Whole-Grain Intake and Cereal Fiber Are Associated with Lower Abdominal Adiposity in Older Adults

Abstract

Foods high in dietary fiber may play an important role in regulating body weight. Few observational studies have examined the relationship between dietary fiber from different sources and body fat in older adults. Our objectives were to examine the associations among grain intake (whole and refined), dietary fiber and fiber sources, and body fat among older adults. We used data from 434 free-living adults (177 men and 257 women) aged between 60 and 80 y. Dietary intake was estimated from a 126-item semiquantitative FFQ. Percent body fat and percent trunk fat mass were measured by whole-body dual-energy X-ray absorptiometry. After adjustment for covariates, whole-grain intake was inversely associated with BMI [26.8 kg/m² (25.7–28.1) vs. 25.8 kg/m² (24.6–27.1), (95% CI); P-trend = 0.08], percent body fat [34.5% (32.7–36.3) vs. 32.1% (30.1–34.1); P-trend = 0.02], and percent trunk fat mass [43.0% (40.4–45.5) vs. 39.4% (36.7–42.1); P-trend = 0.02] in the lowest compared with the highest quartile category of whole-grain intake. Refined grain intake was not associated with any measure of body fat distribution. Cereal fiber was inversely associated with BMI [27.3 kg/m² (26.1–28.6) vs. 25.4 kg/m² (24.3–26.7); P-trend = 0.012], percent body fat [34.7% (32.8–36.6) vs. 31.5% (29.4–33.5); P-trend = 0.004], and percent trunk fat mass [42.8% (40.2–45.4) vs. 37.8% (35.0–40.6); P-trend = 0.001]. No significant association was observed between intakes of total fiber, vegetable or fruit fiber, and body composition measurements. Higher intakes of cereal fiber, particularly from whole-grain sources, are associated with lower total percent body fat and percent trunk fat mass in older adults.

Introduction

In the United States, the prevalence of obesity is increasing across all age groups, including the older segment of the population (1). Based on recent data from the NHANES, an estimated 62.5% of men and 74.9% of women aged between 60–69 y are abdominally obese (2). Abdominal adiposity (i.e. the accumulation of adipose tissue in the abdominal region) may be a more predictive risk factor of chronic disease and mortality than BMI, particularly in older adults (3–5). Lifestyle modifications known to affect body weight can translate into improved health outcomes in this age group. In adults aged ≥60 y with impaired fasting glucose, modest weight loss (∼5–6%) reduced the incidence of type 2 diabetes mellitus by 71% over a 2- to 3-y period (6). In overweight and obese older adults, a diet-induced weight-loss intervention for 18 mo resulted in a significant reduction of adipocyte-derived inflammatory cytokines (7).

Evidence from epidemiological studies has shown that middle-aged adults who consume more whole grains have a lower BMI (8) and central obesity (9,10) and tend to gain significantly less weight over time compared with those individuals mainly consuming refined grains (11). Dietary fiber from cereals, rather than fruits or vegetables, appears to be more protective against the development of several chronic diseases, including metabolic syndrome (12), type 2 diabetes (13–15), and cardiovascular disease (16,17). Dietary fiber from different food sources may affect disease risk via different mechanisms involved in regulating and maintaining body weight. For instance, soluble fiber, derived mainly from fruits and vegetable, may delay gastric emptying and affect insulin secretion and action, whereas insoluble dietary fiber, derived mainly from cereal sources, may activate the release of gut hormones involved in regulating food intake (18,19). Currently, there are limited observational data on the association between different types of grain and sources of dietary fiber and measures of body fat distribution, particularly in older adults.

Our primary objective in this study was to examine the association between whole and refined grain intake and measures of body fat distribution derived from dual energy X-ray absorptiometry (DXA) in older adults (>60 y). Recently, DXA has been shown to provide a valid measure of abdominal adiposity in older adults (20). Because the physiologic consequences of a high-fiber diet may depend on the source of the fiber, a secondary objective was to examine the relations between cereal, fruit and vegetable fiber, and body fat distribution in this study sample.

Materials and Methods

Study population.

This study is an ancillary analysis using data from a 3-y, double-blind, controlled trial examining the effect of vitamin K supplementation on age-related bone turnover, bone loss, and vascular calcification (21). This supplementation trial began in 2002 and dietary, anthropometric, and biochemical measures were collected at baseline on all participants prior to randomization. After exclusion for bone and vascular health-related risk factors as described elsewhere (21), a total of 452 free-living adults (185 men and 267 women) aged between 60–80 y participated in this trial. The majority of the participants recruited were Caucasian (93%). Of the 452 participants who were enrolled in this trial, 4% were excluded from the analysis, because they did not complete an FFQ (n = 6), provided an invalid FFQ (n = 9), or had missing covariate information (n = 3). The final sample consisted of 434 (177 men and 257 women), with a mean age of 68 y. This study was approved by the Tufts Medical Center Institutional Review Board and registered with ClinicalTrials.gov (no. NCT00183001).

Assessment of dietary intake.

Usual dietary intake during the previous year was assessed with the use of a semiquantitative, 126-item FFQ (22). The questionnaires were mailed to the participants before the examination. The participants brought their completed FFQ with them to their appointment for review with a study team member. The FFQ consisted of a list of foods with a standard serving size and a selection of 9 frequency categories ranging from never or <1 serving/mo to >6 servings/d. Nutrient intake was calculated by multiplying the frequency of consumption of each unit of food from the FFQ by the nutrient contents of the specific portion. Dietary information was considered valid if reported energy intakes were ≥2.51 MJ/d (600 kcal/d) for both men and women and if energy intakes were <6.74 MJ/d (4000 kcal/d) for women and <17.54 MJ/d (4200 kcal/d) for men and fewer than 13 food items were left blank.

The FFQ included questions regarding the consumption of whole-grain foods such as cooked and cold breakfast cereals, dark bread, brown rice, popcorn, and other grains (e.g. bulgur, kasha, and couscous). A specific question asked the study participants to provide the brand and type of cold breakfast cereal usually eaten. Breakfast cereal intake was subdivided into whole and refined grain based on the content of whole grain or bran of the cereal, as previously reported (23,24). A breakfast cereal was considered whole grain if it contained ≥25% whole grain or bran by weight. With respect to bran intake, a specific line-item is included in the questionnaire, i.e. “bran, added to food (1 tablespoon).” Thus, bran that was added to foods to increase the fiber content was considered as added bran. Similarly, a line item “wheat germ (1 tablespoon)” is included on the FFQ and is considered “added germ.” Refined-grain foods included cold breakfast cereals (<25% whole grain or bran), white bread, English muffins/bagels, muffins/biscuits, white rice, pasta, pancakes/waffles, crackers, and pizza. In the absence of information on brand names, breakfast cereals were classified as missing.

For comparison purposes, whole grain was also estimated as g/d using a recently developed food composition database for whole grains, as described in detail elsewhere (25). In brief, to estimate whole-grain ingredients, product labels and USDA national nutrient database information were applied to all foods on the FFQ and the percentage of whole-grain concentration per food serving was calculated on a dry weight basis. The whole-grain percentage for each food was multiplied by the gram weight per serving and expressed in g/d. By estimating whole-grain intake as g/d, the use of an arbitrary cut-point to classify a food as a whole-grain food is avoided. The Spearman correlation between whole-grain intake estimated as g/d and servings per day is r = 0.83 (P < 0.0001).

In this study sample, cold and hot breakfast cereals contributed ∼50% of the whole-grain intake. Pearson correlation coefficient between daily intakes of cold and hot breakfast cereals estimates from a FFQ and diet records are high (r ≥ 0.70) (26,27). The correlation coefficient for other sources of whole grains ranged between 0.37 for dark bread and 0.79 for popcorn (27).

The amount of dietary fiber for each food item was calculated and summed into 4 mutually exclusive fiber categories: cereal, fruit, vegetable, and legume. The total fiber for each fiber source was expressed as g/d. In this study, mean intake of legume fiber was low and therefore we did not consider it in subgroup analysis. The Spearman correlation between whole-grain (servings/d) and cereal fiber (g) was r = 0.77 (P < 0.0001) and refined grain (servings/d) and cereal fiber (g) was r = 0.32 (P < 0.0001). Although the relative validity of the FFQ for the different fiber sources has not been reported, this FFQ provided a relatively accurate estimate of dietary fiber intake (r = 0.64) after within-person variation was taken into account (22).

Body composition.

A medical history and physical examination were performed by a nurse practitioner. Body weight was measured to the nearest 0.1 kg with a standard balance beam scale. Height was measured in participants lightly clothed and without shoes to the nearest 0.1 cm using a Harpenden stadiometer. BMI was calculated as kg/m². Percent body fat was measured from a whole-body DXA scan using a GE-Lunar model Prodigy scanner (enCORE 2002; version 6.10.029). The software provides estimates of lean tissue mass, whole-body fat mass, and regional fat mass (trunk, legs, and arms). Abdominal adiposity was manually calculated from the DXA scans by the same trained technician according to the method of Park et al. (28) and percent trunk fat mass was calculated from the grams of fat and lean tissue in the specific region of interest (ROI). Briefly, the specific ROI was defined as the fat mass located between the upper edge of the second lumbar vertebra to above the iliac crest (28). This specific ROI provides a valid estimate of abdominal adiposity. High correlations with measures of abdominal adipose tissue derived from MRI (r = 0.93) have been reported. All scans were analyzed by a single observer. Abdominal adiposity (i. e., trunk fat mass) was reported as percent fat of the ROI. Due to difficulties in the ascertainment of the correct placement of the ROI, percent trunk fat mass measures were available for 373 participants. The Spearman correlation coefficient between percent total body fat and percent abdominal ROI fat was high (r = 0.88; P < 0.0001).

Covariate assessment.

Potential confounding factors included in the statistical model were sex, age, smoking status (yes/no), alcohol intake (g/d), percent energy from total fat, current multivitamin use (yes/no), total energy intake (kcal/d), and physical activity assessed using the Physical Activity Scale for the Elderly, which is designed to measure the level of physical activity in individuals aged ≥65 y (29).

Statistical analyses.

Statistical analyses were conducted using SAS statistical software (version 9.1.3; SAS Institute). Dietary exposures included whole-grain intake, refined grain intake, dietary fiber, and fiber sources (cereal, fruit, and vegetable).

We calculated the age-, sex-, and energy-adjusted means for lifestyle and dietary characteristics across increasing quartile categories of whole-grain, refined grain, dietary fiber, and fiber sources (fruit, vegetable, and cereal) by using SAS PROC GLM. We assessed the significance (defined as a 2-tailed P-value < 0.05) of trends across categories of intake using linear (for continuous lifestyle and dietary variables) or logistic regression (for dichotomous outcome variables). Least squared adjusted means are presented.

General linear models were used to examine the association between dietary intakes and measures of body fat. Dietary intake variables were characterized into quartile categories of intake. To assess trends across quartile categories, we assigned the median intake of each quartile category to individuals with intakes in the category and then included this quartile median variable as a continuous factor in the multiple regression models. The P for trend was the P-value for the associated multiple regression coefficient. In multivariate models, we controlled for sex, age, total energy intake, percent of energy from fat, physical activity, smoking status, alcohol intake, and multivitamin use. For the independent variables that were significantly associated with the outcome (i.e. cereal fiber and whole grains), we further adjusted for the percent of energy from carbohydrates, fruit fiber and vegetable fiber to determine whether these associations were independent of other dietary factors. We tested each association for effect modification by sex and found no significant interaction so data are presented for women and men combined. All correlation coefficients reported were calculated as Spearman rank-order correlation coefficients. Values in the text are mean ± SD unless otherwise noted.

Results

The age of participants was 68 ± 6 y (range: 60–81 y). Based on calculation of whole-grain intake and refined grain intake as servings per day, the intakes were 1.5 ± 1.4 and 1.9 ± 1.4, respectively. The median intakes of added bran and added germ were 0.39 and 0 g/d, respectively, suggesting that these food sources were not major contributors to whole-grain intake. The daily intakes of total dietary fiber, cereal, fruit, and vegetable fiber were 18.6 ± 7.6, 5.6 ± 3.2, 4.2 ± 2.9, and 5.5 ± 2.7 g/d, respectively.

In this study population, based on the servings per week, the major dietary sources of whole grains included dark bread (40%), cold breakfast cereal (33%), hot breakfast cereal (19%), and brown rice (7.5%). The main sources contributing to cereal fiber included cold breakfast cereal (21%), hot breakfast cereal (11%), dark bread (11%), pasta (10%), English muffins/bagels (7%), and pizza (6%). Spearman correlation coefficient between cereal fiber and whole-grains was r = 0.77 (P < 0.0001). After adjustment for energy intake, there was a relatively low correlation between energy-adjusted dietary fiber intake from fruit and vegetables (r = 0.20; P < 0.001) and cereal fiber and fruit fiber intake (r = 0.12; P = 0.01). Energy-adjusted vegetable fiber and cereal fiber were not correlated (r = 0.05; P = 0.29).

Study participant's characteristics across quartile categories of whole and refined grain and dietary fiber and fiber sources are shown in Supplemental Tables 1 and 2. The participants with a higher intake of whole grains were more likely to be women and to take multivitamins and were less likely to smoke. In contrast, participants with a higher intake of refined grains were more likely to be men and be more physically active. Whole- and refined-grain intakes were positively associated with energy intake, but only whole-grain intake was associated with lower fat intake and higher carbohydrate intake. Compared with participants in the lowest quartile category, those in the highest quartile category of dietary fiber were more likely to be women, more likely to take multivitamins, be more physically active, and less likely to smoke compared with those in the lowest quartile category. With respect to the fiber sources, women were more likely to consume higher intakes of fruits and vegetables and participants with higher intakes of vegetable fiber were more likely to engage in regular physical activity. Participants with higher intakes of cereal and fruit fiber intake were more likely to take multivitamins and less likely to be current smokers. Intakes of dietary fiber, cereal, fruit, and vegetable fibers were positively associated with total energy, potassium, and magnesium intakes and inversely associated with energy from total fat. With the exception of vegetable fiber, percent energy from carbohydrate was highest in individuals in the higher quintile categories of intakes. Fruit fiber was inversely, whereas vegetable fiber was positively, associated with alcohol intake.

Multivariate-adjusted mean BMI and measures of body fat across quartile categories of whole-grain and refined grain intakes are shown in Table 1. After adjustment for potential confounding variables associated with diets high in whole grains, a higher intake of whole grains was inversely associated with percent body fat and percent trunk fat mass and a marginally significant association with BMI (P-trend = 0.08). Although a significant inverse association between total bran and total germ was observed with the measures of body fat (data not presented), these components of the grain were highly correlated with each other (total bran and whole grain: r = 0.90; total germ and whole grain: r = 0.81) and due to the small sample size, we could not examine their independent effects. No associations were observed between refined grain intake and any measure of body fat. The relationship between whole-grain intake and measures of body fat were similar, although slightly stronger, based on the more quantitative measure of whole grain as g/d. After adjustment for covariates, whole-grain intake was inversely associated with BMI [26.9 kg/m² (25.8–28.1) vs. 25.4 kg/m² (24.2–26.6), (95% CI); P-trend = 0.006], percent body fat [34.4% (32.6–36.2) vs. 31.7% (29.7–33.6); P-trend = 0.007], and percent trunk fat mass [42.4% (39.9–44.8) vs. 38.4% (35.7–41.1); P-trend = 0.002]. The association between whole-grain intake and the measures of body composition remained significant after adjustment for refined grain intake.

TABLE 1.

Multivariate-adjusted means (95% CI) of body weight, BMI, and body composition measurements by quartiles of whole and refined grain intake¹²

	Q1	Q2	Q3	Q4	P-trend
Whole grain
n	108	109	109	108
Median intake, servings/d	0.21	0.86	1.57	2.86
BMI, kg/m2	26.8 (25.7–28.1)	26.8 (25.6–28.1)	25.9 (24.7–27.1)	25.8 (24.6–27.1)	0.08
Body fat, %	34.5 (32.7–36.3)	33.4 (31.5–35.3)	32.9 (31.0–34.8)	32.1 (30.1–34.1)	0.02
Trunk fat mass, %	43.0 (40.4–45.5)	40.3 (37.7–42.9)	39.5 (36.9–42.1)	39.4 (36.7–42.1)	0.02
Refined grain
n	108	109	109	108
Median intake, servings/d	0.64	1.22	2.00	3.25
BMI, kg/m2	26.4 (25.1–27.7)	26.4 (25.2–27.7)	26.2 (25.1–27.5)	26.5 (25.3–27.8)	0.84
Body fat, %	33.5 (31.5–35.4)	33.6 (31.7–35.5)	33.2 (31.3–35.0)	33.4 (31.4–35.3)	0.86
Trunk fat mass, %	40.5 (37.8–43.2)	40.7 (38.1–43.3)	40.5 (37.9–43.0)	41.4 (38.7–44.2)	0.57

¹Adjusted for age, sex, total energy intake, percent energy from fat, physical activity, smoking, alcohol intake, and multivitamin use.

²Geometric mean (95% CI).

Multivariate-adjusted mean of BMI and measures of body fat across quartile categories of dietary fiber and fiber sources are shown in Table 2. After adjustment for potential confounding variables, cereal fiber intake was inversely associated with BMI (kg/m²), percent body fat, and percent trunk fat mass. Total fiber and fiber from fruits and vegetables was not associated with the measures of body composition. The association between whole-grain intake and cereal fiber and measures of body composition remained significant after adjusting for percentage energy from total carbohydrate intake and adjustment for the other dietary fiber sources, i.e. fruit and vegetable fiber. Adjustment for medication use, including use of statin medication or blood pressure-lowering medication, did not affect these results.

TABLE 2.

Multivariate-adjusted means (95% CI) of body weight, BMI and body composition measurements by quartiles of fiber intake¹²

	Q1	Q2	Q3	Q4	P-trend
Total fiber
n	108	109	109	108
Range	4.4–13.3	13.3–17.6	17.6–22.5	22.5–45.5
Median intake, g/d	10.8	15.6	19.5	26.8
BMI, kg/m2	27.2 (25.9–28.5)	26.0 (24.9–27.2)	25.6 (24.5–26.8)	26.6 (25.2–28.0)	0.59
Body fat, %	34.5 (32.6–36.5)	32.9 (31.1–34.8)	32.5 (30.5–34.4)	33.3 (31.1–35.5)	0.43
Trunk fat mass, %	42.1 (39.5–44.7)	39.2 (36.6–41.9)	40.0 (37.4–42.6)	41.0 (38.0–44.1)	0.77
Cereal fiber
n	108	109	109	108
Range	0.9–3.2	3.2–4.9	5.0–7.1	7.2–24.1
Median intake, g/d	2.4	4.1	5.8	9.3
BMI, kg/m2	27.3 (26.1–28.6)	26.5 (25.3–27.7)	26.0 (24.8–27.2)	25.4 (24.2–26.7)	0.01
Body fat, %	34.7 (32.8–36.6)	33.9 (32.1–35.8)	32.8 (30.9–34.7)	31.5 (29.4–33.5)	0.004
Trunk fat mass, %	42.8 (40.2–45.4)	41.5 (38.9–44.1)	40.2 (37.6–42.8)	37.8 (35.0–40.6)	0.001
Fruit fiber
n	108	109	109	108
Range	0–2.2	2.2–3.7	3.7–5.6	5.6–20.7
Median intake, g/d	1.4	2.9	4.5	7.3
BMI, kg/m2	27.1 (26.0–28.4)	25.5 (24.3–26.7)	25.9 (24.7–27.1)	26.5 (25.2–27.8)	0.79
Body fat, %	34.2 (32.4–36.1)	32.5 (30.5–34.4)	33.0 (31.0–34.9)	33.2 (31.1–35.2)	0.56
Trunk fat mass, %	42.0 (39.5–44.5)	39.3 (36.7–41.9)	40.1 (37.4–42.8)	41.0 (38.2–43.8)	0.88
Vegetable fiber
n	108	109	109	108
Range	0–3.5	3.5–5.1	5.1–7.0	7.0–16.1
Median intake, g/d	2.5	4.3	5.8	8.9
BMI, kg/m2	26.8 (25.6–28.1)	26.3 (25.1–27.6)	25.8 (24.7–27.0)	26.6 (25.4–28.0)	0.89
Body fat, %	33.5 (31.6–35.4)	33.0 (31.0–34.9)	32.9 (31.1–34.8)	34.2 (32.2–36.3)	0.41
Trunk fat mass,%	40.3 (37.7–43.0)	40.2 (37.6–42.9)	40.5 (38.0–43.1)	42.3 (39.4–45.1)	0.19

¹Adjusted for age, sex, total energy intake, percent energy from fat, physical activity, smoking, alcohol intake, and multivitamin use.

²Geometric mean (95% CI).

Discussion

In this cross-sectional study among older adults, higher consumption of whole-grain foods was associated in a dose-dependent manner with a significantly lower percentage of abdominal fat as determined by DXA. No relationship was observed between refined grain intake and measures of body fat. Consistent with other studies (8), individuals with higher intakes of whole-grain foods had significantly lower BMI, independent of other healthy lifestyle behaviors. This inverse association between whole-grain intake and BMI is consistent across studies in older adults, regardless of the dietary method used to assess whole-grain intakes (10,30). Consistent with other studies (30,31), the average intake of whole grains was low in this sample of older adults, with average intake approximating 1.5 serving/d of whole grain. More recently, observational studies have estimated whole-grain intake in g/d, rather than in servings, to avoid the use of an arbitrary cut-point to classify a food as a whole-grain food. A quantitative estimate of whole grains (in g/d), however, may provide a better estimate of whole-grain intake as it takes into consideration the whole-grain content per serving of food. Based on this estimate of whole-grain intake, the predominant source of whole grains was cold breakfast cereals rather than dark bread (29 vs. 12%), reflective of the higher amount of whole grains in breakfast cereals. We observed a significant inverse association between whole-grain intake and BMI, percent body fat, and percent trunk fat mass regardless of the classification, although the trend was stronger with the quantitative estimate of whole-grain intake.

Only 1 prospective study (25) considered whether the physiological consequences of a high-fiber diet were dependent on the food sources. Koh-Banerjee et al. (25) found that both fruit and cereal fiber were inversely related in a dose-dependent manner to weight gain in middle-age men. Notably, in cross-sectional studies, only fiber from cereal sources was significantly associated with blood levels of different adipokines, such as C-reactive protein (32), tumor necrosis factor-α-receptor-2 (32), and adiponectin (33,34). In this study, no relationship was observed between total dietary fiber, fruit and vegetable fiber, and measures of body fat.

As far as we are aware, only 1 cross-sectional study, conducted in Australian adults, observed that a high-dietary fiber intake was inversely associated with fat mass derived from DXA (35). Both dietary fiber and higher intakes of whole grains have been linked to lower (9,10), or less gains in (36,37), abdominal adiposity as defined by waist circumference measures. In the present study, although dietary fiber intake was derived from multiple food sources, a significant association was observed only between cereal fiber and the measures of body fat distribution. There is some scientific evidence, albeit limited, to suggest that higher intakes of cereal fiber may modify body fat distribution. In a recent 12-wk hypocaloric intervention study, free-living obese adults with metabolic syndrome were advised either to avoid whole-grain foods or consume all their recommended grain servings from whole grains (38). Despite similar changes in body weight, waist circumference and percentage body fat (based on DXA), a significantly greater loss in percentage abdominal fat, was observed only in the whole-grain intervention arm. In an intervention study designed to elicit a low postprandial insulin response, consumption of a rye-pasta diet for 12 wk caused a 21% decrease in adipocyte size, even in the absence of changes in body weight (39). The findings of these intervention studies suggest that dietary modification may alter body fat or fat distribution independent of changes in overall body weight. The mechanism whereby cereal fiber, as opposed to other fiber sources, influences body fat distribution remains unknown and the biological mechanism remains speculative. It is possible that the influence of cereal fiber intake on insulin secretion or stimulation of incretin hormones, such as glucagon-like-peptide-1, favors reduced adiposity. Another possible explanation is that high-fiber diets may influence hypothalamic-pituitary-adrenal activity and such alterations in the hypothalamic-pituitary-adrenal axis have been associated with central body fat accumulation (40), although the cause and effect remains unclear.

It is important to recognize the limitations of this study. First, the cross-sectional nature of the present study limits our ability to infer causality between dietary exposures and body fat. Second, the study participants and findings of this study may not be representative of the general population of adults aged ≥60 y due to the exclusion criteria of the parent study. For instance, mean dietary fiber intake in this sample was ∼19 g/d compared with an estimated 13 g/d in a sample (n = 1123) of adults aged 60–80 y participating in NHANES (41), perhaps reflective of a healthier sample. Third, dietary measurement error may have distorted the observed associations between fiber derived from fruit and vegetable intake. Individuals tend to overestimate fruit and vegetable consumption using FFQ (26). Fourth, the DXA provides only a proxy measure of abdominal adiposity and is unable to differentiate subcutaneous fat from visceral fat. However, studies have reported a strong correlation between abdominal fat estimated using the specific ROI from DXA and visceral fat measures by MRI (28) and computed tomography (20,42). Finally, a recognized limitation of all observational studies is that higher whole-grain intake is related to other healthy lifestyle behaviors, such as greater physical activity and less smoking. Although we adjusted for important potential confounding factors, residual confounding caused by lifestyle factors associated with adiposity may still bias the observed association.

Based on current dietary recommendations, individuals >50 y of age should aim to consume between 21 and 30 g/d of dietary fiber and at least one-half of their grain servings should be from whole-grain foods (43). The Modified MyPyramid for Older Adults recommends achieving dietary fiber intake by choosing whole grains, whole fruits and vegetables, and legumes (44) and emphasis should be placed on consuming a variety of foods. The findings of this study suggest that cereal fiber, in particular from whole-grain products, may have an affect on body fat distribution. Further intervention studies are needed to consider how whole-grain foods (rich in cereal fiber) affect the regulation of energy intake and subsequently how different types of whole grains (and sources of cereal fiber) affect body fat distribution.