Download Fiber from Whole Grains, but not Refined Grains

Fiber from Whole Grains, but not Refined Grains, Is
Inversely Associated with All-Cause Mortality in Older
Women: The Iowa Women’s Health Study
David R. Jacobs, Jr, PhD, Mark A. Pereira, PhD, Katie A. Meyer, MPH, and Lawrence H. Kushi, ScD
Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, Minnesota (D.R.J., M.A.P., L.H.K.),
Department of Epidemiology, School of Public Health, Harvard University, Cambridge, Massachusetts (K.A.M.)
Key words: diet, epidemiology, prospective study, whole grain, fiber, mortality, Iowa Women’s Health Study
Background: Inconsistencies in epidemiologic findings relating grain fiber to chronic disease may be
explained by differentiating nutrient-rich fiber derived from whole grain vs. nutrient-poor fiber derived from
refined grain.
Objective: Given that phytochemicals are most varied and abundant in the outer layers of grains, we tested
the hypothesis that whole grain fiber consumption is associated with a reduced mortality risk in comparison to
a similar amount of refined grain fiber.
Design: 11040 postmenopausal women enrolled in the Iowa Women’s Health Study, matched on total grain
fiber intake, but differing in the proportion of fiber consumed from whole vs. refined grain, were followed from
baseline in 1986 through 31 December, 1997, during which time 1341 deaths occurred in 124,823 observed
woman-years.
Results: After multivariate adjustment in proportional hazards regression, women who consumed on average
1.9 g refined grain fiber/2000 kcal and 4.7 g whole grain fiber/2000 kcal had a 17% lower mortality rate
(RR⫽0.83, 95% CI⫽0.73– 0.94) than women who consumed predominantly refined grain fiber: 4.5 g/2000 kcal,
but only 1.3 g whole grain fiber/2000 kcal.
Conclusion: Inferences from studies that have reported associations between grain fiber intake and morbidity or mortality may be limited by not differentiating fiber sources. Future studies should distinguish fiber
from whole vs. refined grains. Public health policy should differentiate whole grains from refined, and
recommend increased consumption of the former.
INTRODUCTION
fiber may be consumed when a large enough quantity of refined
grain is eaten. Furthermore, case-control reports of the association of grain fiber with cancer are inconsistent [5–10]. Therefore, it was hypothesized that the nutrients eaten with fiber
(e.g., antioxidants and minerals) play a role in protection
against chronic disease and that results of studies of grain fiber
intake would depend on the mix of fiber from whole grain vs.
refined grain [4,11].
Whole grain fiber is nutrient-rich [11] and confers health
benefits that may result in increased longevity, while refined
grain fiber is nutrient-poor [11] and offers no such protection
when all other factors are held constant [2]. We studied this
hypothesis, as described below, by comparing the risk of total
mortality between postmenopausal women who reported eating
Twenty-five years ago Burkitt and colleagues [1] hypothesized that the rare occurrence in Africa of “Western” diseases
(e.g., coronary heart disease and colon cancer) was partly due
to a diet characterized by high fiber, minimally processed plant
foods. This hypothesis has often been used to promote the idea
that high fiber diets may help to prevent chronic diseases and
perhaps extend life. Findings [2,3] of reduced risk of cancer,
coronary heart disease and total mortality in people who have
high intake of whole grain, a high fiber food, support the
Burkitt hypothesis. However, an apparently increased cancer
risk associated with refined grain intake has been noted [4] in
case-control studies, even though a fairly large amount of grain
Mark A. Pereira, Ph.D., is now at the Harvard Medical School, Children’s Hospital, Division of Endocrinology, Boston, Massachusetts.
Lawrence H. Kushi, Sc.D., is now at Teachers College, Columbia University, New York, New York.
Address reprint requests to: David R. Jacobs, Jr., Ph.D., Division of Epidemiology, School of Public Health, University of Minnesota, 1300 South Second St., Suite 300,
Minneapolis, MN, 55454.
Journal of the American College of Nutrition, Vol. 19, No. 3, 326S–330S (2000)
Published by the American College of Nutrition
326S
Grain Fiber and Mortality in Women
equivalent amounts of grain fiber, in one case derived primarily
from refined grain foods, in the other primarily from foods
made with whole grain.
METHODS
Subjects
The Iowa Women’s Health Study is a prospective cohort
study of postmenopausal women living in the state of Iowa. In
January, 1986, a random sample of 99,826 women between the
ages of 55 and 69 years with a valid Iowa driver’s license were
mailed a 16-page questionnaire and invited to participate. The
41,836 women who returned the baseline questionnaire compose the study cohort. Responders were slightly leaner, older
and more likely to live in rural, less affluent counties than were
non-respondents [12]. Women were excluded from these analyses if they reported implausibly high (⬎5000 kcal) or low
(⬍600 kcal) energy intake (n⫽538) or left 30 or more items
blank on the food frequency questionnaire (n⫽2,782), leaving
38,740 women. Of these, 11,040 were included in the matched
design, described below under Statistical Analysis.
Data Collection
r⫽0.24. A study of nurses using a similar questionnaire found
high correlations with 28 days of food records for pertinent
food items: 0.75 for cold breakfast cereals, 0.61 for white
bread, and 0.66 for dark bread [17].
The exposure of interest for the current analyses was dietary
fiber from foods that were primarily composed of grains. The
total grain food group intake was subdivided into refined and
whole grains as previously outlined [2,3]. Table 1 lists the
whole and refined grain foods that were queried, with their
respective assumed values for dietary fiber.
Case Ascertainment
Vital status of cohort members was determined through
December 31, 1997, via annual linkage with the State Health
Registry of Iowa, the National Death Index and follow-up
questionnaires mailed in 1987, 1989, 1992 and 1997. During
followup, 4761 of the women died. Of these, 983 deaths were
attributed to coronary heart disease (ICD-9 codes 410 – 414 and
429.2) and 2049 to cancer(ICD-9 codes 140 –239).
Table 1. Whole and Refined Grain Foodsa and their Fiber
Assignments
The baseline questionnaire and measurements were designed to examine factors that may be involved in chronic
disease etiology, including age, body mass index (BMI), waistto-hip ratio (WHR), physical activity, and cigarette smoking.
BMI was calculated as kg/m2 from self-measured weight and
height. Waist to hip ratio was calculated as the average of two
measurements taken by a friend or spouse of participants using
a paper tape measure provided with the questionnaire [13].
Questions on the frequency of moderate and vigorous leisuretime activity [14] were combined to create a three-level physical activity score. Pack-years of smoking were calculated from
information on the intensity and duration of cigarette smoking.
Other participant characteristics, such as age at first live birth or
prevalent hypertension, diabetes, heart disease or cancer were
self-reported.
Food
Whole Grain
Dark bread
Whole grain breakfast cerealb
Popcorn
Cooked oatmeal
Wheat germ
Brown rice
Bran
Other grains (e.g., bulgar, kasha)
Refined Grain
Sweets/desserts
Cookies
Brownies
Donuts
Home baked cakes
Ready make cakes
Pie
White bread, including pita bread
Pasta
English muffins, bagels, or rolls
Refined grain breakfast cerealb
White rice
Muffins or biscuits
Pancakes or waffles
Pizza
Dietary Assessment
A 127-item food frequency questionnaire similar to that
used in the 1984 Nurses’ Health Study assessed habitual dietary
intake over the past year [15]. The validity of the food frequency questionnaire was evaluated in this cohort by comparing nutrient values determined from the questionnaire to values
estimated from the average of five 24-hour dietary recall surveys in 44 participants [16]. The intake of crude fiber was
highly reproducible, r⫽0.8 between two questionnaires administered six months apart; agreement with the average crude
fiber intake reported on five 24-hour recalls administered four
months before the last food frequency questionnaire was lower,
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
Fiber
Assignment
(g/
serving)
1.1
ⱖ3.0
1.1
4.0
0.9
3.3
1.5
3.2
0.6
0.7
0.8
0.5
1.1
2.2
0.6
2.2
1.3
ⱕ2.5
1.0
2.0
2.0
5.2
a
Foods are listed in order of frequency of intake in the cohort.
The brand name or generic cereal specified was coded for whole grain content;
breakfast cereals with a whole grain or bran content of 25% or more by weight
were classified as whole grain.
b
327S
Grain Fiber and Mortality in Women
Statistical Analysis
A matched design was used to assess the hypothesis that, for
constant total grain fiber intake, mortality was lower in women
whose predominant grain fiber source was whole grain foods
than in women whose predominant grain fiber source was
refined grain foods. We compared two groups of women,
matched to consume approximately 6 g/day of total grain fiber
(Table 2). The high refined grain fiber group, who on average
consumed 77% of their grain fiber from refined grain sources,
consisted of 3559 women who reported consuming ⬍3.6
g/2000 kcal of whole grain fiber and ⱖ3.6 g/2000 kcal of
refined grain fiber. The high whole grain fiber group, who
consumed on average 71% of their grain fiber from whole grain
sources, consisted of 7481 women who reported consuming
between 3.6 and 6 g/2000 kcal of whole grain fiber and ⬍3.6
g/2000 kcal of refined grain fiber. Women consuming higher
levels of whole grain fiber were excluded from this group
because there were very few matching women consuming more
than 6 g/2000 kcal from refined grain. The difference in death
rates between the matched groups became larger, the higher the
refined grain fiber cutpoint used to define the matched groups.
The mortality in the high whole grain fiber group relative to
the high refined grain fiber group was studied using proportional hazards regression models which also contained the
independent variables age and energy intake. Because whole
grain intake is strongly associated with a healthier lifestyle [2],
a multivariate adjusted model added the covariates listed in the
footnote to Table 4.
RESULTS
The women who derived 71% of their total grain fiber from
whole grain were older, consumed less total fat and carbohydrate, more total dietary fiber, more fruits and vegetables, red
meat, fish and seafood, and alcohol, were less likely to smoke,
Table 2. Mean ⫾ Standard Deviation of Grain Fiber Intake
(g/2000 kcal) in Iowa Women Aged 55 to 69 in 1986 in
Two Groups Matched on Eating Approximately 6 g/2000
kcal per Day of Total Grain Fiber Intake
Whole grain fiber Refined grain fiber
71% of total grain 77% of total grain
fiber intake*
fiber intake**
Sample size
Whole grain fiber intake
Refined grain fiber intake
Total grain fiber intake
(group matching
criterion)
7481
4.7 ⫾ 0.07
1.9 ⫾ 0.8
3559
1.3 ⫾ 1.0
4.5 ⫾ 0.9
6.6 ⫾ 1.0
5.8 ⫾ 1.2
* 3.6-6 g/2000 kcal of whole grain fiber and ⬍3.6 g/2000 kcal of refined grain
fiber.
** ⬍3.6 g/2000 kcal of whole grain fiber and ⱖ3.6 g/2000 kcal of refined grain
fiber.
328S
Table 3. Baseline Means or Percentages in Iowa Women
Aged 55 to 69 in 1986 in Two Groups Matched on Eating
Approximately 6 g/2000 kcal per Day of Total Grain Fiber
Intake
Number of women
Age (years)
Energy (kcal/day)
Total fat (% kcal)
Saturated fat (% kcal)
Carbohydrate (% kcal)
Total dietary fiber intake
(g/2000 kcal)
Fruit and vegetables
(servings/week)
Red meat (servings/week)
Fish & seafood (servings/
week)
Alcohol intake (g/day)
Cigarette smoking (packyears)
(% ever smoked)
BMI (kg/m2)
Waist-hip ratio
Age at first live birth (years)
Physical activity (% highly
active)
Education (%⬎ high school)
Hormone replacement
therapy (% users)
Vitamin supplement use (%)
Currently married (%)
Prevalence of self-reported
disease (%)
Hypertension
Diabetes
Heart disease
Cancer
Whole grain fiber
71% of total grain
fiber intake*
Refined grain fiber
77% of total grain
fiber intake**
7481
61.8
1838
33.4
11.6
49.3
23.3
3559
61.4†
1847
34.5†
11.6
50.2†
19.2†
47
35†
5.8
1.9
5.3†
1.3†
3.4
8.1
2.3†
11.0†
31
27.0
0.834
20.7
27
37†
27.2
0.846†
20.3
18†
43
41
32†
35†
36
77
27†
74†
36
7
9
10
37
6
10
9
* 3.6-6 g/2000 kcal of whole grain fiber and ⬍3.6 g/2000 kcal of refined grain
fiber.
** ⬍3.6 g/2000 kcal of whole grain fiber and ⱖ3.6 g/2000 kcal of refined grain
fiber.
† p ⬍ 0.001.
Age is unadjusted, energy intake is age adjusted, and all other variables are age
and energy intake adjusted; adjustment is by linear regression.
had a smaller waist to hip ratio, were more physically active,
had more education, were more likely taking hormone replacement therapy and vitamin supplements, and were more likely to
be married. Most of these differences were small between the
groups matched on total grain fiber intake (Table 3).
There were 1341 total deaths among the 11,040 women
studied (124,823 woman-years). Among these, 1240 deaths in
10,444 women had complete data for multivariate analysis.
After multivariate adjustment in proportional hazards regression, the women who consumed on average 1.9 g refined grain
fiber/2000 kcal and 4.7 g whole grain fiber/2000 kcal had a
VOL. 19, NO. 3
Grain Fiber and Mortality in Women
Table 4. Proportional Hazards Regression Analysis of
Mortality in Iowa Women Matched on Total Grain Intake in
1986 and Followed from 1986 to 1997
Total mortality
Coronary heart disease
mortality
Other cardiovascular disease
mortality
All cancer mortality
All other mortality
Hazard Rate
Ratio adjusted
for age and
energy intake
Hazard Rate
Ratio multivariate
adjusted*
0.75 (0.67,0.84)
0.82 (0.64,1.05)
0.83 (0.73,0.94)
0.89 (0.66,1.20)
0.72 (0.54,0.96)
0.75 (0.54,1.06)
0.83 (0.70,0.99)
0.74 (0.64,0.84)
0.92 (0.75,1.12)
0.83 (0.71,0.97)
* Adjusted for age, energy intake, educational status, marital status, hypertension,
diabetes, heart disease, cancer, body mass index, waist to hip ratio, age at first
birth, physical activity score, ever smoker, pack years of cigarette smoking,
alcohol intake, vitamin supplement use, hormone replacement therapy use, percent of energy from total fat, saturated fat, and carbohydrates, intake per 1000
kcal of fiber from nongrain sources, intake of red meat, and intake of fish and
seafood.
17% lower mortality rate (RR⫽0.83, 95% CI⫽0.73– 0.94) than
the women who consumed predominantly refined grain fiber:
4.5 g/2000 kcal, but only 1.3 g whole grain fiber/2000 kcal
(Table 4).
These findings reflected, in part, reductions among 274
coronary heart disease deaths (247 in multivariate analyses),
190 other cardiovascular deaths (180 in multivariate analyses),
566 cancer deaths (527 in multivariate analyses), and 877 other
deaths (813 in multivariate analyses). Compared to the high
refined grain fiber consumers after adjustment for the demographic, behavioral and physiologic characteristics, the high
whole grain fiber women suffered less of each cause of death,
statistically significant for non-cardiovascular, non-cancer mortality (Table 4). Women who were excluded from the matched
design because they consumed more than 6 g/2000 kcal of
whole grain fiber, generally accompanied by small amounts of
refined grain fiber, had statistically significantly reduced total,
coronary heart disease, and all cancer mortality rates, compared
to women who ate little whole grain fiber (data not shown).
COMMENT
In these data, a similar amount of total grain fiber had
dissimilar associations with total mortality, depending on
whether the fiber came from foods that contained primarily
whole grain or refined grain. The fiber only from whole grain
was associated with reduced risk of mortality. These observations imply that consumption of the plant constituents that are
botanically linked to fiber may confer important health benefits
above and beyond effects of the fiber itself. Among these
constituents, found abundantly in the bran and germ of whole
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
grains, are minerals, vitamins, phenolic compounds and phytoestrogens [11]. Although individuals who consume large
quantities of refined grain may consume relatively high
amounts of grain fiber, this fiber derived from cell walls in the
starchy endosperm is nutrient-poor, as it does not include the
biochemicals found in the nutrient-rich bran and germ. This is
not to say that dietary fiber itself is not important, for instance
in modulating the glycemic index [18,19], reducing gastric
transit time and lowering serum cholesterol [20 –23].
Jacobs et al. [4,24] reviewed case-control studies of cancers
and intake of whole grain foods, finding generally reduced risk
associated with higher whole grain intake. Follow-up analyses
in the Iowa Women’s Health Study [2,3] broadened the findings of the reviews on whole grain and cancer. In the Iowa
women studied here, followed for nine years, total, cancer, and
CHD mortality were reduced in those who reported habitual
consumption of whole grain foods. In contrast, these causes of
death tended to be positively associated with refined grain
intake [3].
One interpretation of the findings of these earlier studies is
that whole grain protects against chronic diseases because it is
a “high fiber” food. However, the findings of the current
analyses imply that this may be an oversimplification, because
refined grain also contains grain fiber, although in smaller
concentrations than does whole grain. If enough refined grain
foods are eaten, a fairly large amount of grain fiber can be
consumed, albeit with several times the amount of starch in
comparison to whole grain foods. Taken together, the findings
of previous studies and of this matched analysis suggest that the
potential health effects of grain fiber may depend on its source.
The group-matched design used here did not work perfectly,
in that the mean level of total grain fiber differed between the
matched groups by 0.8 g/2000 kcal. Perfect matching was not
possible because of the inverse correlation between whole grain
fiber intake and refined grain fiber intake (r⫽⫺0.24). Findings
were robust to modifications in cutpoints defining the matched
groups; the estimate of mortality difference between matched
groups was enhanced, but the number of subjects in the higher
refined grain fiber group dropped rapidly, the higher the cutpoint chosen for higher vs. lower refined grain fiber (data not
shown). We also considered a proportional hazards regression
strategy in the whole cohort, with the two fiber terms included
as continuous variables. However, we determined that the full
cohort analysis strategy was more effective for whole grain
fiber than for refined grain fiber. The whole grain fiber proportional hazards regression coefficient was inverse in all analyses, but the refined grain fiber regression coefficient varied
greatly and had a large standard error. The limited range of
intake of refined grain fiber intake led to a substantially less
reliable estimation of the association of mortality with refined
grain fiber than with whole grain fiber. This problem of limited
range of refined grain fiber intake was circumvented in the
matched design. The problem of limited range of intake may
329S
Grain Fiber and Mortality in Women
have contributed to inconsistent findings concerning total grain
fiber in populations that consume little whole grain [5–10,25].
Future studies should attempt to separate grain fiber-containing foods into those that are nutrient-rich (whole grains)
and those that are nutrient-poor (refined grains). Although
further basic research, including feeding studies in animals and
humans, on the effects of nutrients and constituents associated
with fiber is needed, as recently suggested by Jacobs et al.,
[2,3] and Willett [26], public health recommendations should
more clearly differentiate between carbohydrates that are nutrient-rich (e.g., at least part whole grain breads and breakfast
cereals) and those that nutrient-poor, due to refining or other
processing.
REFERENCES
1. Burkitt DP, Walker AR, Painter NS: Dietary fiber and disease.
JAMA 229:1068–1074, 1974.
2. Jacobs Jr DJ, Meyer KA, Kushi LH, Folsom AR: Whole-grain
intake may reduce the risk of ischemic heart disease death in
postmenopausal women: The Iowa Women’s Health Study. Am J
Clin Nutr 68:248–257, 1998.
3. Jacobs Jr DJ, Meyer KA, Kushi LH, Folsom AR: Is whole grain
intake associated with reduced total and cause specific death rates
in older women? The Iowa Women’s Health Study. Am J Public
Health 89:322–329, 1999.
4. Jacobs Jr DR, Slavin J, Marquart M: Whole grain intake and
cancer: a review of the literature. Nutr Cancer 24:221–229, 1995.
5. Willett WC, Stampfer MJ, Colditz GA, Rosner BA, Speizer FE:
Relation of meat, fat and fiber intake to the risk of colon cancer in
a prospective study among women. N Engl J Med 323:1664–1672,
1990.
6. Freudenheim JL, Graham S, Marshall JR, Haughey BP, Wilkinson
G: A case-control study of diet and rectal cancer in western New
York. Am J Epidemiol 131:612–624, 1990.
7. Potter JP, McMichael AJ: Diet and cancer of the colon and rectum:
a case-control study. J Nat Cancer Inst 76:557–569, 1986.
8. McKeown-Eyssen GE, Bright-See E: Dietary factors in colon
cancer: international relationships. Nutr Cancer 6:160–170, 1984.
9. Meyer F, White E: Alcohol and nutrients in relation to colon
cancer in middle-aged adults. Am J Epidemiol 138:225–236, 1993.
10. Pryor M, Slattery ML, Robinson LM, Egger M: Adolescent diet
and breast cancer in Utah. Cancer Res 49:2161–2167, 1989.
11. Slavin J, Jacobs DR, Jr., Marquart L: Whole-grain consumption
and chronic disease: protective mechanisms. Nutr Cancer 27:14–
21, 1997.
12. Folsom AR, Prineas RJ, Kay SA, Soler JT: Body fat distribution
and self-reported prevalence of hypertension, heart disease, and
330S
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
other heart disease in older women. Int J Epidemiol 18:361–367,
1989.
Kushi LH, Kaye SA, Folsom AR, Soler JT, Prineas RJ: Accuracy
and reliability of self-measurement of body girths. Am J Epidemiol
128:740–748, 1988.
Kushi LH, Fee RM, Folsom AR, Mink PJ, Andersen KE, Sellers
TA. Physical activity and mortality in postmenopausal women.
JAMA 277:1287–1292, 1997.
Willett WC, Sampson L, Browne MI, Stampfer MJ, Rosner B,
Hennekens CH, Speizer FE: The use of a self-administered questionnaire to assess diet four years in the past. Am J Epidemiol
127:188–199, 1988.
Munger RG, Folsom AR, Kushi LH, Kaye SA, Sellers TA: Dietary
assessment of older Iowa women with a food frequency questionnaire: nutrient intake, reproducibility, and comparison with 24hour dietary recall interviews. Am J Epidemiol 136:192–200,
1992.
Salvini S, Hunter DJ, Sampson L, Stampfer MJ, Colditz GA,
Rosner B, Willett WC: Food-based validation of a dietary questionnaire: the effects of week-to-week variation in food consumption. Int J Epidemiol 18:858–867, 1989.
Jenkins DJ, Jenkins AL: Dietary fiber and the glycemic response.
Proc Soc Exp Biol Med 180:422–431, 1985.
Wolever TMS, Jenkins DJA, Vuksan V, Jenkins AL, Buckley GC,
Wong GS, Josse RG: Beneficial effect of a low glycaemic index
diet in Type 2 diabetes. Diabet Med 9:451–458, 1992.
McIntyre A, Gibson PR, Young GP: Butyrate production from
dietary fiber and protection against large bowel cancer in a rat
model. Gut 34:386–391, 1993.
Cummings J, Bingham S, Heaton K, Eastwood M: Fecal weight,
colon cancer risk, and dietary intake of nonstarch polysaccharide
(dietary fiber). Gastroenterology 103:1783–1787, 1992.
Stephen AM: Whole grains—impact of consuming whole grains
on physiological effects of dietary fiber and starch. Crit Rev Food
Sci Nutr 34:499–511, 1994.
Ripsin CM, Keenan JM, Jacobs DR, Elmer PJ, Welch RR, Van
Horn L, Liu K, Turnbull WH, Thye FW, Kestin M: Oat products
and lipid lowering: a meta-analysis. JAMA 267:3317–3325, 1992.
Jacobs Jr DR, Marquart M, Slavin J, Kushi LH: Whole grain intake
and cancer: an expanded review and meta-analysis. Nutr Cancer
30:85–96, 1998.
Fuchs CS, Giovannucci EL, Colditz GA, Hunter DJ, Stampfer MJ,
Rosner B, Speizer FE, Willett WC: Dietary fiber and the risk of
colorectal cancer and adenoma in women. New Eng J Med 340:
169–176, 1999.
Willett WC: The dietary pyramid: does the foundation need repair?
Am J Clin Nutr 68:2128–2129, 1998.
Received February 2000.
VOL. 19, NO. 3