Abstract
Background: Fruit, vegetable, and dietary fiber intake have been associated with lower risk of cardiovascular disease (CVD); however, little is known about their role in obesity prevention.
Objective: Our goal was to investigate whether intake of fruits, vegetables, and dietary fiber is associated with weight change and the risk of becoming overweight and obese.
Methods: We studied 18,146 women aged ≥45 y from the Women’s Health Study free of CVD and cancer with an initial body mass index (BMI) of 18.5 to <25 kg/m2. Fruit, vegetable, and dietary fiber intakes were assessed at baseline through a 131-item food-frequency questionnaire, along with obesity-related risk factors. Women self-reported body weight on annual questionnaires.
Results: During a mean follow-up of 15.9 y, 8125 women became overweight or obese (BMI ≥25 kg/m2). Intakes of total fruits and vegetables, fruits, and dietary fiber were not associated with the longitudinal changes in body weight, whereas higher vegetable intake was associated with greater weight gain (P-trend: 0.02). In multivariable analyses, controlling for total energy intake and physical activity along with other lifestyle, clinical, and dietary factors, women in the highest vs. lowest quintile of fruit intake had an HR of 0.87 (95% CI: 0.80, 0.94; P-trend: 0.01) of becoming overweight or obese. No association was observed for vegetable or dietary fiber intake. The association between fruit intake and risk of becoming overweight or obese was modified by baseline BMI (P-interaction: <0.0001) where the strongest inverse association was observed among women with a BMI <23 kg/m2 (HR: 0.82; 95% CI: 0.71, 0.94).
Conclusion: Our results suggest that greater baseline intake of fruit, but not vegetables or fiber, by middle-aged and older women with a normal BMI at baseline is associated with lower risk of becoming overweight or obese.
Keywords: vegetables, fruits, dietary fiber, overweight, obesity, epidemiology
Introduction
The high prevalence of overweight and obesity is a serious public health challenge that remains one of the most common causes of morbidity and mortality (1). Data from the National Health and Nutrition Examination Survey showed that more than one-third (34.9%) of the US population were obese in 2011–2012 (2). Diet and physical activity are considered key measures to prevent weight gain and development of obesity (3). Fruit and vegetable consumption has been associated with a lower risk of cardiovascular disease (CVD) (4, 5) and cancer (6, 7); however, less is known about its role in obesity prevention. Fruits and vegetables contain dietary fiber and a wide range of vitamins, minerals, and phytochemicals that may counteract weight gain over time (8). Moreover, fruit and vegetable consumption contributes to satiety and replaces more caloric-dense foods to hinder weight gain (9).
Cross-sectional studies investigating the association between fruit and vegetable consumption and body weight have reported inconsistent results (10–15). Only a few prospective cohort studies have investigated whether increases in fruit and vegetable consumption reduce weight gain over time (16–19). The majority of these studies have reported inverse associations between high fruit and vegetable intake and weight gain (16–18). Two small-scale randomized trials investigated the effect of fruit and vegetable intake on weight loss as the primary outcome of interest and reported nonsignificant decreases in body weight at the end of the trials (20, 21). Secondary analyses of randomized trials have reported inverse associations between a diet characterized by increased consumption of fruits and vegetables and weight loss (22–25), but no trial to our knowledge has been specifically designed to investigate the effect of fruit and vegetable intake on weight change or development of obesity.
We therefore aimed to examine the prospective association of fruit, vegetable, and dietary fiber intake with weight gain and risk of becoming overweight or obese in a cohort of middle-aged and elderly women who were initially normal weight.
Methods
Starting in 1992, 39,876 women randomly assigned into the Women’s Health Study, a 2 × 2 × 2 factorial trial of low-dose aspirin, vitamin E, and β-carotene in the primary prevention of CVD and cancer among female US health professionals, completed baseline questionnaires asking about their medical history and lifestyle factors. Women were aged ≥45 y, postmenopausal or not intending to become pregnant, and had no history of myocardial infarction, stroke, transient ischemic attack, or cancer (except nonmelanoma skin cancer). Women were also asked to complete a 131-item validated semiquantitative FFQ and 39,310 (98.6%) women responded (26, 27).
At baseline, women self-reported weight (in pounds) and height (in inches). For this analysis, we excluded women with a baseline BMI (calculated as body weight in kilograms divided by the square of height in meters) outside the range of 18.5–25 kg/m2 (n = 20,106). We also excluded women who reported a history of diabetes (n = 1143) or CVD (n = 13) at baseline, had missing information on fruit and vegetable intake, responded to an insufficient number of food items, or had a total energy intake outside the range of 600–3500 kcal/d (n = 468). Thus, 18,146 women were included in the current study and were followed through 28 February 2012. During follow-up, body weight was updated on annual questionnaires at years 2, 3, 5, 6, and 9. When the trial ended, ∼85% of the initially normal-weight women agreed to continue observational follow-up, and body weight was additionally updated at years 11, 12, 13, 14, 15, 16, and 17. Written informed consent was obtained from all participants, and this research was approved by the institutional review board of Brigham and Women’s Hospital, Boston, MA.
Intakes of fruits, vegetables, and dietary fiber.
On the FFQ, women answered questions regarding how often they consumed a specific food item, on average, during the previous year. There were 9 possible response categories ranging from “never or less than once per month” to “6+ per day.” The FFQ included 16 fruit and 28 vegetable items. We calculated the mean daily intake (servings/d) of total fruits and vegetables by summing the frequency of consuming individual food items. We also calculated the sum of subgroups of specific types of fruits, including citrus fruits (grapefruit, grapefruit juice, orange, and orange juice), berries (blueberries and strawberries), and fruit juice (grapefruit juice, orange juice, apple juice, and other juice), as well as vegetables, including green leafy vegetables (spinach, kale, and lettuce), cruciferous vegetables (broccoli, cabbage, cauliflower, and Brussels sprouts), dark yellow vegetables (carrots, yellow squash, yams, and sweet potatoes), legumes (beans, peas, string beans, and tofu), and tomatoes (tomatoes, tomato sauce, and tomato juice). Dietary fiber intake was calculated by multiplying the frequency of consumption of each food item by the fiber content of each specific portion derived from the USDA (28) and from manufacturers using the Association of Official Analytical Chemists’ method (29–31). Dietary fiber intake was energy-adjusted with use of the residual method (32). Assessment of fruit, vegetable, and dietary fiber intake using FFQs has been validated by comparing four 1-wk diet records in similar health professional cohorts such as the Nurses’ Health Study (26) and the Health Professionals’ Follow-up Study (33).
Development of overweight or obesity during follow-up.
We categorized women as normal weight (BMI: 18.5 to <25 kg/m2), overweight (BMI: 25 to <30 kg/m2), or obese (BMI: ≥30 kg/m2) at each follow-up time (34). We defined women who became overweight or obese as those who had a normal BMI at baseline and then newly reported an overweight or obese BMI during follow-up. Women were followed from baseline to the date of the questionnaire on which they had become overweight or obese, their date of death, or end of follow-up, whichever occurred first. We also censored participants on the date of diabetes during follow-up to avoid possible influence on weight control because of a new diabetes diagnosis. The high validity of self-reported body weight has been demonstrated in the Nurses’ Health Study, with a Pearson correlation coefficient of 0.97 between self-reported and standardized measurements of weight taken by technicians (27).
Statistical analyses.
All statistical analyses were performed with SAS version 9.2 (SAS Institute). We categorized women into quintiles of fruit, vegetable, and dietary fiber intake. Age-standardized mean values and SDs for continuous variables and percentages for categorical variables were calculated and compared across quintiles. Linear trends were assessed with use of ANOVA for continuous and categorical variables.
We calculated body weight change between baseline and each follow-up time. Only women with reported values on weight were included in the analysis at the different time points. The PROC MIXED models with an unstructured covariance matrix for repeated measures was used to investigate the association between intakes of fruits and vegetables with longitudinal changes in body weight as a continuous variable. We used Cox proportional hazards models to calculate HRs with 95% CIs (35). In our multivariable models, we adjusted for different lifestyle, clinical, and dietary factors including baseline age (years, continuous), smoking status (never, past, current <15 cigarettes/d, or ≥15 cigarettes/d), physical activity [multiple of resting metabolic rate (MET) hours based on absolute intensity of the activity], postmenopausal status (no, yes, biologically uncertain, or unclear), hormone replacement therapy use (never, past, or current), history of hypertension (yes and no), history of hypercholesterolemia (yes and no), alcohol intake [rarely/never, 1–3 drinks/mo (1 drink has been estimated to contain 13.2-g ethanol), 1–6 drinks/wk, or ≥1 drink/wk], and caloric intake (kcal/d). We performed a separate multivariable-adjusted model by adding baseline BMI (kg/m2, continuous) because adjusting for baseline status may induce biased statistical association in analysis of change (36). Linear trends across quintiles were tested with use of the median value of each quintile as a continuous variable. We further examined the shape of association between types of fruits and vegetables and risk of becoming overweight or obese with use of a restricted cubic spline with 3 knots (at the 10th, 50th, and 90th percentile) (37).
To investigate effect modification by important lifestyle factors related to fruit and vegetable intake and risk of becoming overweight or obese, we performed stratified analyses by age, baseline BMI, smoking status, and physical activity. We tested for interactions using Wald tests.
The proportional hazards assumption was tested by entering the product of baseline intakes of fruits, vegetables, and dietary fiber and the natural logarithm of time in the model. Because there was evidence of violation of this assumption we performed sensitivity analyses on the risk of becoming overweight or obese by ending the follow-up after 8 y. Two-sided P values < 0.05 were considered statistically significant.
Results
During a mean follow-up of 15.9 y, 8125 women became overweight or obese. In Table 1, we present baseline characteristics of the women. Those in the highest quintile of fruit intake compared to the lowest were older, had lower BMI, more physically active, more likely to use hormone replacement therapy, less likely to have hypercholesterolemia, more likely to consume alcohol, and more likely to ever use multivitamins regularly; they also had higher intake of dietary fiber and total energy intake. A similar pattern of associations with baseline characteristics was observed for quintiles of vegetable intake.
TABLE 1.
Baseline characteristics according to fruit and vegetable intake in 18,146 middle-aged and older women with normal weight1
| Total fruit intake (servings/d) |
Total vegetable intake (servings/d) |
|||||||
| Characteristics | Q1: <1.0 | Q3: 1.7 to <2.3 | Q5: ≥3.1 | P-trend | Q1: <2.0 | Q3: 3.0 to <3.9 | Q5: ≥5.4 | P-trend |
| n | 3604 | 3654 | 3567 | 3605 | 3682 | 3548 | ||
| Lifestyle | ||||||||
| Age, y | 52.0 ± 6.2 | 53.9 ± 6.9 | 55.5 ± 7.7 | <0.0001 | 52.9 ± 6.8 | 54.0 ± 7.1 | 54.6 ± 7.4 | <0.0001 |
| BMI, kg/m2 | 22.4 ± 1.6 | 22.5 ± 1.7 | 22.3 ± 1.6 | 0.0007 | 22.4 ± 1.7 | 22.4 ± 1.6 | 22.4 ± 1.6 | 0.77 |
| Current smokers | 27 | 11 | 7 | <0.0001 | 19 | 13 | 10 | <0.0001 |
| Physical activity, total MET hours | 11.8 ± 16.5 | 16.5 ± 17.9 | 23.2 ± 24.5 | <0.0001 | 12.3 ± 16.1 | 16.3 ± 17.2 | 23.2 ± 24.3 | <0.0001 |
| Hormone therapy | 42 | 46 | 45 | 0.004 | 43 | 45 | 47 | 0.02 |
| Postmenopausal | 53 | 54 | 54 | 0.77 | 54 | 52 | 55 | 0.06 |
| Hypercholesterolemia | 26 | 24 | 23 | 0.01 | 25 | 25 | 24 | 0.50 |
| Hypertension | 17 | 16 | 16 | 0.65 | 16 | 16 | 16 | 0.48 |
| ≥1 Alcoholic drink/mo2 | 58 | 64 | 62 | <0.0001 | 55 | 64 | 65 | <0.0001 |
| Multivitamin use | 26 | 31 | 33 | <0.0001 | 31 | 30 | 32 | 0.74 |
| Dietary | ||||||||
| Total fiber, g/d | 12.5 ± 5.2 | 18.5 ± 6.0 | 27.1 ± 9.0 | <0.0001 | 12.0 ± 4.5 | 18.2 ± 5.5 | 28.3 ± 8.7 | <0.0001 |
| Energy intake, kcal/d | 1397 ± 463 | 1691 ± 436 | 2062 ± 497 | <0.0001 | 1365 ± 424 | 1683 ± 446 | 2064 ± 522 | <0.0001 |
Values are means ± SDs or percentages and are standardized to the age distribution of the study population. MET, resting metabolic rate; Q, quintile.
Ethanol content has been estimated to be 13.2 g for 360 mL (12 oz) of beer, 10.8 g for 120 mL (4 oz) of red or white wine, and 15.1 g for 45 mL of liquor.
In Table 2, the longitudinal changes in body weight during a 17-y period are presented according to quintiles of consumption of total fruits and vegetables, fruits, vegetables, and dietary fiber. There were no statistically significant associations observed for intakes of total fruits and vegetables, fruits, and dietary fiber, whereas vegetable intake was associated with greater weight gain (P-trend: 0.02). When adding red meat to the multivariable-adjusted model the results were similar. We further investigated body weight changes during a 5-y period (Supplemental Table 1) in which there were no statistically significant trends observed for weight gain across quintiles of total fruits and vegetables, fruits, vegetables, and dietary fiber.
TABLE 2.
Mean change (95% CI) in body weight between baseline and last follow-up time point (year 17) according to quintiles of fruit, vegetable, and dietary fiber intake among middle-aged and older women1
| Weight change (kg/17 y) |
||||
| n | Age adjusted | Multivariable adjusted2 | Multivariable adjusted2 + BMI | |
| Fruits and vegetables, servings/d | ||||
| Q1: <3.5 | 3573 | 1.58 (1.47, 1.70) | 1.83 (1.66, 2.01) | 1.82 (1.65, 2.00) |
| Q2: 3.5 to <4.9 | 3680 | 1.43 (1.31, 1.54) | 1.72 (1.55, 1.89) | 1.71 (1.54, 1.88) |
| Q3: 4.9 to <6.3 | 3672 | 1.50 (1.38, 1.61) | 1.81 (1.64, 1.98) | 1.80 (1.63, 1.97) |
| Q4: 6.3 to <8.2 | 3668 | 1.58 (1.47, 1.69) | 1.92 (1.75, 2.09) | 1.90 (1.73, 2.07) |
| Q5: ≥8.2 | 3553 | 1.50 (1.37, 1.61) | 1.87 (1.69, 2.05) | 1.86 (1.67, 2.04) |
| P-trend | 0.81 | 0.25 | 0.25 | |
| Fruits, servings/d | ||||
| Q1: <1.0 | 3604 | 1.53 (1.41, 1.65) | 1.80 (1.63, 1.97) | 1.78 (1.61, 1.95) |
| Q2: 1.0 to <1.7 | 3644 | 1.56 (1.44, 1.67) | 1.87 (1.70, 2.04) | 1.85 (1.68, 2.02) |
| Q3: 1.7 to <2.3 | 3654 | 1.60 (1.48, 1.71) | 1.92 (1.75, 2.09) | 1.91 (1.74, 2.09) |
| Q4: 2.3 to <3.1 | 3677 | 1.46 (1.35, 1.57) | 1.79 (1.61, 1.96) | 1.77 (1.60, 1.95) |
| Q5: ≥3.1 | 3567 | 1.43 (1.32, 1.55) | 1.78 (1.59, 1.96) | 1.77 (1.59, 1.96) |
| P-trend | 0.11 | 0.46 | 0.59 | |
| Vegetables, servings/d | ||||
| Q1: <2.0 | 3605 | 1.51 (1.39, 1.63) | 1.76 (1.58, 1.93) | 1.75 (1.58, 1.92) |
| Q2: 2.0 to <3.0 | 3649 | 1.43 (1.31, 1.54) | 1.71 (1.54, 1.88) | 1.70 (1.53, 1.87) |
| Q3: 3.0 to <3.9 | 3682 | 1.54 (1.43, 1.66) | 1.86 (1.69, 2.03) | 1.84 (1.67, 2.01) |
| Q4: 3.9 to <5.4 | 3662 | 1.55 (1.44, 1.67) | 1.90 (1.73, 2.07) | 1.88 (1.71, 2.05) |
| Q5: ≥5.4 | 3548 | 1.54 (1.43, 1.66) | 1.93 (1.75, 2.11) | 1.91 (1.73, 2.08) |
| P-trend | 0.28 | 0.01 | 0.02 | |
| Dietary fiber, mg/d | ||||
| Q1: <12.4 | 3630 | 1.56 (1.44, 1.68) | 1.86 (1.68, 2.04) | 1.84 (1.66, 2.02) |
| Q2: 12.4 to <16.1 | 3615 | 1.53 (1.41, 1.65) | 1.83 (1.66, 2.01) | 1.82 (1.65, 1.99) |
| Q3: 16.1 to <19.7 | 3642 | 1.55 (1.44, 1.67) | 1.86 (1.69, 2.03) | 1.85 (1.68, 2.02) |
| Q4: 19.7 to <24.7 | 3629 | 1.48 (1.36, 1.59) | 1.80 (1.62, 1.97) | 1.78 (1.61, 1.96) |
| Q5: ≥24.7 | 3630 | 1.46 (1.34, 1.57) | 1.78 (1.59, 1.97) | 1.77 (1.58, 1.96) |
| P-trend | 0.15 | 0.44 | 0.45 | |
Q, quintile.
Adjusted for age, randomization treatment assignment, physical activity, history of hypercholesterolemia or hypertension, smoking status, postmenopausal status, postmenopausal hormone use, alcohol use, multivitamin use, and energy intake.
We computed the mean body weight during the follow-up among women who consumed <5 and ≥5 servings/d of fruits and vegetables. Body weight increased over time in both groups; however, the greatest increase was observed among women consuming <5 servings/d of fruits and vegetables. In mixed modeling analyses, there was a statistically significant interaction between fruit and vegetable intake and follow-up time on the changes in body weight (P < 0.0001).
In age-adjusted analyses, we observed lower risks of becoming overweight or obese for higher intake compared to the lowest intake of total fruits and vegetables, fruits, and dietary fiber (Table 3). No association was observed for vegetable intake. Adjusting for other lifestyle, clinical, and dietary factors slightly attenuated the associations. Women in the highest quintile of fruit intake had a lower risk, whereas women in the highest quintile of vegetable intake had a higher risk of becoming overweight or obese. No significant associations were found for total fruit and vegetable intake and dietary fiber intake. When adding BMI to the model, we observed a statistically significant 9% (95% CI: 1%, 16%; P-trend: 0.19) lower risk for total fruit and vegetable intake and a 13% (95% CI: 6%, 20%; P-trend: 0.01) lower risk for fruit intake. For vegetable intake, a statistically significant inverse association was only found in the second quintile vs. the lowest quintile. The results were similar when mutually adjusting for fruit and vegetable intake in the same model. Moreover, no association was observed for dietary fiber intake. We next added red meat to the multivariable-adjusted model, and the results were attenuated for total fruit and vegetable intake (HR in quintile 5 vs. quintile 1: 0.95; 95% CI: 0.87, 1.04) and fruit intake (HR in quintile 5 vs. quintile 1: 0.91; 95% CI: 0.83, 0.98). The results for vegetable intake and dietary fiber intake were similar.
TABLE 3.
HRs (95% CIs) of becoming overweight or obese according to quintiles of fruit, vegetable, and dietary fiber intake among middle-aged and older women1
| Age adjusted |
Multivariable adjusted |
||||
| n | Cases | HR (95% CI) | HR (95% CI)2 | HR (95% CI)2 + BMI | |
| Fruits and vegetables, servings/d | |||||
| Q1: <3.5 | 3573 | 1726 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 3.5 to <4.9 | 3680 | 1705 | 0.96 (0.90, 1.03) | 0.99 (0.93, 1.06) | 0.88 (0.82, 0.94) |
| Q3: 4.9 to <6.3 | 3672 | 1600 | 0.91 (0.85, 0.97) | 0.95 (0.89, 1.03) | 0.92 (0.86, 0.99) |
| Q4: 6.3 to <8.2 | 3668 | 1577 | 0.91 (0.85, 0.98) | 0.98 (0.91, 1.05) | 0.93 (0.86, 1.00) |
| Q5: ≥8.2 | 3553 | 1517 | 0.91 (0.85, 0.97) | 1.01 (0.93, 1.10) | 0.91 (0.84, 0.99) |
| P-trend | 0.003 | 0.80 | 0.19 | ||
| Fruits, servings/d | |||||
| Q1: <1.0 | 3604 | 1781 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 1.0 to <1.7 | 3644 | 1694 | 0.94 (0.88, 1.01) | 0.97 (0.91, 1.04) | 0.86 (0.80, 0.92) |
| Q3: 1.7 to <2.3 | 3654 | 1648 | 0.92 (0.86, 0.98) | 0.96 (0.89, 1.02) | 0.93 (0.87, 1.00) |
| Q4: 2.3 to <3.1 | 3677 | 1586 | 0.88 (0.82, 0.94) | 0.92 (0.86, 0.99) | 0.92 (0.85, 0.99) |
| Q5: ≥3.1 | 3567 | 1416 | 0.81 (0.75, 0.87) | 0.86 (0.79, 0.93) | 0.87 (0.80, 0.94) |
| P-trend | <0.0001 | 0.0001 | 0.01 | ||
| Vegetables, servings/d | |||||
| Q1: <2.0 | 3605 | 1659 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 2.0 to <3.0 | 3649 | 1617 | 0.95 (0.89, 1.02) | 0.98 (0.91, 1.05) | 0.86 (0.80, 0.92) |
| Q3: 3.0 to <3.9 | 3682 | 1674 | 1.01 (0.94, 1.08) | 1.07 (1.00, 1.15) | 1.05 (0.97, 1.12) |
| Q4: 3.9 to <5.4 | 3662 | 1618 | 0.98 (0.91, 1.05) | 1.06 (0.98, 1.14) | 1.00 (0.93, 1.07) |
| Q5: ≥5.4 | 3548 | 1557 | 0.99 (0.92, 1.06) | 1.11 (1.03, 1.20) | 0.99 (0.91, 1.07) |
| P-trend | 0.94 | 0.003 | 0.22 | ||
| Dietary fiber, mg/d | |||||
| Q1: <12.4 | 3630 | 1727 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 12.4 to <16.1 | 3615 | 1660 | 0.98 (0.91, 1.04) | 1.00 (0.93, 1.07) | 1.18 (1.10, 1.27) |
| Q3: 16.1 to <19.7 | 3642 | 1627 | 0.94 (0.88, 1.00) | 0.97 (0.90, 1.04) | 1.07 (0.99, 1.16) |
| Q4: 19.7 to <24.7 | 3629 | 1587 | 0.93 (0.87, 1.00) | 0.98 (0.90, 1.06) | 1.05 (0.96, 1.14) |
| Q5: ≥24.7 | 3630 | 1524 | 0.90 (0.84, 0.96) | 0.96 (0.88, 1.06) | 1.00 (0.91, 1.10) |
| P-trend | 0.0008 | 0.40 | 0.13 | ||
All statistical tests were conducted with use of Cox proportional hazards regression models. Q, quintile; ref, reference.
Adjusted for age, randomization treatment assignment, physical activity, history of hypercholesterolemia or hypertension, smoking status, postmenopausal status, postmenopausal hormone use, alcohol use, multivitamin use, and energy intake.
Because we found that the associations of fruit and vegetable intake with weight change varied by follow-up time, we performed sensitivity analyses on the risk of becoming overweight or obese by ending the follow-up after 8 y. In the multivariable-adjusted model including BMI, similar associations were observed for total fruit and vegetable intake (HR in the highest quintile: 0.94; 95% CI: 0.86, 1.04; P-trend: 0.69), fruit intake (HR in the highest quintile: 0.88; 95% CI: 0.80, 0.97; P-trend: 0.08), vegetable intake (HR in the highest quintile: 1.01; 95% CI: 0.92, 1.11; P-trend: 0.12), and dietary fiber intake (HR in the highest quintile: 1.07; 95% CI: 0.95, 1.20; P-trend: 0.82).
We further investigated the association between specific types of fruits and vegetables (Table 4) and the risk of becoming overweight or obese. In multivariable-adjusted models including BMI, we observed statistically significant inverse associations for intakes of citrus fruits, fruit juices, and legumes. Intakes of berries, green leafy vegetables, and cruciferous vegetables were associated with a higher risk of becoming overweight or obese. When mutually adjusting for types fruits and vegetables a stronger association was observed for legumes (HR in quintile 5 vs. quintile 1: 0.82; 95% CI: 0.76, 0.89), whereas the results were attenuated for citrus fruits (HR in quintile 5 vs. quintile 1: 0.89; 95% CI: 0.81, 0.97) and fruit juices (HR in quintile 5 vs. quintile 1: 0.81; 95% CI: 0.79, 0.95). We also investigated the association with tofu intake and observed that women who consumed ≥1 serving/wk had lower risk of becoming overweight or obese (HR: 0.78; 95% CI: 0.68, 0.89; P-trend: <0.0001) compared with those who consumed none or <1 time/mo of tofu products. Further adjustment for BMI slightly attenuated this association (HR: 0.84; 95% CI: 0.73, 0.96; P-trend: 0.007). We next investigated the shape of association between types of fruits and vegetables and risk of becoming overweight or obese by performing restricted cubic spline analysis (Supplemental Figure 1). We did not perform splines for the intakes of berries and tomatoes because their distributions were narrow. We found evidence of departure from linearity for citrus fruits (P < 0.0001), fruit juices (P < 0.0001), and legumes (P = 0.001).
TABLE 4.
HRs (95% CIs) of becoming overweight or obese according to subgroups of intakes of fruits and vegetables among middle-aged and older women1
| Age adjusted |
Multivariable adjusted |
||||
| n | Cases | HR (95% CI) | HR (95% CI)2 | HR (95% CI)2 + BMI | |
| Citrus fruits, servings/d | |||||
| Q1: <0.2 | 3919 | 1913 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.2 to <0.5 | 3683 | 1750 | 0.96 (0.90, 1.02) | 0.99 (0.93, 1.06) | 1.01 (0.95, 1.08) |
| Q3: 0.5 to <0.9 | 3578 | 1642 | 0.93 (0.87, 0.99) | 0.97 (0.91, 1.04) | 0.84 (0.78, 0.90) |
| Q4: 0.9 to <1.3 | 3433 | 1440 | 0.85 (0.79, 0.91) | 0.90 (0.84, 0.97) | 0.91 (0.85, 0.98) |
| Q5: ≥1.3 | 3521 | 1375 | 0.78 (0.73, 0.84) | 0.83 (0.77, 0.90) | 0.84 (0.78, 0.90) |
| P-trend | <0.0001 | <0.0001 | <0.0001 | ||
| Berries, servings/d | |||||
| Q1: <0.2 | 4369 | 1979 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.2 to <0.3 | 5080 | 2313 | 1.00 (0.94, 1.06) | 1.02 (0.96, 1.08) | 1.12 (1.06, 1.20) |
| Q3: 0.3 to <0.4 | 2697 | 1117 | 0.87 (0.81, 0.94) | 0.92 (0.85, 0.99) | 1.03 (0.96, 1.11) |
| Q4: 0.4 to <0.7 | 3442 | 1553 | 1.00 (0.93, 1.07) | 1.04 (0.97, 1.11) | 1.12 (1.05, 1.20) |
| Q5: ≥0.7 | 2514 | 1145 | 1.03 (0.95, 1.10) | 1.10 (1.02, 1.18) | 1.21 (1.11, 1.30) |
| P-trend | 0.47 | 0.02 | <0.0001 | ||
| Fruit juices, servings/d | |||||
| Q1: <0.1 | 4130 | 1986 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.1 to <0.3 | 3287 | 1606 | 0.99 (0.93, 1.06) | 0.99 (0.93, 1.06) | 0.98 (0.91, 1.04) |
| Q3: 0.3 to <0.5 | 3615 | 1636 | 0.91 (0.85, 0.97) | 0.91 (0.85, 0.97) | 0.79 (0.73, 0.84) |
| Q4: 0.5 to <0.8 | 3573 | 1488 | 0.84 (0.78, 0.89) | 0.85 (0.80, 0.91) | 0.88 (0.82, 0.95) |
| Q5: ≥0.8 | 3518 | 1402 | 0.77 (0.72, 0.82) | 0.78 (0.73, 0.84) | 0.81 (0.76, 0.88) |
| P-trend | <0.0001 | <0.0001 | <0.0001 | ||
| Green leafy vegetables, servings/d | |||||
| Q1: <0.3 | 4085 | 1870 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.3 to <0.5 | 3672 | 1669 | 0.97 (0.90, 1.03) | 0.99 (0.93, 1.06) | 1.10 (1.03, 1.17) |
| Q3: 0.5 to <0.8 | 3238 | 1430 | 0.93 (0.87, 1.00) | 0.98 (0.91, 1.05) | 1.09 (1.01, 1.17) |
| Q4: 0.8 to <1.1 | 3723 | 1633 | 0.95 (0.89, 1.01) | 1.01 (0.94, 1.08) | 1.11 (1.04, 1.19) |
| Q5: ≥1.1 | 3420 | 1521 | 0.97 (0.90, 1.03) | 1.08 (1.00, 1.16) | 1.15 (1.07, 1.24) |
| P-trend | 0.34 | 0.03 | 0.0005 | ||
| Cruciferous vegetables, servings/d | |||||
| Q1: <0.2 | 4644 | 2099 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.2 to <0.3 | 2931 | 1297 | 0.97 (0.90, 1.04) | 1.00 (0.93, 1.07) | 1.11 (1.04, 1.20) |
| Q3: 0.3 to <0.4 | 3593 | 1614 | 1.02 (0.96, 1.09) | 1.05 (0.98, 1.12) | 1.11 (1.04, 1.19) |
| Q4: 0.4 to <0.7 | 3416 | 1485 | 0.98 (0.91, 1.04) | 1.03 (0.96, 1.10) | 1.09 (1.02, 1.17) |
| Q5: ≥0.7 | 3561 | 1629 | 1.06 (1.00, 1.13) | 1.13 (1.05, 1.20) | 1.16 (1.09, 1.25) |
| P-trend | 0.06 | 0.0006 | 0.0005 | ||
| Dark yellow vegetables, servings/d | |||||
| Q1: <0.1 | 3685 | 1762 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.1 to <0.3 | 4152 | 1883 | 0.95 (0.89, 1.01) | 0.96 (0.90, 1.03) | 1.14 (1.06, 1.22) |
| Q3: 0.3 to <0.4 | 3827 | 1719 | 0.94 (0.88, 1.01) | 0.98 (0.91, 1.05) | 1.14 (1.06, 1.22) |
| Q4: 0.4 to <0.8 | 3025 | 1300 | 0.90 (0.84, 0.96) | 0.94 (0.88, 1.02) | 1.10 (1.02, 1.19) |
| Q5: ≥0.8 | 3451 | 1460 | 0.88 (0.82, 0.95) | 0.95 (0.88, 1.02) | 1.04 (0.96, 1.12) |
| P-trend | 0.0004 | 0.21 | 0.62 | ||
| Legumes, servings/d | |||||
| Q1: <0.2 | 4871 | 2214 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.2 to <0.3 | 3493 | 1575 | 0.99 (0.93, 1.06) | 1.00 (0.94, 1.07) | 0.94 (0.89, 1.01) |
| Q3: 0.3 to <0.4 | 2647 | 1153 | 0.94 (0.88, 1.01) | 0.94 (0.88, 1.01) | 0.92 (0.86, 0.99) |
| Q4: 0.4 to <0.6 | 4103 | 1874 | 1.02 (0.96, 1.09) | 1.03 (0.96, 1.10) | 0.84 (0.79, 0.90) |
| Q5: ≥0.6 | 3032 | 1309 | 0.96 (0.90, 1.03) | 0.98 (0.91, 1.05) | 0.87 (0.81, 0.94) |
| P-trend | 0.60 | 0.90 | <0.0001 | ||
| Tomatoes, servings/d | |||||
| Q1: <0.2 | 5283 | 2378 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2: 0.2 to <0.3 | 2251 | 977 | 0.92 (0.85, 0.99) | 0.94 (0.87, 1.01) | 0.92 (0.85, 0.99) |
| Q3: 0.3 to <0.6 | 4998 | 2228 | 1.00 (0.94, 1.06) | 1.04 (0.98, 1.10) | 0.90 (0.84, 0.95) |
| Q4: 0.6 to <0.9 | 2485 | 1103 | 0.97 (0.90, 1.04) | 1.03 (0.96, 1.11) | 0.97 (0.90, 1.04) |
| Q5: ≥0.9 | 3122 | 1434 | 1.04 (0.97, 1.11) | 1.12 (1.03, 1.20) | 1.02 (0.95, 1.09) |
| P-trend | 0.16 | 0.0004 | 0.32 | ||
All statistical tests were conducted with use of Cox proportional hazards regression models. Q, quintile; ref, reference.
Adjusted for age, randomization treatment assignment, physical activity, history of hypercholesterolemia or hypertension, smoking status, postmenopausal status, postmenopausal hormone use, alcohol use, multivitamin use, and energy intake.
In stratified analyses, we investigated whether the association between fruit and vegetable intake varied by factors such as baseline age, smoking status, physical activity, and BMI. The association seemed to vary by age (P-interaction: <0.0001), and inverse associations were observed among women 50 to <60 y of age (HR in quintile 5: 0.86; 95% CI: 0.76, 0.98; P-trend: 0.31) and >60 y of age (HR in quintile 5: 0.81; 95% CI: 0.64, 1.03; P-trend: 0.009) but not among women aged <50 y (HR in quintile 5: 1.07; 95% CI: 0.94, 1.21; P-trend: 0.38). In stratified analyses by smoking status, we observed a statistically significant inverse association among never smokers (HR in quintile 5: 0.86; 95% CI: 0.76, 0.97; P-trend: 0.35) but not among past smokers (HR in quintile 5: 1.15; 95% CI: 1.00, 1.32; P-trend: 0.37) and current smokers (HR in quintile 5: 0.87; 95% CI: 0.69, 1.09; P-trend: 0.53; P-interaction: <0.0001). The association between fruit and vegetable intake and risk of becoming overweight or obese also varied by physical activity where an inverse association was observed among women with <7.5 MET hours/wk (HR in quintile 5: 0.88; 95% CI: 0.77, 1.01; P-trend: 0.18) but not among women with ≥7.5 MET hours/wk (HR in quintile 5: 1.06; 95% CI: 0.94, 1.19; P-trend: 0.63). The test for interaction was statistically significant (P < 0.0001). In Figure 1, we present subgroup analyses by baseline BMI for intakes of total fruit and vegetables, fruits, and vegetables. The association between fruit intake and the risk of becoming overweight or obese was stronger among women who had a BMI <23 kg/m2, whereas no significant associations were observed among women with a BMI ranging from 23 to <24 or 24 to <25 kg/m2 (P-interaction: <0.0001). In stratified analyses for vegetable intake, among women with baseline BMI <23 kg/m2 and from 23 to <24 kg/m2, a nonsignificant higher risk of becoming overweight or obese was observed in the highest quintile compared to the lowest. Among women with baseline BMI ranging from 24 to <25 kg/m2, a lower risk of becoming overweight or obese was observed in the highest quintile of vegetable intake, however, the test for interaction was not significant (P = 0.29).
FIGURE 1.
The risk of becoming overweight or obese according to quintiles of total fruit and vegetable (A), fruit (B), and vegetable (C) intakes (servings/d) by subgroups of baseline BMI (<23, 23 to <24, and ≥24 kg/m2 among 18,146 middle-aged and older women). BMI <23 (n = 10,712); BMI 23 to <24 (n = 3669); BMI ≥24 (n = 3765). See Table 2 for category cutoffs. The Cox proportional hazards regression models were used to calculate relative risks (dots) and 95% CIs (bars). All relative risk estimates are adjusted for age, randomization treatment assignment, BMI, physical activity, history of hypercholesterolemia or hypertension, smoking status, postmenopausal status, postmenopausal hormone use, alcohol use, multivitamin use, and energy intake. The tests for interaction between intakes of total fruit and vegetable, fruit, and vegetables and BMI on risk of becoming overweight or obese were tested with the Wald test. Ref, reference.
Discussion
In this prospective cohort of middle-aged and older women, high fruit, but not vegetable, intake was inversely associated with the risk of becoming overweight or obese. No association was observed for dietary fiber intake. Moreover, we found no significant associations of fruit and dietary fiber intake with weight gain, whereas higher vegetable intake was associated with greater weight gain during 15.9 y of follow-up.
Although several observation studies (38–44) have investigated how different food patterns are associated with weight gain and risk of becoming overweight or obese, few have specifically investigated the impact of fruit and vegetable intake (16–19). In the European Prospective Investigation into Cancer and Nutrition Study, fruit and vegetable intake was not associated with weight change during a mean of 5 y of follow-up in 373,803 women and men aged between 25 and 70 y (17). However, in stratified analyses, an inverse association was observed between high fruit intake and weight change among women who were initially aged >50 y, of normal weight, never smokers, or had a low prudent dietary pattern score. The Nurses’ Health Study examined 74,063 women followed for 12 y (16), comparing women with the largest increase vs. decrease in fruit and vegetable intake, and the RR of becoming obese was 0.76 (95% CI: 0.69, 0.86; P-trend: <0.0001). In 79,236 women and men of the Cancer Prevention Study II, higher vegetable consumption was associated with lower odds of gaining weight over a 10-y period (18).
Randomized trials testing diets high in fruits and vegetables—typically among overweight or obese individuals—have been linked with weight loss in primary and secondary analyses (20–25). Two randomized trials, one including 45 women (21) and the other 90 women and men (20), have investigated the effect of fruit and vegetable intake on weight loss as the primary endpoint. In these 2 trials, following the participants for a 10-wk and 8-wk period, respectively, nonsignificant decreases in body weight were reported for the groups with diets higher in fruits and vegetables (20, 21). In a trial of 658 overweight or obese women and men followed for 6 mo, increasing intakes of fruit, vegetables, fiber, vitamins, and minerals reduced body weight by 5.1–6.1 kg (23). Another randomized trial of 97 obese women reduced the energy density of the diet through increased intake of fruits and vegetables and decreased fat intake, and the diet intervention resulted in both weight loss and maintenance (24). In a small randomized trial of 49 women and men, adding apples and pears but not oats reduced weight by 0.93 kg and 0.84 kg, respectively, during a 10-wk period (25).
Fruits and vegetables may prevent weight gain through several mechanisms. Micronutrients, e.g., potassium and magnesium, and calcium seem to be beneficial in weight control (45). Polyphenols, a group of bioactive compounds found in fruits and vegetables, may prevent weight gain through their antioxidant and anti-inflammatory properties by increasing thermogenesis and energy expenditure (46). Plant sterols have been suggested to interfere with fat absorption and thereby prevent weight gain. Moreover, fruits and vegetables also have low energy density and are high in water and dietary fiber content, which affect satiety (9). On the other hand, in our analyses, we observed only a modest lower risk of becoming overweight or obese for fruit intake, and this did not extend to analyses for weight change. Moreover, higher vegetable intake was associated with a slightly higher risk of weight gain; however, the magnitude of effect between the lowest and highest intakes was very small, for which these results should be interpreted with caution. Taken together, our results do not strongly support a diet characterized by high fruit and vegetable intake to lessen weight gain or lower the risk of becoming overweight or obese.
Because baseline body weight is very likely to affect the risk of becoming overweight or obese, we investigated how our observed findings were affected by baseline BMI. We found that the association between fruit intake and risk of becoming overweight or obese may be modified by baseline BMI. The biological mechanism behind this observation remains unclear but highlights that increasing fruit intake may prevent weight gain especially among women within more favorable levels of normal BMI. We also investigated effect modification by other factors such as age, smoking, and physical activity. In these analyses, there was evidence that the association between fruit and vegetable consumption and the risk of becoming overweight or obese could be modified by these factors.
Our study has several strengths. We studied a large prospective cohort with high rates of follow-up, during which women provided high-quality self-reports of lifestyle, clinical, and dietary factors, as well as updated body weight. However, there are also important potential limitations to consider. Random misclassification of self-reported body weight may have biased our risk estimates. Moreover, our results may also be biased by measurement error in self-reported fruit and vegetable consumption. The validity and reproducibility of the FFQ was not investigated in the current study population, however, previous studies in similar study populations have demonstrated reasonable validity and reproducibility of intake of fruits, vegetables, and dietary fiber (26, 27). Moreover, we only measured fruit and vegetable consumption at baseline and therefore cannot rule out whether women changed their diet during the follow-up. Although we had available information on a wide range of potential confounders, we cannot rule out uncontrolled confounding by other factors. We cannot exclude the possibility that the large number of comparisons may have generated false statistically significant associations. We also did not have information on food preparation methods, which may have influenced our results. Finally, women in the Women’s Health Study were predominantly Caucasian health care professionals, which may limit the generalizability of our results to other populations.
In conclusion, in this prospective study of middle-aged and older women with a normal BMI at baseline, higher self-reported baseline intake of fruits, but not vegetables or fiber, was associated with lower risk of becoming overweight or obese.
Supplementary Material
Acknowledgments
SR and HDS designed the research; SR, LW, I-ML, JEB, and HDS conducted the research; SR, LW, I-ML, JEM, JEB, and HDS analyzed the data; SR and HDS wrote the paper; and HDS had primary responsibility for the final content. All authors read and approved the final manuscript.
References
- 1.WHO. Obesity: preventing and managing the global epidemic. Geneva (Switzerland): WHO; 2000. [PubMed]
- 2.Ogden CLCM, Kit BK, Flegal KM. Prevalence of obesity in the United States, 2009–2010. NCHS data brief, no. 82. Hyattsville (MD): National Center for Health Statistics; 2012. [PubMed]
- 3.Kumanyika SK, Obarzanek E, Stettler N, Bell R, Field AE, Fortmann SP, Franklin BA, Gillman MW, Lewis CE, Poston WC II, et al. . Population-based prevention of obesity: the need for comprehensive promotion of healthful eating, physical activity, and energy balance: a scientific statement from American Heart Association Council on Epidemiology and Prevention, Interdisciplinary Committee for Prevention (formerly the expert panel on population and prevention science). Circulation 2008;118:428–64. [DOI] [PubMed] [Google Scholar]
- 4.Mente A, de Koning L, Shannon HS, Anand SS. A systematic review of the evidence supporting a causal link between dietary factors and coronary heart disease. Arch Intern Med 2009;169:659–69. [DOI] [PubMed] [Google Scholar]
- 5.He FJ, Nowson CA, MacGregor GA. Fruit and vegetable consumption and stroke: meta-analysis of cohort studies. Lancet 2006;367:320–6. [DOI] [PubMed] [Google Scholar]
- 6.Key TJ. Fruit and vegetables and cancer risk. Br J Cancer 2011;104:6–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Boeing H, Bechthold A, Bub A, Ellinger S, Haller D, Kroke A, Leschik-Bonnet E, Muller MJ, Oberritter H, Schulze M, et al. . Critical review: vegetables and fruit in the prevention of chronic diseases. Eur J Nutr 2012;51:637–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rebello CJ, Liu AG, Greenway FL, Dhurandhar NV. Dietary strategies to increase satiety. Adv Food Nutr Res 2013;69:105–82. [DOI] [PubMed] [Google Scholar]
- 9.Rolls BJ, Ello-Martin JA, Tohill BC. What can intervention studies tell us about the relationship between fruit and vegetable consumption and weight management? Nutr Rev 2004;62:1–17. [DOI] [PubMed] [Google Scholar]
- 10.Serdula MK, Byers T, Mokdad AH, Simoes E, Mendlein JM, Coates RJ. The association between fruit and vegetable intake and chronic disease risk factors. Epidemiology 1996;7:161–5. [DOI] [PubMed] [Google Scholar]
- 11.Trudeau E, Kristal AR, Li S, Patterson RE. Demographic and psychosocial predictors of fruit and vegetable intakes differ: implications for dietary interventions. J Am Diet Assoc 1998;98:1412–7. [DOI] [PubMed] [Google Scholar]
- 12.Williams DE, Wareham NJ, Cox BD, Byrne CD, Hales CN, Day NE. Frequent salad vegetable consumption is associated with a reduction in the risk of diabetes mellitus. J Clin Epidemiol 1999;52:329–35. [DOI] [PubMed] [Google Scholar]
- 13.Flood A, Velie EM, Chaterjee N, Subar AF, Thompson FE, Lacey JV Jr, Schairer C, Troisi R, Schatzkin A. Fruit and vegetable intakes and the risk of colorectal cancer in the Breast Cancer Detection Demonstration Project follow-up cohort. Am J Clin Nutr 2002;75:936–43. [DOI] [PubMed] [Google Scholar]
- 14.Bazzano LA, He J, Ogden LG, Loria CM, Vupputuri S, Myers L, Whelton PK. Fruit and vegetable intake and risk of cardiovascular disease in US adults: the first National Health and Nutrition Examination Survey Epidemiologic Follow-up Study. Am J Clin Nutr 2002;76:93–9. [DOI] [PubMed] [Google Scholar]
- 15.Rissanen TH, Voutilainen S, Virtanen JK, Venho B, Vanharanta M, Mursu J, Salonen JT. Low intake of fruits, berries and vegetables is associated with excess mortality in men: the Kuopio Ischaemic Heart Disease Risk Factor (KIHD) Study. J Nutr 2003;133:199–204. [DOI] [PubMed] [Google Scholar]
- 16.He K, Hu FB, Colditz GA, Manson JE, Willett WC, Liu S. Changes in intake of fruits and vegetables in relation to risk of obesity and weight gain among middle-aged women. Int J Obes Relat Metab Disord 2004;28:1569–74. [DOI] [PubMed] [Google Scholar]
- 17.Vergnaud AC, Norat T, Romaguera D, Mouw T, May AM, Romieu I, Freisling H, Slimani N, Boutron-Ruault MC, Clavel-Chapelon F, et al. . Fruit and vegetable consumption and prospective weight change in participants of the European Prospective Investigation into Cancer and Nutrition-Physical Activity, Nutrition, Alcohol, Cessation of Smoking, Eating Out of Home, and Obesity study. Am J Clin Nutr 2012;95:184–93. [DOI] [PubMed] [Google Scholar]
- 18.Kahn HS, Tatham LM, Rodriguez C, Calle EE, Thun MJ, Heath CW Jr. Stable behaviors associated with adults’ 10-year change in body mass index and likelihood of gain at the waist. Am J Public Health 1997;87:747–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Parker DR, Gonzalez S, Derby CA, Gans KM, Lasater TM, Carleton RA. Dietary factors in relation to weight change among men and women from two southeastern New England communities. Int J Obes Relat Metab Disord 1997;21:103–9. [DOI] [PubMed] [Google Scholar]
- 20.Whybrow S, Harrison CL, Mayer C, James Stubbs R. Effects of added fruits and vegetables on dietary intakes and body weight in Scottish adults. Br J Nutr 2006;95:496–503. [DOI] [PubMed] [Google Scholar]
- 21.Wagner M. Effects of nutrition education and fruit and vegetable consumption on knowledge of antioxidants and biomarkers of inflammation and chronic disease. Fargo (ND): North Dakota State University; 2012.
- 22.Lanza E, Schatzkin A, Daston C, Corle D, Freedman L, Ballard-Barbash R, Caan B, Lance P, Marshall J, Iber F, et al. . Implementation of a 4-y, high-fiber, high-fruit-and-vegetable, low-fat dietary intervention: results of dietary changes in the Polyp Prevention Trial. Am J Clin Nutr 2001;74:387–401. [DOI] [PubMed] [Google Scholar]
- 23.Ledikwe JH, Rolls BJ, Smiciklas-Wright H, Mitchell DC, Ard JD, Champagne C, Karanja N, Lin PH, Stevens VJ, Appel LJ. Reductions in dietary energy density are associated with weight loss in overweight and obese participants in the PREMIER trial. Am J Clin Nutr 2007;85:1212–21. [DOI] [PubMed] [Google Scholar]
- 24.Ello-Martin JA, Roe LS, Ledikwe JH, Beach AM, Rolls BJ. Dietary energy density in the treatment of obesity: a year-long trial comparing 2 weight-loss diets. Am J Clin Nutr 2007;85:1465–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.de Oliveira MC, Sichieri R, Venturim Mozzer R. A low-energy-dense diet adding fruit reduces weight and energy intake in women. Appetite 2008;51:291–5. [DOI] [PubMed] [Google Scholar]
- 26.Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, Hennekens CH, Speizer FE. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol 1985;122:51–65. [DOI] [PubMed] [Google Scholar]
- 27.Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB, Willett WC. Validity of self-reported waist and hip circumferences in men and women. Epidemiology 1990;1:466–73. [DOI] [PubMed] [Google Scholar]
- 28.USDA. Composition of foods—raw, processed, and prepared, 1963–1988. Agricultural handbook no. 8 series. Washington (DC): Government Printing Office; 1989.
- 29.Prosky L, Asp NG, Furda I, Devries JW, Schweizer TF, Harland BF. Determination of total dietary fiber in foods, food-products, and total diets—interlaboratory study. J Assoc Off Anal Chem 1984;67:1044–52. [Google Scholar]
- 30.Prosky L, Asp NG, Furda I, Devries JW, Schweizer TF, Harland BF. Determination of total dietary fiber in foods and food-products—collaborative study. J Assoc Off Anal Chem 1985;68:677–9. [PubMed] [Google Scholar]
- 31.Prosky L, Asp NG, Schweizer TF, Devries JW, Furda I. Determination of insoluble, soluble, and total dietary fiber in foods and food-products—interlaboratory study. J Assoc Off Anal Chem 1988;71:1017–23. [PubMed] [Google Scholar]
- 32.Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986;124:17–27. [DOI] [PubMed] [Google Scholar]
- 33.Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol 1992;135:1114–26. [DOI] [PubMed] [Google Scholar]
- 34.Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert Panel on the Identification, Evaluation, and Treatment of Overweight in Adults. Am J Clin Nutr 1998;68:899–917. [DOI] [PubMed] [Google Scholar]
- 35.Cox D. Regression models and life-tables. J R Stat Soc B 1972;34:187–220. [Google Scholar]
- 36.Glymour MM, Weuve J, Berkman LF, Kawachi I, Robins JM. When is baseline adjustment useful in analyses of change? An example with education and cognitive change. Am J Epidemiol 2005;162:267–78. [DOI] [PubMed] [Google Scholar]
- 37.Harrell FE Jr, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst 1988;80:1198–202. [DOI] [PubMed] [Google Scholar]
- 38.McNaughton SA, Mishra GD, Stephen AM, Wadsworth ME. Dietary patterns throughout adult life are associated with body mass index, waist circumference, blood pressure, and red cell folate. J Nutr 2007;137:99–105. [DOI] [PubMed] [Google Scholar]
- 39.Schulze MB, Fung TT, Manson JE, Willett WC, Hu FB. Dietary patterns and changes in body weight in women. Obesity (Silver Spring) 2006;14:1444–53. [DOI] [PubMed] [Google Scholar]
- 40.Newby PK, Weismayer C, Akesson A, Tucker KL, Wolk A. Longitudinal changes in food patterns predict changes in weight and body mass index and the effects are greatest in obese women. J Nutr 2006;136:2580–7. [DOI] [PubMed] [Google Scholar]
- 41.Togo P, Osler M, Sorensen TI, Heitmann BL. A longitudinal study of food intake patterns and obesity in adult Danish men and women. Int J Obes Relat Metab Disord 2004;28:583–93. [DOI] [PubMed] [Google Scholar]
- 42.Newby PK, Muller D, Hallfrisch J, Qiao N, Andres R, Tucker KL. Dietary patterns and changes in body mass index and waist circumference in adults. Am J Clin Nutr 2003;77:1417–25. [DOI] [PubMed] [Google Scholar]
- 43.Quatromoni PA, Copenhafer DL, D'Agostino RB, Millen BE. Dietary patterns predict the development of overweight in women: the Framingham Nutrition Studies. J Am Diet Assoc 2002;102:1239–46. [DOI] [PubMed] [Google Scholar]
- 44.Togo P, Osler M, Sorensen TI, Heitmann BL. Food intake patterns and body mass index in observational studies. Int J Obes Relat Metab Disord 2001;25:1741–51. [DOI] [PubMed] [Google Scholar]
- 45.Abete I, Astrup A, Martinez JA, Thorsdottir I, Zulet MA. Obesity and the metabolic syndrome: role of different dietary macronutrient distribution patterns and specific nutritional components on weight loss and maintenance. Nutr Rev 2010;68:214–31. [DOI] [PubMed] [Google Scholar]
- 46.Wang S, Moustaid-Moussa N, Chen L, Mo H, Shastri A, Su R, Bapat P, Kwun I, Shen CL. Novel insights of dietary polyphenols and obesity. J Nutr Biochem 2014;25:1–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.