of Medicine, Boston University School of Medicine (LD, LLM, and CE); the Division of Epidemiology, University of Minnesota, Minneapolis (DKA); the Department of Psychiatry, University of Utah, Salt Lake City (HC); and the Division of Biostatistics, Washington University, St Louis (MAP).
2 Supported in part by the National Heart, Lung, and Blood Institute (cooperative agreement grants U01 HL56563, U01 HL56564, U01 HL56565, U01 HL56566, U01 HL56567, U01 HL56568, and U01 HL56569). 3 Reprints not available. Address correspondence to L Djoussé, Boston University School of Medicine, Room B-612, 715 Albany Street, Boston, MA 02118-2526. E-mail: ldjousse{at}bu.edu.
ABSTRACT
Background: An elevated LDL-cholesterol concentration is associated with an increased risk of cardiovascular disease. The association between fruit and vegetable consumption and LDL has been inconsistent.
Objective: The objective was to determine whether a high intake of fruit and vegetables is inversely associated with LDL concentrations.
Design: We used data collected from 4466 subjects in the National Heart, Lung, and Blood Institute Family Heart Study to study the association between fruit and vegetable consumption and serum LDL. We used a food-frequency questionnaire to assess fruit and vegetable intakes and regression models to estimate adjusted mean LDL according to fruit and vegetable consumption.
Results: The mean (±SD) age of the men (n = 2047) was 51.5 ± 14.0 y and that of the women (n = 2419) was 52.2 ± 13.7 y. The average daily serving of fruit and vegetables was 3.2 ± 1.7 for men and was 3.5 ± 1.8 for women. Fruit and vegetable consumption was inversely related to LDL: in the categories 0–1.9, 2.0–2.9, 3.0–3.9, and =" BORDER="0">4 servings/d, multivariate-adjusted mean (95% CI) LDL concentrations were 3.36 (3.28, 3.44), 3.35 (3.27, 3.43), 3.26 (3.17, 3.35), and 3.17 (3.09, 3.25) mmol/L, respectively, for men (P for trend < 0.0001) and 3.35 (3.26, 3.44), 3.22 (3.14, 3.30), 3.21 (3.13, 3.29), and 3.11 (3.04, 3.18), respectively, for women (P for trend < 0.0001). This association was observed across categories of age, education, smoking status, physical activity, and tertiles of Keys score. Exclusion of subjects with prevalent diabetes mellitus or coronary artery disease did not alter these results significantly.
Conclusion: Consumption of fruit and vegetables is inversely related to LDL in men and women.
Key Words: LDL • fruit • vegetables • lipids • National Heart • Lung • and Blood Institute Family Heart Study
INTRODUCTION
An elevated LDL-cholesterol concentration is a risk factor for coronary artery disease (1–5). Plasma concentrations of LDL are influenced by both genetic and environmental factors. Although it is difficult to alter genetic factors, modifiable environmental factors such as smoking or dietary patterns could be targeted in preventive interventions aimed at lowering LDL. The usual guidelines recommend a reduction in dietary saturated fat and cholesterol intakes as a way to prevent hypercholesterolemia; however, only limited data are available on the benefits of fruit and vegetable consumption on plasma concentrations of LDL in a community-based population. Data on the effects of fruit and vegetable intakes on LDL are inconsistent. In the Dietary Approaches to Stop Hypertension (DASH) trial (6), a diet high in fruit and vegetables was not associated with a significant reduction in LDL compared with the control diet, although the trend suggests a decrease in plasma LDL concentrations. In contrast, in the Indian Diet Heart Study (7), fruit and vegetable consumption decreased LDL concentrations by 7%. In a randomized trial, fruit and vegetable intake was associated with a reduction in LDL among patients with acute myocardial infarction after 12 wk of intervention (8). Fruit and vegetables are rich in dietary fiber, which has been shown to decrease LDL concentrations (9–11).
In the current study we used data from 4466 adult participants of the National Heart, Lung, and Blood Institute (NHLBI) Family Heart Study to evaluate whether higher intakes of fruit and vegetables is inversely related to LDL concentrations in men and women, independent of other risk factors.
SUBJECTS AND METHODS
Study population
The NHLBI Family Heart Study is a multicenter, population-based study designed to identify and evaluate genetic and nongenetic determinants of cardiovascular disease. A detailed description of the methods and design was published previously (12). Subjects in this study are members of families from previously established population-based cohort studies: The Framingham Heart Study in Framingham, MA; the Atherosclerosis Risk in Communities Study cohorts in North Carolina and Minnesota; and the Utah Health Family Tree Study in Salt Lake City. Briefly, in 1993–1995, participants (probands) in each of the 4 studies were selected at random and invited to furnish an updated family health history that contained information on their parents, children, and siblings. Of the families who furnished data, 588 were chosen at random and 657 were chosen because of higher than expected coronary artery disease rates among their family members. All members of these families, including the spouses of the probands, were invited to 1 of the 4 study clinics for evaluation. The evaluation included a detailed medical and lifestyle history that was obtained through an interview. Informed consent was obtained from each participant, and the study protocol was reviewed and approved by the institutional review boards of each of the participating institutions.
Blood collection and assays
Fasting blood samples for lipid measurements were collected into evacuated tubes containing no additives. Blood samples were then spun at 3000 x g for 10 min at 4 °C. Sera were stored at -70 °C until shipped periodically to a central laboratory at the Fairview–University Medical Center in Minneapolis for processing. LDL was estimated by using the method of Friedewald et al (13), except for subjects with triacylglycerol concentrations >400 mg/dL, whose LDL concentrations were measured by ultracentrifugation.
Assessment of fruit and vegetable intakes
Trained interviewers obtained information on usual consumption of fruit and vegetables with the use of a 100-item food-frequency questionnaire (FFQ) modified from the FFQ developed by Willett et al (14, 15). The reproducibility and validity of the FFQ was documented elsewhere (16). For fruit consumption, each participant was asked about how often, on average, he or she consumed fruit during the previous year (items 9 through 14 on the FFQ). For vegetable consumption, subjects were asked about their average consumption of vegetables such as beans, broccoli, cabbage, cauliflower, carrots, corn, spinach, squash, and tomatoes (items 15 through 25 on the FFQ). Response categories ranged from almost never, 1–3/mo, 1/wk, 2–4/wk, 5–6/wk, 1/d, 2–3/d, 4–6/d, to >6/d. The portion size of each fruit and vegetable was specified to facilitate determination of the number of typical servings and nutrient content.
Other variables
Information on the dietary consumption of total fat, saturated fat, and cholesterol was obtained by using the same FFQ. Intake of specific nutrients was computed by multiplying the frequency of consumption of an item by the nutrient content of specified portions. Nutrient composition was obtained from the Harvard University FOOD COMPOSITION DATABASE, derived from US Department of Agriculture sources (17), and from manufacturer information.
Anthropometric data were collected while the subjects were wearing scrub suits. A balance scale was used to measure body weight, and height was measured with a wall-mounted vertical ruler. Demographic data and information on alcohol intake, cigarette smoking, and physical activity (minutes per day of leisure activity) were obtained by interview.
Statistical analysis
Because eating habits differ between men and women, we initially analyzed the data stratified by sex and created the following categories of fruit and vegetable consumption: 0–1.9, 2.0–2.9, 3.0–3.9, and =" BORDER="0">4 servings/d. Because the subjects were not independent in this study, we used a general linear model (PROC MIXED; SAS Institute, Cary, NC) to estimate adjusted mean LDL concentrations across categories of fruit and vegetable intake. This method corrects the variance of the point estimates for familial clustering. Adjustment was made for age, age squared, study center, risk group (high risk compared with random), body mass index, energy intake (quintiles), smoking status (never, former, and current smokers), dietary cholesterol (quintiles), history of coronary artery disease, and diabetes mellitus. Additional adjustment for education (high school graduate or less, vocational school, and college or more), physical activity (minutes per day of leisure activity), use of vitamin supplements, consumption of breakfast cereals, and total fat (quintiles), saturated fat (quintiles), and polyunsaturated fat (quintiles) intakes did not alter the results significantly. All sex-specific analyses were initially conducted within the random and high-risk groups separately, but, because the results were similar, we combined the high-risk and random groups.
Subjects with higher intakes of fruit and vegetables were older and had different lifestyle habits and lower Keys score than did those who consumed fewer servings of fruit and vegetables per day. The Keys score correlates changes in fatty acid intake with changes in serum cholesterol and is computed as follows:
RESULTS
Of the 4466 subjects, 45.8% were men. The average age of the men was 51.5 ± 14.0 y (range: 25.2–91.0 y) and of the women was 52.2 ± 13.7 y (range: 25.2–93.6 y). The mean consumption of fruit and vegetables was 3.2 ± 1.7 servings/d for men (range: 0–15.2 servings/d) and 3.5 ± 1.8 servings/d for women (range: 0–14.1 servings/d). The baseline characteristics of the male and female participants according to their consumption of fruit and vegetables are shown in Tables 1 and 2, respectively. For both men and women, a higher intake of fruit and vegetables was associated with older age (P < 0.0001), higher energy intake (P < 0.0001), a lower percentage of energy from total fat (P < 0.0001), lower Keys score (P < 0.0001), a higher intake of dairy products (P < 0.0005), use of vitamin supplements (P < 0.0001), more physical activity (P < 0.0001), a lower prevalence of smoking (P = 0.001), and a higher prevalence of coronary artery disease (P < 0.05) and diabetes mellitus (P < 0.0002). In men, a higher consumption of fruit and vegetables was also associated with a lower percentage of energy from saturated fat (P = 0.024) and higher educational attainment (P = 0.003). In women, a higher intake of fruit and vegetables was also associated with a higher intake of dietary cholesterol (P < 0.0001).
View this table:
TABLE 1. . Baseline characteristics of male participants in the National Heart, Lung, and Blood Institute Family Heart Study according to fruit and vegetable intakes
View this table:
TABLE 2. . Baseline characteristics of female participants in the National Heart, Lung, and Blood Institute Family Heart Study according to fruit and vegetable intakes
In the categories 0–1.9, 2.0–2.9, 3.0–3.9, and =" BORDER="0">4 servings of fruit and vegetables per day, adjusted mean (±SE) LDL concentrations were 3.36 ± 0.04, 3.35 ± 0.04, 3.26 ± 0.04, and 3.17 ± 0.06 mmol/L, respectively, in men (P for trend < 0.0001) and 3.35 ± 0.05, 3.22 ± 0.04, 3.21 ± 0.04, and 3.11 ± 0.04 mmol/L, respectively, in women (P for trend < 0.0001) in a model that adjusted for age, age squared, field center, risk group, body mass index, energy intake, smoking status, dietary cholesterol, and prevalence of diabetes mellitus and coronary artery disease (Table 3). Additional adjustment for education, physical activity, and saturated, polyunsaturated, and total fat intakes did not alter the results significantly. Furthermore, exclusion of 751 subjects with diabetes mellitus, coronary artery disease, or both did not change the results (data not shown). We also looked at HDL, LDL:HDL, and triacylglycerol as endpoints. Although the consumption of fruit and vegetables was not associated with HDL (P for trend = 0.57 for men and 0.97 for women) or triacylglycerol (P for trend = 0.83 for men and 0.60 for women) concentrations, fruit and vegetables consumption was inversely related to LDL:HDL. From the lowest to the highest category of fruit and vegetable consumption, multivariate-adjusted mean LDL:HDL values were 3.21 ± 0.05, 3.19 ± 0.05, 3.16 ± 0.05, and 3.03 ± 0.05, respectively, for men (P for trend = 0.006). Corresponding values for women were 2.52 ± 0.05, 2.40 ± 0.04, 2.42 ± 0.04, and 2.36 ± 0.04 (P for trend = 0.020). Because the intake of fruit and vegetables was inversely related to the intake of energy from saturated fat, we conducted subanalyses that were restricted to subjects above the 75th percentile of energy intake from saturated fat; we found an inverse relation between fruit and vegetable consumption and LDL cholesterol concentration (P for trend < 0.0001; data not shown).
View this table:
TABLE 3. . Adjusted LDL cholesterol according to fruit and vegetable intakes in the National Heart, Lung, and Blood Institute Family Heart Study1
To assess residual confounding by dietary fat intake, we stratified by tertiles of Keys score. From the lowest to the highest category of fruit and vegetable consumption, multivariate-adjusted mean (±SE) LDL-cholesterol concentrations were 3.43 ± 0.07, 3.37 ± 0.06, 3.26 ± 0.06, and 3.19 ± 0.05 mmol/L, respectively, for the lowest tertile of Keys score (P for trend = 0.0004); 3.31 ± 0.05, 3.28 ± 0.05, 3.24 ± 0.05, and 3.11 ± 0.05 mmol/L, respectively, for the second tertile of Keys score (P for trend = 0.002); and 3.38 ± 0.05, 3.24 ± 0.05, 3.23 ± 0.06, and 3.14 ± 0.06 mmol/L, respectively, for the third tertile of Keys score (P for trend = 0.0008). In addition, when stratified by age, education, smoking status, physical activity, and use of vitamin supplements, the relation between fruit and vegetable intakes and LDL-cholesterol concentrations persisted (data not shown).
DISCUSSION
In this cross-sectional study, we found that consumption of fruit and vegetables was inversely related to LDL-cholesterol concentrations in men and women, independent of age, Keys score, smoking status, exercise, educational attainment, and use of vitamin supplements. Subjects in the highest fruit and vegetable intake groups had LDL concentrations that were
Our study has some limitations. Given the cross-sectional design, we cannot infer causality between fruit and vegetable consumption and lower concentrations of LDL cholesterol. However, dietary patterns in adults are relatively stable, and the exclusion of those with prevalent coronary artery disease and diabetes mellitus did not alter the results significantly, making it more likely that dietary intakes reported over the past year may reflect the subjects' usual diets. Frequencies of fruit and vegetable consumption were self-reported; thus, reporting bias might have affected our estimates of the effect.
Could the observed inverse association between fruit and vegetable consumption and LDL-cholesterol concentrations simply be attributed to a lower intake of saturated fat and dietary cholesterol in subjects with higher fruit and vegetable intakes? Our data indicate that a higher intake of fruit and vegetables was associated with lower Keys score and a lower percentage of energy from saturated fat, especially in men. This suggests that residual confounding by saturated fat could have biased our estimates. However, dietary cholesterol did not differ across categories of fruit and vegetable intakes in men and was in the opposite direction in women, because a higher intake of fruit and vegetables was related to higher dietary cholesterol in women (Table 2). In addition, the inverse association between fruit and vegetable consumption and LDL-cholesterol concentrations was observed across all tertiles of Keys score and among subjects whose energy intake from saturated fat was above the 75th percentile of the total population. This finding is not consistent with the fact that subjects with a higher intake of fruit and vegetables were more likely to eat less saturated fat and dietary cholesterol. Thus, our findings are less likely to be attributable to the effect of substituting saturated fat and dietary cholesterol with fruit and vegetables. The wide age range, the availability of data across many centers, and the large sample size are strengths of the study. In conclusion, our data show that the consumption of fruit and vegetables is associated with lower concentrations of LDL cholesterol in a dose-response manner.
ACKNOWLEDGMENTS
This paper is presented on behalf of the investigators of the NHLBI Family Heart Study. The participating institutions and principal staff of the study are as follows. Forsyth County, University of North Carolina, and Wake Forest University Field Center: Gerardo Heiss, Stephen Rich, Greg Evans, James Pankow, HA Tyroler, Jeannette T Bensen, Catherine Paton, Delilah Posey, and Amy Haire; University of Minnesota Field Center: Donna K Arnett, Aaron R Folsom, Larry Atwood, James Peacock, and Greg Feitl; Boston University, Framingham Field Center: R Curtis Ellison, Richard H Myers, Yuqing Zhang, Andrew G Bostom, Luc Djoussé, Jemma B Wilk, and Greta Lee Splansky; University of Utah Field Center: Steven C Hunt, Roger R Williams (deceased), Paul N Hopkins, Hilary Coon, and Jan Skuppin; Coordinating Center, Washington University, St Louis: Michael A Province, DC Rao, Ingrid B Borecki, Yuling Hong, Mary Feitosa, Jeanne Cashman, and Avril Adelman; Central Biochemistry Laboratory, University of Minnesota, Minneapolis: John H Eckfeldt, Catherine Leiendecker-Foster, Michael Y Tsai, and Greg Rynders; Central Molecular Laboratory, University of Utah, Salt Lake City: Mark F Leppert, Jean-Marc Lalouel, Tena Varvil, and Lisa Baird; NHLBI Project Office: Phyliss Sholinsky, Millicent Higgins (retired), Jacob Keller (retired), Sarah Knox, and Lorraine Silsbee.
LD designed the project, completed the data analyses, and drafted the manuscript. DKA and MAP participated in the study design, the data collection, and the critical review of the manuscript. HC and LLM participated in the data analyses and the critical review of the manuscript. RCE participated in the study design, the data collection, the data analyses, and the critical review of the manuscript. None of the authors had a conflict of interest to disclose.
REFERENCES