Beneficial Effects of High Dietary Fiber Intake in Patients with Type 2 Diabetes Mellitus

Abstract

The effect of increasing the intake of dietary fiber on glycemic control in patients with type 2 diabetes mellitus is controversial.
In a randomized, crossover study, we assigned 13 patients with type 2 diabetes mellitus to follow two diets, each for six weeks: a diet containing moderate amounts of fiber (total, 24 g; 8 g of soluble fiber and 16 g of insoluble fiber), as recommended by the American Diabetes Association (ADA), and a high-fiber diet (total, 50 g; 25 g of soluble fiber and 25 g of insoluble fiber), containing foods not fortified with fiber (unfortified foods). Both diets, prepared in a research kitchen, had the same macronutrient and energy content. We compared the effects of the two diets on glycemic control and plasma lipid concentrations.
Compliance with the diets was excellent. During the sixth week, the high-fiber diet, as compared with the the sixth week of the ADA diet, mean daily preprandial plasma glucose concentrations were 13 mg per deciliter [0.7 mmol per liter] lower (95 percent confidence interval, 1 to 24 mg per deciliter [0.1 to 1.3 mmol per liter]; P=0.04) and mean median difference, daily urinary glucose excretion 1.3 g (0.23; 95 percent confidence interval, 0.03 to 1.83 g; P= 0.008). The high-fiber diet also lowered the area under the curve for 24-hour plasma glucose and insulin concentrations, which were measured every two hours, by 10 percent (P=0.02) and 12 percent (P=0.05), respectively. The high-fiber diet reduced plasma total cholesterol concentrations by 6.7 percent (P=0.02), triglyceride concentrations by 10.2 percent (P=0.02), and very-low-density lipoprotein cholesterol concentrations by 12.5 percent (P=0.01).
A high intake of dietary fiber, particularly of the soluble type, above the level recommended by the ADA, improves glycemic control, decreases hyperinsulinemia, and lowers plasma lipid concentrations in patients with type 2 diabetes.

Full text

1392
·
May 11, 2000
The New England Journal of Medicine
BENEFICIAL EFFECTS OF HIGH DIETARY FIBER INTAKE IN PATIENTS
WITH TYPE 2 DIABETES MELLITUS
M
ANISHA
C
HANDALIA
, M.D., A
BHIMANYU
G
ARG
, M.D., D
IETER
L
UTJOHANN
, P
H
.D., K
LAUS
VON
B
ERGMANN
, M.D.,
S
COTT
M. G
RUNDY
, M.D., P
H
.D.,
AND
L
INDA
J. B
RINKLEY
, R.D.
A
BSTRACT
Background
The effect of increasing the intake of
dietary fiber on glycemic control in patients with type
2 diabetes mellitus is controversial.
Methods
In a randomized, crossover study, we as-
signed 13 patients with type 2 diabetes mellitus to
follow two diets, each for six weeks: a diet containing
moderate amounts of fiber (total, 24 g; 8 g of soluble
fiber and 16 g of insoluble fiber), as recommended by
the American Diabetes Association (ADA), and a high-
fiber diet (total, 50 g; 25 g of soluble fiber and 25 g
of insoluble fiber) containing foods not fortified with
fiber (unfortified foods). Both diets, prepared in a re-
search kitchen, had the same macronutrient and en-
ergy content. We compared the effects of the two diets
on glycemic control and plasma lipid concentrations.
Results
Compliance with the diets was excellent.
During the sixth week of the high-fiber diet, as com-
pared with the sixth week of the ADA diet, mean dai-
ly preprandial plasma glucose concentrations were 13
mg per deciliter (0.7 mmol per liter) lower (95 per-
cent confidence interval, 1 to 24 mg per deciliter [0.1
to 1.3 mmol per liter]; P=0.04) and mean daily uri-
nary glucose excretion was 1.3 g lower (median dif-
ference, 0.23 g; 95 percent confidence interval, 0.03
to 1.83; P=0.008). The high-fiber diet also lowered the
area under the curve for 24-hour plasma glucose and
insulin concentrations, which were measured every
two hours, by 10 percent (P=0.02) and 12 percent
(P=0.05), respectively. The high-fiber diet reduced
plasma total cholesterol concentrations by 6.7 per-
cent (P=0.02), triglyceride concentrations by 10.2
percent (P=0.02), and very-low-density lipoprotein
cholesterol concentrations by 12.5 percent (P=0.01).
Conclusions
A high intake of dietary fiber, partic-
ularly of the soluble type, above the level recommend-
ed by the ADA, improves glycemic control, decreases
hyperinsulinemia, and lowers plasma lipid concen-
trations in patients with type 2 diabetes. (N Engl J
Med 2000;342:1392-8.)
©2000, Massachusetts Medical Society.
From the Department of Internal Medicine (M.C., A.G., S.M.G., L.J.B.)
and the Center for Human Nutrition (A.G., S.M.G.), University of Texas
Southwestern Medical Center, Dallas; the Department of Veterans Affairs
Medical Center, Dallas (M.C., A.G., S.M.G.); and the Department of Clin-
ical Pharmacology, Rheinische Friedrich-Wilhelms-Universität, Bonn, Ger-
many (D.L., K.B.). Address reprint requests to Dr. Garg at the Center for
Human Nutrition, University of Texas Southwestern Medical Center, 5323
Harry Hines Blvd., Dallas, TX 75390.
IETARY guidelines for patients with dia-
betes mellitus were revised by the Amer-
ican Diabetes Association (ADA) earlier
this year.
1
The ADA recommends that the
composition of the diet be individualized on the basis
of a nutritional assessment and the outcomes desired.
Consistent with the previous recommendations of
the ADA,
2
the new guidelines advise replacing satu-
D
rated fat with carbohydrates. However, on the basis of
previous studies,
3-10
an alternative approach of replac-
ing saturated fat with cis monounsaturated fat was
also included in the recommendations.
1
This new ap-
proach is further supported by epidemiologic studies
that have shown the healthful effects of diets rich in cis
monounsaturated fat in Mediterranean countries.
11,12
Another, less strongly emphasized aspect of Med-
iterranean diets is the high intake of fruits, vegetables,
and grains that are rich sources of dietary fiber.
13,14
The ADA recommended a moderate increase in the
intake of dietary fiber to 20 to 35 g per day because
of the cholesterol-lowering effects of soluble fiber.
However, the effects of dietary fiber on glycemic con-
trol were considered inconsequential.
1
Furthermore,
the expert panel of the ADA considered it difficult
to achieve a high dietary intake of soluble fiber with-
out consuming foods or supplements fortified with
fiber.
1
We therefore designed the present study to de-
termine the effects on glycemic control and plasma
lipid concentrations of increasing the intake of die-
tary fiber in patients with type 2 diabetes exclusively
through the consumption of foods not fortified with
fiber (unfortified foods) to a level beyond that rec-
ommended by the ADA. In addition, we studied the
effects of such an intervention on the intestinal ab-
sorption of cholesterol and the fecal excretion of ster-
ols in an attempt to uncover the mechanisms by which
a high-fiber diet lowers plasma cholesterol.
METHODS
Patients
We studied 12 men and 1 woman (9 non-Hispanic whites and
4 blacks) with type 2 diabetes at the general clinical research center
of the University of Texas Southwestern Medical Center at Dallas.
The protocol for the study was approved by the institutional re-
view board of the medical center, and each patient gave written
informed consent. In all patients the onset of diabetes was insid-
ious; the disease developed in most of the patients after 40 years
of age. Their mean (±SD) age was 61±9 years (range, 45 to 70).
Their mean body weight was 93.5±12.7 kg, and the mean body-
mass index (the weight in kilograms divided by the square of the
height in meters) was 32.3±3.9. Three patients were treated with
diet alone, and the other 10 patients were treated with 2.5 to 20
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.

BENEFICIAL EFFECTS OF HIGH DIETARY FIBER INTAKE IN PATIENTS WITH TYPE 2 DIABETES MELLITUS
Volume 342 Number 19
·
1393
mg of glyburide daily in addition to diet. The dose of glyburide
was not changed during the study.
On entry into the study, the patients’ plasma cholesterol and tri-
glyceride concentrations after an overnight fast ranged from 151
to 324 mg per deciliter (3.90 to 8.38 mmol per liter) and 67 to
390 mg per deciliter (0.76 to 4.40 mmol per liter), respectively,
and their fasting plasma glucose concentrations were less than
200 mg per deciliter (11.1 mmol per liter). Their glycosylated he-
moglobin values ranged from 6.0 to 9.8 percent. Two patients
had a history of coronary heart disease, but none had recently had
a myocardial infarction, unstable angina pectoris, or congestive
heart failure. Also, none had thyroid, renal, or hepatic disease. None
of the patients were receiving lipid-lowering therapy.
Design of the Study
All the patients were first admitted to the general clinical re-
search center for five days (the base-line period), during which a
detailed history was taken, a physical examination was performed,
and laboratory tests were performed. After the base-line period, all
the patients received both the ADA diet and the high-fiber diet,
each diet for a period of six weeks. Six patients received the high-
fiber diet first, and the other seven received the ADA diet first.
There was a median interval of seven days between the two study
periods, during which the patients were instructed to consume an
isocaloric diet. During the last week of each dietary period (days
36 to 42), the patients were hospitalized for evaluation.
On weekdays, all the patients ate at least one meal (breakfast,
lunch, or dinner) at the general clinical research center. Other meals
were supplied in packages so that they could be consumed at home.
The dietitian monitored compliance by interviewing the patients.
The patients were instructed to bring back any unconsumed food
and to maintain a constant level of physical activity throughout the
study.
Blood for lipid analyses was drawn, after an overnight fast, daily
for two days before the institution of the study diet and daily on
days 38 through 42 during both dietary periods. Plasma glucose
was measured at 7 and 11 a.m. and at 4 and 8 p.m. each day dur-
ing the base-line period and on days 38 through 42 of both die-
tary periods. Glycosylated hemoglobin was measured during the
base-line period and at the end of each dietary period. On the
last day of each dietary period, blood samples were obtained ev-
ery two hours for measurements of plasma glucose and insulin.
On days 38 through 42, patients collected 24-hour urine speci-
mens for quantitative determination of glucose.
To permit us to determine fecal sterol balance and the percentage
of cholesterol absorption, each patient took a capsule containing
30 mg of sitostanol, 3 mg of [26,26, 26,27, 27,27-
2
H
6
]-cholesterol,
and 3 mg of [5,6,22,23-
2
H
4
]-sitostanol (Medical Isotopes, Pelham,
N.H.) three times a day on days 36 through 42. Fecal samples
were collected on day 35 or 36 and on the last three days of each
dietary period. Fecal samples were frozen within 12 hours after
collection and were pooled for analysis of small aliquots.
Diets
The composition of the study diets is shown in Table 1. The
composition of the diets was calculated by means of a software
program based on the Department of Agriculture Handbook Se-
ries 8 (Nutriplanner, Practocare, San Diego, Calif.).
15
The content
of total as well as soluble and insoluble dietary fiber was estimated
according to the data provided in the
CRC Handbook of Dietary
Fiber in Human Nutrition.
16
Both diets consisted of unfortified
foods. The patients were allowed some choices of food items.
Both diets provided 15 percent of the total energy as protein, 55
percent as carbohydrate, and 30 percent as fat; saturated, cis mono-
unsaturated, and polyunsaturated fats accounted for 7 percent, 17
percent, and 6 percent of the total energy, respectively.
The high-fiber diet provided 50 g of total fiber per day; soluble
and insoluble fiber content provided 25 g each. The ADA diet
contained 24 g of total fiber per day, with 8 g as soluble fiber and
16 g as insoluble fiber. Unfortified foods, particularly those rich
*ADA denotes American Diabetes Association.
T
ABLE
1.
C
OMPOSITION
OF
THE
S
TUDY
D
IETS
.
C
ONSTITUENT
ADA
D
IET
*
H
IGH
-F
IBER
D
IET
Carbohydrate (% of total energy) 55 55
Protein (% of total energy) 15 15
Fat (% of total energy)
Saturated
Cis monounsaturated
Polyunsaturated
30
7
17
6
30
7
17
6
Cholesterol (mg/day) 300 297
Fiber (g/day)
Total
Soluble
Insoluble
24
8
16
50
25
25
*Each menu provided 2308 kcal per day. ADA denotes American Dia-
betes Association.
T
ABLE
2.
S
AMPLE
M
ENUS
OF
THE
S
TUDY
D
IETS
.*
ADA D
IET
H
IGH
-F
IBER
D
IET
FOOD WEIGHT FOOD WEIGHT
grams grams
Breakfast
Orange juice 220 Orange sections 300
White grits 50 Oatmeal 50
Egg substitute 40 Scrambled egg 37
Olive oil 10 Olive oil 10
Decaffeinated coffee 2 Decaffeinated cof fee 2
Lunch
Ham (5% fat) 50 Ham (5% fat) 52
Mayonnaise 6 Mayonnaise 12
Iceberg lettuce 15 Iceberg lettuce 10
Fresh tomato 30 Fresh tomato 15
Low-sodium bread 60 Whole-wheat bread 60
Corn (canned) 140 Corn (canned) 40
Cider vinegar 5 Green peas (canned) 110
Dehydrated onion 2 Dehydrated onion 2
Olive oil 10 Olive oil 10
Fresh green pepper 10 Fresh green pepper 10
Fresh celery 15 Fresh celer y 15
Fruit cocktail (canned) 105 Fresh papaya 250
Instant tea 2 Instant tea 2
Oatmeal raisin cookie 20
Dinner
Chicken breast (skinned) 90 Chicken breast (skinned) 90
Bran flakes 10 Bran flakes 10
Low-sodium bread 20 Oat bran 5
Parmesan cheese 1 Parmesan cheese 1
Whole egg 1 Egg substitute 10
Tomato (canned) 120 Tomato (canned) 105
Low-fat cheese 11 Low-fat cheese 19
Spaghetti 45 Spaghetti 19
Green beans 75 Zucchini 195
Olive oil 17 Olive oil 19
Whole-wheat bread 21 Whole-wheat bread 30
Graham crackers 21 Fresh peaches 300
Instant tea 2 Instant tea 2
Bedtime snack
Mozzarella cheese 30 Fruit cocktail (canned) 200
Low-sodium bread 30 Cherries (canned) 100
Pineapple juice 190 Granola 15
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.

1394
·
May 11, 2000
The New England Journal of Medicine
in soluble fiber, such as cantaloupe, grapefruit, orange, papaya, rai-
sins, lima beans, okra, sweet potato, winter squash, zucchini, gra-
nola, oat bran, and oatmeal, were used to achieve high-fiber intake.
No fiber supplements were used. Sample menus of both the study
diets are shown in Table 2. The individual foods were weighed
daily during meal preparation in the research kitchen of the gen-
eral clinical research center.
Biochemical Analyses
Fasting plasma total cholesterol, lipoprotein cholesterol, and tri-
glycerides were measured according to the procedures of the Lip-
id Research Clinics.
17
Cholesterol and triglycerides were measured
enzymatically with the use of kits (Boehringer Mannheim, Indi-
anapolis). Very-low-density lipoproteins (VLDLs) (density, less
than 1.006 g per milliliter) were removed by ultracentrifugation,
and cholesterol was measured in the VLDL fraction and the in-
franatant. High-density-lipoprotein (HDL) cholesterol was meas-
ured enzymatically after lipoproteins containing apolipoprotein B
had been precipitated with heparin–manganese. Cholesterol in
the low-density lipoprotein (LDL) fraction was estimated to be
the difference between the cholesterol content of the infranatant
and that of the HDL fraction.
Plasma and urinary glucose were measured by the glucose ox-
idase method (Beckman Glucose Analyzer, Beckman Instruments,
Fullerton, Calif.). Glycosylated hemoglobin was measured with
ion-exchange high-performance liquid chromatography (Bio-Rad
Laboratories, Hercules, Calif.). Plasma insulin was measured by
radioimmunoassay.
18,19
Pooled fecal samples collected within the last week of each die-
tary period were prepared for analysis of neutral and acidic fecal
sterols as described previously.
20
Gas–liquid chromatography of
neutral and acidic fecal sterols was performed on a gas chromato-
graph (model HP5890, Hewlett–Packard, Palo Alto, Calif.)
equipped with an automatic sample injector. Cholesterol absorp-
tion was measured during the same period from fecal samples by
gas–liquid chromatography and mass spectrometry.
21
Statistical Analysis
To compare the two study periods and to assess the effect of
the sequence in which the patients received the high-fiber and
ADA diets, we used repeated-measures analysis of variance.
22
For
skewed data, we used the Wilcoxon signed-rank test to compare
the two dietary periods.
23
RESULTS
The compliance with both the study diets was ex-
cellent, according to interviews with the patients and
estimates of the energy content of any leftover foods.
Three patients reported consuming extra food on
one day during the study, two while eating the high-
fiber diet and one the ADA diet. The patients com-
mented about the larger quantities of food in the
high-fiber diet, but they consumed all the food giv-
en to them. The results are presented irrespective of
the order of the diets, because the sequence of the
diets had no effect on the results.
During the last week of each study period, the pa-
tients in both groups had similar daily energy intakes
and body weights and received a similar dose of gly-
buride (Table 3). The mean plasma glucose concen-
tration was lower (by 13 mg per deciliter [0.7 mmol
per liter], or 8.9 percent) when patients completed
the high-fiber diet than when they completed the
ADA diet (P=0.04), and mean daily urinary glucose
excretion was 1.3 g lower (P=0.008). Daily plasma
glucose concentrations were 10 percent lower with
the high-fiber diet than with the ADA diet (values
for the area under the curve, 3743±944 vs. 3365±
1003 mg·hour per deciliter [207.8±52.4 vs. 186.8±
55.7 mmol·hour per liter]; P=0.02), and plasma in-
sulin concentrations were 12 percent lower (values
for the area under the curve, 1107±650 vs. 971±
491 µU· hour per milliliter [6642±3900 vs. 5826±
2946 pmol·hour per liter]; P=0.05) (Fig. 1). Gly-
*Plus–minus values are means ±SD. ADA denotes American Diabetes Association, and CI confi-
dence interval.
†An analysis of variance was used to compare the two diets, except for urinary glucose, for which
the Wilcoxon signed-rank test was used.
‡The values are averages of plasma glucose concentrations measured at 7 and 11 a.m. and at 4 and
8 p.m. each day for five days during hospitalization. To convert values for plasma glucose to milli-
moles per liter, multiply by 0.056.
§The values are averages of five daily urine collections during hospitalization.
T
ABLE
3.
M
ETABOLIC
V
ARIABLES
DURING
THE
L
AST
W
EEK
OF
THE
S
TUDY
P
ERIODS
(D
AYS
38
THROUGH
42).*
V
ARIABLE
ADA D
IET
H
IGH
-F
IBER
D
IET
D
IFFERENCE
BETWEEN
D
IETS
(95% CI) P V
ALUE
†
Energy intake (kcal/day) 2308±236 2308±236 — 1.00
Weight (kg) 90.7±13.3 90.5±12.7 ¡0.2 (¡1.1 to 0.6) 0.60
Dose of glyburide (mg/day) 10.0±8.7 10.0±8.7 — 1.00
Plasma glucose (mg/deciliter)‡ 142±36 130±38 ¡13 (¡24 to ¡1) 0.04
Urinary glucose (g/day)
Mean
Median§
2.3±4.3
0.76
1.0±1.9
0.0
—
¡0.23 (¡1.83 to ¡0.03)
—
0.008
Glycosylated hemoglobin (%) 7.2±1.3 6.9±1.2 ¡0.3 (¡0.6 to 0.1) 0.09
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.

BENEFICIAL EFFECTS OF HIGH DIETARY FIBER INTAKE IN PATIENTS WITH TYPE 2 DIABETES MELLITUS
Volume 342 Number 19
·
1395
cosylated hemoglobin values were slightly lower af-
ter the high-fiber diet (P=0.09).
As compared with the ADA diet, the high-fiber
diet resulted in a lower fasting plasma total cholester-
ol concentration (by 6.7 percent, P=0.02), a lower
plasma triglyceride concentration (by 10.2 percent,
P=0.02), and a lower plasma VLDL cholesterol con-
centration (by 12.5 percent, P=0.01) (Table 4). The
fasting plasma LDL cholesterol concentration was
6.3 percent lower with the high-fiber diet (P=0.11).
There were no significant differences between the
two diets in terms of the fasting plasma HDL cho-
lesterol concentration.
As compared with the ADA diet, the high-fiber
diet decreased gastrointestinal absorption of choles-
terol by 10 percent (48.5±9.6 vs. 43.7±7.4 percent;
95 percent confidence interval for the decrease, 0.6
to 9.0 percent; P=0.03) and increased fecal acidic
sterol excretion by 41 percent (895±301 vs. 1258±
458 mg per day; 95 percent confidence interval for
the increase, 137 to 589 mg per day; P=0.005), but
did not significantly affect the excretion of neutral
Figure 1.
Mean (±SE) 24-Hour Profile of Plasma Glucose Concentrations (Panel A) and Insulin Concen-
trations (Panel B) during the Last Day of the American Diabetes Association (ADA) Diet and the Last
Day of the High-Fiber Diet in 13 Patients with Type 2 Diabetes Mellitus.
The arrows indicate the times at which the main meals and a snack were consumed during the day.
To convert values for glucose to millimoles per liter, multiply by 0.056. To convert values for insulin to
picomoles per liter, multiply by 6.
0
100
7 a.m.
7 a.m.
20
40
60
80
9 a.m.
11 a.m.
1 p.m.
3 p.m.
5 p.m.
7 p.m.
9 p.m.
11 p.m.
1 a.m.
3 a.m.
5 a.m.
Time
B
High-fiber diet
ADA diet
Plasma Insulin (mU/ml)
100
240
7 a.m.
7 a.m.
120
140
160
180
200
220
9 a.m.
11 a.m.
1 p.m.
3 p.m.
5 p.m.
7 p.m.
9 p.m.
11 p.m.
1 a.m.
3 a.m.
5 a.m.
A
High-fiber diet
P<0.05 P<0.05
P<0.05
ADA diet
Plasma Glucose (mg/dl)
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.

1396
·
May 11, 2000
The New England Journal of Medicine
sterols (1052±375 vs. 1122±565 mg per day; 95
percent confidence interval for the difference, ¡194
to 334 mg per day; P=0.60).
DISCUSSION
The intake of dietary fiber among people living in
Western countries remains low, and according to the
Third National Health and Nutrition Examination
Survey (NHANES), it averages 17 g per day in the
United States.
24
Although patients with diabetes are
advised to increase their intake of dietary fiber, in
the NHANES study, their average daily intake was
found to be only 16 g.
24
Why the intake of dietary
fiber in patients with diabetes remains low — despite
its well-documented effect of lowering plasma cho-
lesterol concentrations — remains unexplained. It is
possible that the controversy about whether there
are beneficial effects of dietary fiber on glycemic con-
trol reduces the enthusiasm of physicians and dieti-
tians for recommending high-fiber diets. The main
purpose of our study was to investigate the effects
on glycemic control of increasing the intake of die-
tary fiber. To avoid the confounding effects of con-
comitant changes in energy and macronutrients, the
two study diets were isocaloric and the macronutri-
ent composition of the diets was identical. Further-
more, unfortified foods were used as the source of
dietary fiber.
Most important, we found that the high-fiber diet
improved glycemic control, as evidenced by decreases
in the mean daily preprandial and 24-hour plasma
glucose concentrations. Urinary glucose excretion
was also lowered by the high-fiber diet. The high-
fiber diet lowered glycosylated hemoglobin values
slightly but not significantly. The high-fiber diet also
lowered 24-hour plasma insulin concentrations.
The results of previous studies that evaluated the
role of dietary fiber on glycemic control in patients
with type 2 diabetes were inconsistent. In some of the
studies, the lack of control for concomitant changes
in the intake of macronutrients makes the data diffi-
cult to interpret. For example, in the study by Kiehm
et al.
25
and in that by Simpson et al.,
26
the high-fiber
diet had a lower fat and higher carbohydrate content
than the low-fiber diet. In other studies, the inter-
pretation of the results was confounded by the short
duration of the dietary intervention,
27-29
the lack of
random assignment of the sequence of the high-fiber
and low-fiber diets,
27,29
and unexplained weight loss
during the high-fiber diet.
29
Only a few well-controlled studies have evaluated
the glycemic effects of increasing the intake of die-
tary fiber with the use of either preparations of re-
fined concentrated fiber or unfortified food, and the
results have been inconsistent.
1,30
For example, diets
that included 15 to 21 g of guar-gum fiber or oat-bran
concentrate per day had no effect on glycemic con-
trol
31,32
or resulted in only a slight improvement.
33,34
In randomized, crossover trials of six weeks’ duration
in which the intake of dietary fiber was increased by
16 g per 1000 kcal through the consumption of
foods prepared in a research kitchen or by 14 g per
day through dietary instruction, there was no im-
provement in glycemic control.
35,36
In contrast, in-
creasing dietary fiber by 23 g for three weeks and by
30 g for six weeks resulted in decreased fasting and
postprandial plasma glucose concentrations.
37,38
We
found that an increase in the intake of total dietary
fiber, which consisted predominantly of soluble fi-
ber, significantly improved glycemic control and de-
creased the degree of hyperinsulinemia in patients
with type 2 diabetes.
Our study also demonstrates the feasibility of
achieving a high intake of dietary soluble fiber by con-
suming unfortified foods. Our patients accepted the
high-fiber diet well and had few side effects; there-
fore, we recommend that patients with diabetes be
encouraged to use unfortified foods instead of less
palatable purified-fiber preparations and supplements
to increase their intake of dietary fiber.
The mechanisms of the improved glycemic con-
trol associated with high fiber intake remain unde-
fined. Whether this effect is due to an increase in
soluble fiber, insoluble fiber, or both is unclear. Be-
sides causing increased fecal excretion of bile acids,
dietary fiber may cause malabsorption of fat.
39
How-
ever, in our study, the patients’ weight did not change
with the high-fiber diet, which suggests that the de-
gree of reduction in the absorption of fat was insig-
nificant. Another possibility is that dietary fiber im-
proves glycemic control by reducing or delaying the
absorption of carbohydrates.
As expected, the high-fiber diet reduced plasma
total cholesterol concentrations by 6.7 percent, a
*Plus–minus values are means ±SD. To convert values for cholesterol
and triglycerides to millimoles per liter, multiply by 0.026 and 0.011, re-
spectively. ADA denotes American Diabetes Association, CI confidence in-
terval, VLDL very-low-density lipoprotein, LDL low-density lipoprotein,
and HDL high-density lipoprotein.
†An analysis of variance was used to compare the two diets.
T
ABLE
4.
F
ASTING
P
LASMA
L
IPID AND LIPOPROTEIN
CONCENTRATIONS DURING THE LAST WEEK
OF THE STUDY PERIODS (DAYS 38 THROUGH 42).*
VARIABLE
ADA
DIET
HIGH-
FIBER DIET
DIFFERENCE BETWEEN
DIETS (95% CI)
P
VALUE†
mg/dl
Plasma total cholesterol 210±33 196±31 ¡14 (¡27 to ¡2) 0.02
Plasma triglycerides 205±95 184±76 ¡21 (¡37 to ¡4) 0.02
Plasma VLDL cholesterol 40±19 35±16 ¡5 (¡9 to ¡1) 0.01
Plasma LDL cholesterol 142±29 133±29 ¡9 (¡22 to 3) 0.11
Plasma HDL cholesterol 29±7 28±4 ¡1 (¡4 to 3) 0.80
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.

BENEFICIAL EFFECTS OF HIGH DIETARY FIBER INTAKE IN PATIENTS WITH TYPE 2 DIABETES MELLITUS
Volume 342 Number 19 ·1397
finding consistent with the results of previous reports
of the cholesterol-reducing effects of soluble but not
insoluble fiber.40,41 Therefore, the lowering of cho-
lesterol can be attributed primarily to an average in-
crease of 17 g in the intake of soluble fiber. Previous
studies in normal subjects have reported no effects
of the amount of dietary fiber on plasma triglyceride
concentrations.42 In our study, the decrease in plas-
ma triglyceride and VLDL cholesterol concentrations
during the high-fiber diet could have been due to
the improvement in glycemic control.
The mechanisms of the reduction in plasma cho-
lesterol concentrations induced by the increased di-
etary fiber intake are controversial, however. The in-
crease in bile-acid excretion probably explains most of
the reduction, and the reduction in cholesterol ab-
sorption may also have contributed to this finding.
Previous studies have also reported a variable increase
in bile-acid excretion resulting from the consump-
tion of pectin,39,43 oat bran,4 4,45 bagasse,46 and diets
with a mixture of soluble fiber and insoluble fiber,47
but not psyllium.48 In contrast, Kesaniemi et al.47 re-
ported that a high-fiber diet did not change choles-
terol absorption in normal subjects. However, the
high-fiber diet they used included 26 g of fiber, and
it did not lower plasma cholesterol concentrations.47
In conclusion, an increase in the intake of dietary
fiber, predominantly of the soluble type, by patients
with type 2 diabetes mellitus improved glycemic con-
trol and decreased hyperinsulinemia in addition to
the expected lowering of plasma lipid concentrations.
Therefore, dietary guidelines for patients with diabe-
tes should emphasize an overall increase in dietary fi-
ber through the consumption of unfortified foods,
rather than the use of fiber supplements.
Supported in part by grants (M01-RR00633 and HL-29252) from the
National Institutes of Health and by research grants from the Bundesmin-
isterium für Bildung, Forschung, Wissenschaft und Technologie (01EC9402)
and the Deutsche Forschungsgemeinschaft (BE 1673/1-1).
We are indebted to Angela Osborn, Travis Petricek, and the nurs-
ing and dietetic ser vice of the General Clinical Research Center of
the University of Texas Southwestern Medical Center, Dallas, for
their excellent technical support and to Beverley Adams-Huet, M.S.,
for statistical analysis.
REFERENCES
1. Nutrition recommendations and principles for people with diabetes
mellitus. Diabetes Care 2000;23:S43-S46.
2. American Diabetes Association. Nutritional recommendations and prin-
ciples for individuals with diabetes mellitus: 1986. Diabetes Care 1987;10:
126-32.
3. Garg A, Bonanome A, Grundy SM, Zhang Z-J, Unger RH. Compari-
son of a high-carbohydrate diet with high-monounsaturated-fat diet in pa-
tients with non-insulin-dependent diabetes mellitus. N Engl J Med 1988;
319:829-34.
4. Rivellese AA, Giacco R, Genovese S, et al. Effects of changing amount
of carbohydrate in diet on plasma lipoproteins and apolipoproteins in type
II diabetic patients. Diabetes Care 1990;13:446-8.
5. Parillo M, Rivellese AA , Ciardullo AV, et al. A high-monounsaturated-
fat/low-carbohydrate diet improves peripheral insulin sensitivity in non-
insulin-dependent diabetic patients. Metabolism 1992;41:1373-8.
6. Rasmussen OW, Thomsen C, Hansen KW, Vesterlund M, Winther E,
Hermansen K. Effects on blood pressure, glucose, and lipid levels of a
high-monounsaturated fat diet compared with a high-carbohydrate diet in
NIDDM subjects. Diabetes Care 1993;16:1565-71.
7. Campbell LV, Marmot PE, Dyer JA, Borkman M, Storlien LH. The
high-monounsaturated fat diet as a practical alternative for NIDDM. Dia-
betes Care 1994;17:177-82.
8. Lerman-Garber I, Ichazo-Cerro S, Zamora-Gonzalez J, Cardoso-
Saldana G, Posadas-Romero C. Effect of a high-monounsaturated fat diet
enriched with avocado in NIDDM patients. Diabetes Care 1994;17:311-
5.
9. Garg A, Grundy SM, Unger RH. Comparison of effects of high and
low carbohydrate diets on plasma lipoproteins and insulin sensitivity in pa-
tients with mild NIDDM. Diabetes 1992;41:1278-85.
10. Garg A, Bantle JP, Henry RR, et al. Effects of var ying carbohydrate
content of diet in patients with non-insulin-dependent diabetes mellitus.
JAMA 1994;271:1421-8.
11. Keys A, Menotti A, Karvonen MJ, et al. The diet and 15-year death
rate in the Seven Countries Study. Am J Epidemiol 1986;124:903-15.
12. Menotti A, Keys A, Aravanis C, et al. Seven Countries Study: first 20-
year mortality data in 12 cohorts of six countries. Ann Med 1989;21:175-
9.
13. Nestle M. Mediterranean diets: historical and research overview. Am J
Clin Nutr 1995;61:Suppl:1313S-1320S.
14. Kromhout D, Keys A, Aravanis C, et al. Food consumption patterns
in the 1960s in seven countries. Am J Clin Nutr 1989;49:889-94.
15. Department of Agriculture, Agricultural Research Service. Composi-
tion of foods. Agriculture handbook No. 8. Series 8-1 to 8-16. Washing-
ton, D.C.: Government Printing Office, 1976-1987.
16. Schakel S, Sievert YA, Buzzard IM. Dietary fiber values for common
foods. In: Spiller GA , ed. CRC handbook of dietary fiber in human nutri-
tion. 2nd ed. Boca Raton, Fla.: CRC Press, 1993:567-93.
17. Public Health Service. Manual of laboratory operations: lipid research
clinics program: lipid and lipoprotein analysis. 2nd ed. Washington, D.C.:
Government Printing Office, 1982.
18. Herbert V, Lau KS, Gottlieb CW, Bleicher SJ. Coated charcoal immu-
noassay of insulin. J Clin Endocrinol Metab 1965;25:1375-84.
19. Yalow RS, Berson SA . Immunoassay of endogenous plasma insulin in
man. J Clin Invest 1960;39:1157-75.
20. Czubayko F, Beumers B, L ammsfuss S, Lutjohann D, von Bergmann
K. A simplified micro-method for quantification of fecal excretion of neu-
tral and acidic sterols for outpatient studies in humans. J Lipid Res 1991;
32:1861-7.
21. Lutjohann D, Meese CO, Crouse JR III, von Bergmann K. Evaluation
of deuterated cholesterol and deuterated sitostanol for measurement of
cholesterol absorption in humans. J Lipid Res 1993;34:1039-46.
22. Jones B, Kenward MG. Design and analysis of crossover trials. L on-
don: Chapman & Hall, 1989.
23. Conover WJ. Practical nonparametric statistics. 2nd ed. New York:
John Wiley, 1980:288-92.
24. National Health and Nutrition Examination Survey III, 1988-94.
NCHS CD-ROM series 11. No. 2A. ASCII version. Hyattsville, Md.: Na-
tional Center for Health Statistics, April 1998.
25. Kiehm TG, Anderson JW, Ward K. Beneficial effects of a high carbo-
hydrate, high fiber diet on hyperglycemic diabetic men. Am J Clin Nutr
1976;29:895-9.
26. Simpson HCR , Simpson RW, Lousley S, et al. A high carbohydrate
leguminous fibre diet improves all aspects of diabetic control. Lancet 1981;
1:1-5.
27. Rivellese A, R iccardi G, Giacco A, et al. Ef fect of dietary fibre on glu-
cose control and serum lipoproteins in diabetic patients. L ancet 1980;2:
447-50.
28. Riccardi G, R ivellese A, Pacioni D, Genovese S, Mastranzo P, Mancini
M. Separate influence of dietary carbohydrate and fibre on the metabolic
control in diabetes. Diabetologia 1984;26:116-21.
29. O’Dea K, Traianedes K, Ireland P, et al. The effects of diet differing
in fat, carbohydrate, and fiber on carbohydrate and lipid metabolism in
type II diabetes. J Am Diet Assoc 1989;89:1076-86.
30. Nuttall FQ. Dietary fiber in the management of diabetes. D iabetes
1993;42:503-8.
31. Holman RR, Steemson J, Darling P, Turner RC. No glycemic benefit
from guar administration in NIDDM. Diabetes Care 1987;10:68-71.
32. Uusitupa M, Siitonen O, Savolainen K, Silvasti M, Penttila I, Parvi-
ainen M. Metabolic and nutritional effects of long-term use of guar gum
in the treatment of noninsulin-dependent diabetes of poor metabolic con-
trol. Am J Clin Nutr 1989;49:345-51.
33. Aro A, Uusitupa M, Voutilainen E, Hersio K, Korhonen T, Siitonen
O. Improved diabetic control and hypocholesterolaemic effect induced by
long-term dietary supplementation with guar gum in t ype 2 (insulin-inde-
pendent) diabetes. Diabetologia 1981;21:29-33.
34. Pick ME, Hawrysh ZJ, Gee MI, Toth E, Garg ML , Hardin RT. Oat
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.

1398 ·May 11, 2000
The New England Journal of Medicine
bran concentrate bread products improve long-term control of diabetes: a
pilot study. J Am Diet Assoc 1996;96:1254-61.
35. Hollenbeck CB, Coulston AM, Reaven GM. To what extent does in-
creased dietary fiber improve glucose and lipid metabolism in patients with
noninsulin-dependent diabetes mellitus (NIDDM)? Am J Clin Nutr 1986;
43:16-24.
36. Manhire A, Henry CL, Hartog M, Heaton KW. Unrefined carbohy-
drate and dietary fibre in treatment of diabetes mellitus. J Hum Nutr 1981;
35:99-101.
37. Karlstrom B, Vessby B, Asp NG, et al. Effects of an increased content
of cereal fibre in the diet of type 2 (non-insulin-dependent) diabetic pa-
tients. Diabetologia 1984;26:272-7.
38. Hagander B, Asp NG, Efendic S, Nilsson-Ehle P, Schersten B. Dietary
fiber decreases fasting blood glucose levels and plasma LDL concentration
in noninsulin-dependent diabetes mellitus patients. Am J Clin Nutr 1988;
47:852-8.
39. Kay RM, Truswell AS. Effect of citrus pectin on blood lipids and fecal
steroid excretion in man. Am J Clin Nutr 1977;30:171-5.
40. Jenkins DJ, Newton C, Leeds AR , Cummings JH. Effect of pectin,
guar gum, and wheat fibre on serum-cholesterol. Lancet 1975;1:1116-7.
41. Hillman LC, Peters SG, Fisher CA , Pomare EW. The effects of the fi-
ber components pectin, cellulose and lignin on serum cholesterol levels.
Am J Clin Nutr 1985;42:207-13.
42. Brown L, Rosner B, Willett W W, Sacks FM. Cholesterol-lowering ef-
fects of dietary fiber: a meta-analysis. Am J Clin Nutr 1999;69:30-42.
43. Cummings JH, Southgate DAT, Branch WJ, et al. The digestion of
pectin in the human gut and its effect on calcium absorption and large
bowel function. Br J Nutr 1979;41:477-85.
44. Kirby RW, Anderson JW, Sieling B, et al. Oat-bran intake selectively
lowers serum low-density lipoprotein cholesterol concentrations of hyper-
cholesterolemic men. Am J Clin Nutr 1981;34:824-9.
45. Zhang JX, Hallmans G, Andersson H, et al. Effect of oat bran on plas-
ma cholesterol and bile acid excretion in nine subjects with ileostomies. Am
J Clin Nutr 1992;56:99-105.
46. Walters RL, Baird IM, Davies PS, et al. Effects of two types of dietary
fibre on faecal steroid and lipid excretion. BMJ 1975;2:536-8.
47. Kesaniemi YA, Tarpila S, Miettinen TA. Low vs high dietary fiber and
serum, biliary, and fecal lipids in middle-aged men. Am J Clin Nutr 1990;
51:1007-12.
48. Abraham ZD, Mehta T. Three-week psyllium-husk supplementation:
effect on plasma cholesterol concentrations, fecal steroid excretion, and
carbohydrate absorption in men. Am J Clin Nutr 1988;47:67-74.
POSTING PRESENTATIONS AT MEDICAL MEETINGS ON THE INTERNET
Posting an audio recording of an oral presentation at a medical meeting on the
Internet, with selected slides from the presentation, will not be considered prior
publication. This will allow students and physicians who are unable to attend the
meeting to hear the presentation and view the slides. If there are any questions
about this policy, authors should feel free to call the Journal’s Editorial Offices.
Downloaded from www.nejm.org by MR KEVIN PATRICK on November 17, 2003.
Copyright © 2000 Massachusetts Medical Society. All rights reserved.