ArticlePDF AvailableLiterature Review

Effects of Dietary Fiber and Carbohydrate on Glucose and Lipoprotein Metabolism in Diabetic Patients

Authors:

Abstract

Dietary recommendations for the treatment of diabetic patients issued by national and international diabetes associations consistently emphasize the need to increase carbohydrate consumption. However, these recommendations have been questioned on the basis of growing evidence that, in both insulin-dependent and non-insulin-dependent diabetic patients, a high-carbohydrate diet does not offer any advantage in terms of blood glucose and plasma lipid concentrations compared with a high-fat (mainly unsaturated) diet. It has been shown repeatedly that a high-carbohydrate diet increases plasma insulin and triglyceride levels and can deteriorate blood glucose control in the postprandial period. However, much of the controversy between advocates and detractors of dietary carbohydrate can be settled by taking into account dietary fiber. Several studies have shown that the adverse metabolic effects of high-carbohydrate diets are neutralized when fiber and carbohydrate are increased simultaneously in the diet for diabetic patients. In particular, these studies demonstrated that a high-carbohydrate/high-fiber diet significantly improves blood glucose control and reduces plasma cholesterol levels in diabetic patients compared with a low-carbohydrate/low-fiber diet. In addition, a high-carbohydrate/high-fiber diet does not increase plasma insulin and triglyceride concentrations, despite the higher consumption of carbohydrates. Unfortunately, dietary fiber represents a heterogenous category, and there is still much to understand as to which foods should be preferred to maximize the metabolic effects of fiber. There are indications that only water-soluble fiber is active on plasma glucose and lipoprotein metabolism in humans. Therefore, in practice, the consumption of legumes, vegetables, and fruits--rich in water-soluble fiber--should be particularly encouraged. The mechanisms by which dietary fiber exerts its hypoglycemic and hypolipidemic activities are unknown. However, the ability of dietary fiber to retard food digestion and nutrient absorption certainly has an important influence on lipid and carbohydrate metabolism. The beneficial effects of high-fiber foods are also exerted by some foods not particularly rich in fiber. The fiber content and physical form of the food can influence the accessibility of nutrients by digestive enzymes, thus delaying digestion and absorption. The identification of these foods with a low-glycemic response would help enlarge the list of foods particularly suitable for diabetic patients. In conclusion, a diet low in cholesterol and saturated fat should be recommended to all diabetic patients to prevent cardiovascular disease. A balanced increase in consumption of fiber-rich foods and unsaturated fat is the most rational way to replace foods rich in saturated fat and cholesterol in the diabetic diet.
Effects of Dietary Fiber
and Carbohydrate on Glucose
and Lipoprotein Metabolism
in Diabetic Patients
Gabriele
Riccardl,
MD
Angela
A.
Wvellese, MD
Dietary recommendations for the treatment of
diabetic patients issued by national and international
diabetes associations consistently emphasize the
need to increase carbohydrate consumption.
However, these recommendations have been
questioned on the basis of growing evidence that, in
both insulin-dependent ana non-insulin-dependent
diabetic patients, a high-carbohydrate diet does not
offer any advantage in terms of blood glucose and
plasma lipid concentrations compared with a high-fat
(mainly unsaturated) diet. It has been shown
repeatedly that a high-carbohydrate diet increases
plasma insulin and triglyceride levels and can
deteriorate blood glucose control in the postprandial
period.
However, much of the controversy between
advocates and detractors of dietary carbohydrate can
be settled by taking into account dietary fiber.
Several studies have shown that the adverse
metabolic effects of high-carbohydrate diets are
neutralized when fiber and carbohydrate are
increased simultaneously in the diet for diabetic
patients. In particular, these studies demonstrated
that a high-carbohydrate/high-fiber diet significantly
improves blood glucose control and reduces plasma
cholesterol levels in diabetic patients compared with
a low-carbohydrate/low-fiber diet. In addition, a
high-carbohydrate/high-fiber diet does not increase
plasma insulin and triglyceride concentrations,
despite the higher consumption of carbohydrates.
Unfortunately, dietary fiber represents a
heterogenous category, and there is still much to
understand as to which foods should be preferred to
maximize the metabolic effects of fiber. There are
From the Institute of Internal Medicine and Metabolic Diseases, Second Medi-
cal School, University of Naples, Naples, Italy.
Address correspondence and reprint requests to Gabriele Riccardi, MD,
Institute of Internal Medicine and Metabolic Diseases, Second Medical School,
University of Naples, Via S. Pansini, 5, 80131 Naples, Italy.
indications that only water-soluble fiber is active on
plasma glucose and lipoprotein metabolism in
numans. Therefore, in practice, the consumption of
legumes, vegetables, and fruits—rich in water-soluble
fiber—should be particularly encouraged. The
mechanisms by wnich dietary fiber exerts its
hypoglycemic and hypolipidemic activities are
unknown.
However, the ability of dietary fiber to
retard food digestion and nutrient absorption
certainly has an important influence on fipid and
carbohydrate metabolism. The beneficial effects of
high-fiber foods are also exerted by some foods not
particularly rich in fiber. The fiber content and
physical form of the food can influence the
accessibility of nutrients by digestive enzymes, thus
delaying digestion and absorption. The identification
of these foods with a low-glycemic response would
help enlarge the list of foods particularly suitable for
diabetic patients. In conclusion, a diet low in
cholesterol and saturated fat should be
recommended to all diabetic patients to prevent
cardiovascular disease. A balanced increase in
consumption of fiber-rich foods and un saturated fat
is the most rational way to replace foods rich in
saturated fat and cholesterol in the diabetic diet.
Diabetes Care 14:1115-25,1991
T
he approach to the dietary treatment of pa-
tients with diabetes mellitus has drastically
changed in the last decade. Diabetologists are
becoming increasingly more aware that the
diabetic diet should not only help improve blood glu-
cose control but also help prevent cardiovascular dis-
ease.
There is much evidence that diabetic patients are
at high risk of cardiovascular disease and that this risk
DIABETES CARE, VOL. 14, NO. 12, DECEMBER 19911115
DIETARY FIBER, CARBOHYDRATE, AND LIPID METABOLISM
increases in the presence of one or more of the other
risk factors (1-3). In particular, there are data support-
ing a direct linear relationship between plasma choles-
terol concentrations and cardiovascular mortality in
diabetic and nondiabetic individuals (4). This indicates
that there is no clear threshold separating individuals
with high or normal plasma cholesterol values; thus,
the lower the plasma cholesterol levels, the better the
prognosis for the patient.
The most effective dietary measure to decrease
plasma cholesterol levels in humans is to reduce the
intake of saturated fat and cholesterol (5). Therefore, a
diet low in cholesterol and saturated fat is recom-
mended to all diabetic patients (regardless of their
plasma cholesterol levels) to prevent or delay the oc-
currence of cardiovascular disease (6-7).
TOWARD
A
DIETARY STRATEGY
FOR
TREATMENT
OF
DIABETIC PATIENTS
Many diabetic patients, particularly those with non-
insulin-dependent diabetes (NIDDM), are overweight
and would benefit from a reduction in caloric intake
due to a lower consumption of foods rich in saturated
fat or cholesterol. However, in patients of normal
weight, foods rich in saturated fat and cholesterol
should be partially replaced by alternative sources of
calories. Because it does not seem prudent to increase
protein intake, which is already high in western diets,
there are two possible alternatives to replace saturated
fat in the diabetic diet: 7) carbohydrates and 2) unsatu-
rated fats. The first approach is usually referred to as
the "low-fat/high-carbohydrate diet" (total and
satu-
rated fats are reduced and replaced by carbohydrate),
whereas the alternative represents the so-called "fat-
modified diet" (total fat is not reduced because
satu-
rated fat is replaced by unsaturated fat) (Fig. 1). Both
approaches include a reduction of dietary cholesterol.
All current dietary recommendations for diabetic pa-
tients are based on the low-fat/high-carbohydrate diet
(Table 1). However, the statement of the American
Diabetes Association is more stringent because it limits
in every case the intake of total fat to <30% of total
energy intake (6). Conversely, the European document,
although advising a similar amount of total fat, allows
a higher intake of total fat, provided that monounsat-
urated and not saturated fatty acids are increased (7).
The various documents do not state clearly the rea-
sons for preferring the low-fat/high-carbohydrate diet
for the treatment of diabetic patients. However, this
choice has been strongly influenced by the adoption of
similar policies by other scientific societies aiming at
cardiovascular disease prevention in the general public
(8,9).
The underlying reason is probably the assump-
tion that, not only does the intake of saturated fat and
cholesterol but the intake of total fat also needs to be
reduced to lower plasma cholesterol concentrations.
CHOL: 300mg/lttay
too
OOOmg/ttayXOmoMay
UN8ATURATED
FATS
CARBOHYDRATES
PROTONS
FATWET
LOW FAT
MET
FIG.
1. Schematic representation of nutrient distribution
in diet habitually consumed in western countries and 2
alternative dietary approaches that can be used to re-
duce plasma cholesterol (chol) levels in diabetic and
nondiabetic individuals.
This is not necessarily true because epidemiological
intercultural comparisons and dietary experiments
have clearly shown that among the various types of
dietary fat, only saturated fat can increase plasma cho-
lesterol levels (10-12). Conversely, the intake of unsat-
urated fat (monounsaturated and polyunsaturated) has
no significant hypercholesterolemic effect in humans
(13).
Therefore, there is a rationale for proposing the fat-
modified diet as a suitable alternative to the low-fat
diet to achieve a substantial reduction of plasma and
low-density lipoprotein cholesterol (LDL-chol) concen-
TABLE 1
Recommended composition of weight-maintaining diet
for diabetic patients
Protein
Fat
Total
Saturated
Monounsaturated
Polyunsaturated
Cholesterol
Carbohydrate
Sucrose (added)
Fiber
Alcohol
Sodium
American Diabetes
Association
(1986)
10-15%
<30%
<10%
<13%
<7%
<300 mg/day
55-60%
modest
40 g/day
<10 g/day
<3 g/day
European Association
for the Study of
Diabetes (1988)
Decreased
<30%*
<10%
<10%
<10%
<300 mg/day
50-60%
<30 g/day
40 g/day
<25-30 g/dayt
<6 g/day
"Intake of total fat may be higher, provided that only monounsat-
urated fat is increased.
tSame intake recommended to general population (ref. 8)
1116DIABETES CARE, VOL. 14, NO. 12, DECEMBER 1991
G. R1CCARDI AND
AA
RIVELLESE
trations in diabetic and in nondiabetic individuals. Fur-
thermore, this alternative approach is justified because
the treatment of diabetic patients with low-fat/high-
carbohydrate diets has been questioned because of
possible adverse metabolic effects (14). In particular, it
has been claimed that low-fat/high-carbohydrate diets
would deteriorate blood glucose control, adversely af-
fect plasma lipoproteins (increasing triglycerides and
decreasing high-density lipoproteins [HDL]), and ele-
vate the concentration of plasma insulin (considered as
a possible cardiovascular risk factor) (15-17). Based
on these considerations, an official document of the
National Institutes of Health (Bethesda, MD) suggested
that low-fat/high-carbohydrate diets may be suitable
for some but not all NIDDM patients, and, therefore,
in some cases a fat-modified diet would be more ap-
propriate (18).
The comparison between high-carbohydrate/low-fat
diets and fat-modified diets in diabetic patients has
been attempted in many studies that gave conflicting
results. However, to interpret correctly the existing
evidence, attention should be paid to the exact com-
position of the two diets being compared.
High-carbohydrate/low-fat diets may contain differ-
ent proportions of monodisaccarides, starch, or fiber,
whereas in fat-modified diets, the proportion of mono-
unsaturated/C20:5o)3 polyunsaturated/C20:5w6 poly-
unsaturated fatty acids may vary. Each of these dietary
constituents has different effects on glucose and lipid
metabolism, therefore, the comparison between
high-
carbohydrate/low-fat and fat-modified diets may favor
one or the other dietary approach according to the
types of carbohydrates and fats used.
This might explain many of the discrepancies be-
tween the results of different studies. In most trials
showing a superiority of the low-fat/high-carbohydrate
diet, the diet also had lower saturated fat and a higher
fiber content (19,20). Conversely, when the results
were not in favor of the low-fat/high-carbohydrate diet,
a substantial proportion of the carbohydrate quota in
this diet was monodisaccarides (21). Furthermore,
some of these dietary experiments were not relevant to
the clinical experience because the diets used were
not physiological in terms of both the use of unnatural
foods (liquid-formula diets) and in the choice of a very
unbalanced diet composition (>70% carbohydrate
with little or no fat) (22,23).
Therefore, despite many published studies compar-
ing fat-modified with high-carbohydrate/low-fat diets,
only a few are clinically relevant and also give an
unequivocal answer as to which of these two dietary
approaches yields a more pronounced cholesterol-low-
ering activity, induces a greater improvement of blood
glucose control, and has a more beneficial influence
on the overall cardiovascular risk-factor profile in dia-
betic patients.
For a clinically relevant comparison, the diets to be
tested should have some specific prerequisites. The
major fat component in the fat-modified diet has nec-
essarily to be monounsaturated fat because consump-
tion of polyunsaturated fat >10% of the daily energy
intake is not recommended (8,9,24). This is motivated
by the lack of available evidence on the long-term
safety of very-high-polyunsaturated-fat diets. On the
other hand, diets rich in monounsaturated fat are cer-
tainly safe because they have been consumed for cen-
turies in Mediterranean countries where cardiovascular
and total mortalities are particularly low (25). In
addi-
tion,
in contrast with previous beliefs, experimental
data in hyperlipidemic individuals have shown that
diets rich in either mono-unsaturated or polyunsaturated
fat exert similar effects on plasma lipoproteins (5,26,27).
In relation to the low-fat/high-carbohydrate diet,
there are only two possible alternatives that are
clini-
cally relevant: /) a diet in which starchy foods repre-
sent the major source of dietary carbohydrate; 2) a
low-fat/high-carbohydrate diet in which the carbohy-
drate quota derives mostly from unprocessed vegeta-
bles rich in fiber. These two possibilities should be
considered separately because there is much evidence
that dietary fiber influences blood glucose and lipid
metabolism in both diabetic and nondiabetic individu-
als (28,30). It can be anticipated that much of the
controversy about low-fat/high-carbohydrate diets
could be settled if the metabolic effects of dietary fiber
in diabetic patients were taken into account. Also, the
range of variations in the proportion of carbohydrate to
fat in the diets tested has to be necessarily limited.
From a practical point of view, it is irrelevant to com-
pare an almost fat-free diet with a diet with >50% fat.
The really meaningful question in terms of clinical
implementation is whether carbohydrate and fat
should represent 60 and 20 or 40 and 40%, respec-
tively, of total energy intake (14).
HIGH STARCH
OR
HIGH MONOUNSATURATED FAT?
The comparison between a high-starch and a
high
monounsaturated fat diet has been attempted by our
group,
in both insulin-dependent diabetic (IDDM; 31)
and NIDDM patients (16). In one diet, the contribution
of carbohydrate and fat to the daily calorie intake was
60 and 20%, respectively (low-fat/high-carbohydrate
diet),
whereas in the other diet, carbohydrate was re-
duced to 40%, and fat was increased to 40% (fat-
modified diet). This dietary variation was obtained by
substituting olive oil (rich in monounsaturated fat) with
an isoenergetic amount of starchy foods (bread, pota-
toes,
rice). Both diets were low in saturated fat (—10%
calories) and cholesterol
(—350
mg/day) and had
simi-
lar contents of calories, monodisaccarides, fiber, and
protein.
The only difference between the diets was in
the amounts of monounsaturated fat and complex di-
gestible carbohydrate, respectively. Both diets con-
sisted only of natural foods distributed in various daily
DIABETES CARE, VOL. 14, NO. 12, DECEMBER 19911117
DIETARY FIBER, CARBOHYDRATE, AND LIPID METABOLISM
POST PRANDIAL BLOOO GLUCOSE
FIG.
2. Mean ± SE daily blood glucose profiles re-
corded 15 days after either a high-starch (•) or high
monounsaturated fat (•) diet in insulin-dependent dia-
betic patients (n = 8) (ref. 31).
menus. Patients were randomly assigned to one of the
diets,
which was followed for 15 days, then crossed
over to the other diet, which was also followed for 15
days.
In IDDM patients, the insulin dosage was not modi-
fied throughout the study and at the end of each
dietary period, blood glucose was continuously moni-
tored for 24 h. Moreover, a fasting blood sample was
taken for lipoprotein analysis. Blood glucose concen-
trations throughout the 24 h were consistently higher
after the high-starch diet; however, the greatest differ-
ence was reached after meals, particularly after lunch
(the main meal in Italy); at this time, blood glucose
concentration was on average —5.6 mM higher than
after the monounsaturated fat-rich diet. In addition,
plasma glucose excursions during the day were much
higher after the starch-rich diet (Fig. 2). In contrast,
plasma lipoproteins were not significantly modified by
diet; however, patients had reasonably good blood
glucose control and were normolipidemic.
The same diets were also evaluated in NIDDM pa-
tients treated with diet alone or diet plus hypoglycemic
drugs,
with a similar experimental design. In this case,
at the end of the two dietary periods, the glycemic
control was evaluated by measuring blood glucose
concentrations in the fasting state and 2 h after lunch.
At both times, plasma insulin was also measured,
whereas plasma lipoproteins were analyzed in a fast-
ing patient.
Even in these patients, blood glucose control was
worse after the high-starch diet; in particular, post-
prandial blood glucose concentrations were signifi-
cantly higher after the high-starch diet, whereas, fast-
ing blood glucose levels did not vary with the two
diets (Fig. 3). In addition, the starch-rich diet compared
with the diet rich in monounsaturated fat induced
significantly higher plasma insulin levels—particularly
p<0.02
POST PRANDIAL PLASMA NSUUN
p< O.OO2
1
fASTMG BLOOD GLUCOSEFASTING PLASMA INSULIN
I'
JL
FIG.
3. Mean ± SE fasting and postprandial blood glu-
cose and plasma insulin concentrations measured after
15 days on either a high-starch (open bars) or high-
monounsaturated fat (hatched bars) diet in non-insulin-
dependent diabetic patients (n = 18).
in the postprandial phase (Fig. 3)—and significantly
higher plasma triglyceride levels (Fig. 4). Conversely,
total and HDL-chol concentrations were similar at the
end of the two dietary periods (Fig. 4).
Similar experiments by other groups have yielded
substantially similar results (15). Moreover, it has been
clearly demonstrated in IDDM patients that the
amount of carbohydrate in the meal influences post-
prandial blood glucose concentrations, even if the
amount of insulin injected is modified according to the
carbohydrate content of the meal (32,33).
The results we obtained in NIDDM patients treated
with diet and/or hypoglycemic drugs are in close
agreement with those of Carg et al. (15), who evalu-
ated the metabolic effects of partial substitution of
monounsaturated fat for starch in the diet of NIDDM
patients treated with insulin. This dietary maneuver
resulted in better blood glucose control, reduced
insu-
lin requirements, lower concentrations of plasma tri-
glycerides, lower levels of very-low-density lipoprotein
(VLDL), and higher levels of HDL. Conversely, plasma
levels of total and LDL-chol did not vary. These results
agree also with those of Coulston et al. (34).
There are some indications in NIDDM patients with
1118DIABETES CARE, VOL. 14, NO. 12, DECEMBER 1991
G.
RICCARDI AND A.A. RIVELLESE
PLASMA CHOLESTEROLPLASMA TI06LYCERIDES
p<O-OOI
X,
HDL-CHOLESTEROL
FIG.
4. Mean ± SE fasting plasma lipid levels measured
after 15 days on either high-starch (open bars) or high-
mononsaturated fat (hatched bars) diet in non-insulin-
dependent diabetic patients (n = 18). HDL, high-density
lipoprotein.
that might have contributed to the improvement of
insulin sensitivity.
Available evidence clearly indicates that an increase
in the intake of complex digestible carbohydrate is not
always the best way to replace saturated fat and cho-
lesterol in the diabetic diet. In IDDM patients, high
consumption of starchy foods significantly deteriorates
blood glucose control and increases the glycemic ex-
cursions throughout the day. This can only be partially
prevented by increasing the dosage of insulin.
In NIDDM patients, a high-starch diet deteriorates
blood glucose control (although to a lesser extent and
only in the postprandial phase) but increases plasma
insulin concentrations, elevates plasma triglycerides,
lowers plasma levels of HDL (in some patients), and
worsens peripheral insulin sensitivity. Such untoward
effects of the high-starch diet can be more pronounced
in particular groups of patients
(i.e.,
hypertriglyceri-
demic, severely hyperglycemic).
a mild form of the disease that the effects of a
high-
carbohydrate diet on blood glucose control will not be
as deleterious as in the more severe cases (35). How-
ever, this hypothesis still deserves adequate testing.
Another important parameter that needs to be evalu-
ated in NIDDM patients treated with a low-fat/high-
carbohydrate diet or a fat-modified diet is insulin sen-
sitivity. It has been claimed that a higher carbohydrate
consumption leads to a general increase in the ability
of insulin to stimulate glucose disposal. Impaired
insu-
lin sensitivity represents a key metabolic derangement
in diabetes and seems to also be involved in the
etiology of its cardiovascular complications (36,37).
Therefore, the possibility of improving insulin sensitiv-
ity by means of a low-fat/high-carbohydrate diet has
always been a strong argument in favor of the adop-
tion of this type of diet by NIDDM patients. However,
evidence supporting this concept suffers from major
flaws due to poor diet comparability (differences in the
polyunsaturated-saturated fat ratio, fiber content), ex-
treme dietary variations (>75% carbohydrate), use of
liquid-formula diets, and inadequate methods to assess
insulin sensitivity (oral glucose tolerance tests, insulin
infusion test) (19-23,38,39).
In our dietary experiment on the metabolic effects of
a high-starch diet compared with a high-monounsat-
urated fat diet, we evaluated insulin sensitivity by the
euglycemic-hyperinsulinemic clamp technique in a
subgroup of nine NIDDM patients. Contrary to our
expectations, insulin sensitivity did not improve after
the high-starch diet; the opposite was true because
insulin-stimulated glucose disposal (M value) was sig-
nificantly increased after the high-monounsaturated fat
diet compared with the high-carbohydrate diet
ROLE
OF
DIETARY FIBER
IN
TREATMENT
OF
DIABETIC PATIENTS
-1
(mean ± SE 6.2 ± 0.6 vs. 4.8 ± 0.6 mg
kg"1 min
P < 0.05; 40). This finding is consistent with all other
metabolic effects of the high-monounsaturated fat diet
(lower plasma glucose, insulin, and triglyceride levels)
Results of studies comparing a fat-modified diet with a
low-fat/high-carbohydrate diet have been clearly in fa-
vor of the former. However, this does not imply that
the outcome has to be the same if the fat-modified diet
is compared with a low-fat/high-carbohydrate diet in
which unprocessed vegetable foods (fiber rich), instead
of starch, represent the major carbohydrate source.
Studies show that dietary fiber has important beneficial
effects on blood glucose and lipoprotein metabolism
(30).
There are many definitions of dietary fiber; however,
the one that seems more appropriate from a physiolog-
ical and clinical point of view was proposed in the
1970s by Trowell (41), who first suggested that dietary
fiber may play a role in diabetes. According to this
definition,
dietary fiber is represented by plant poly-
saccharides that are resistant to hydrolysis by the di-
gestive enzymes of humans.
The important role of dietary fiber in the treatment
of diabetic patients was indicated by the pioneer
stud-
ies of Kiehm et al. (42), who showed a drastic reduc-
tion of the insulin dosage and a dramatic improvement
of blood glucose control in insulin-treated diabetic
patients by prescribing a high-carbohydrate/high-fiber
diet. However, from that study, it was not clear to
what extent the effects of the high-carbohydrate/high-
fiber diet should be ascribed to the high-fiber content
or to other dietary modifications, including changes in
the type and amount of carbohydrate and fat. There-
fore,
we decided to undertake a dietary experiment to
evaluate the separate influence of dietary fiber and
carbohydrate on blood glucose and lipoprotein metab-
olism (28,43). In six IDDM and eight NIDDM patients,
we compared a fat-modified diet in which olive oil
DIABETES CARE, VOL. 14, NO. 12, DECEMBER 19911119
DIETARY FIBER, CARBOHYDRATE, AND LIPID METABOLISM
-1
£ -2
-3
\\\\\\\\\\vi
T
Fasting2-h Poet-prandialAverage dally
FIG.
5. Mean ± SE variations in blood glucose levels
obtained by changing from low-carbohydrate/low-fiber
to high-carbohydrate/high-fiber diet in insulin-depen-
dent (stippled
bars)
and non-insulin-dependent (hatched
bars) diabetic patients. This dietary maneuver had a sig-
nificant effect on 2-h postprandial (P < 0.01) and daily
average (P < 0.005) blood glucose concentrations but
not on fasting glucose levels (NS); the effect was not
significantly different for insulin-dependent non-insulin-
dependent diabetic patients (ref. 43).
was the major source of fat, with a low-fat/high-carbo-
hydrate diet in which the high-carbohydrate quota was
obtained by increasing the consumption of fiber-rich
foods (legumes, fruit, vegetables). In the first diet, car-
bohydrate and fat represented 42 and 37% of the total
energy intake, whereas the other diet had proportions
of these dietary constituents equal to 53 and 30%,
respectively. Both diets were low in saturated fat and
cholesterol and had a similar content of calories, mono-
disaccarides, and protein. To evaluate separately the
effects of dietary fiber and carbohydrate on glucose
and lipoprotein metabolism, a third diet was included
in the experiment. This was also a low-fat/high-carbo-
hydrate diet (53% carbohydrate, 30% fat), but in this
case,
the carbohydrate quota was mainly fiber-de-
pleted starchy foods (bread, potatoes). Thus, fiber con-
tent was 20 g/day in the fat-modified diet, 54 g/day in
the fiber-enriched low-fat/high-carbohydrate diet, and
16 g/day in the fiber-depleted low-fat/high-carbohy-
drate diet. The three diets, which consisted exclusively
of natural foods, were consumed in random order for
consecutive periods of 10 days while the patients were
hospitalized in a metabolic ward.
The comparison between the fat-modified diet and
the fiber-rich low-fat/high-carbohydrate diet was
clearly in favor of the latter. In both IDDM and
NIDDM patients, the fiber-rich low-fat/high-carbohy-
drate diet produced a significant decrease in postpran-
dial and daily average blood glucose concentrations;
fasting glucose levels were also reduced but the differ-
ence did not reach statistical significance (Fig. 5). Also
LDL-chol concentrations were significantly reduced af-
ter the fiber-rich low-fat/high-carbohydrate diet com-
pared with the fat-modified diet in both types of dia-
betic patients; conversely, VLDL triglyceride levels
0.6
-0.5
-1
ilia
_L
LDL CholesterolHDL Cholesterol VLDL Trlglycerides
FIG.
6. Mean ± SE variations in serum lipoprotein com-
position obtained by changing from low-carbohydrate/
low-fiber to high-carbohydrate/high-fiber diet in insulin-
dependent (stippled bars) and non-insulin-dependent
(hatched bars) diabetic patients. This dietary maneuver
had a significant effect on low-density lipoprotein (LDL)
(P < 0.001) and high-density lipoprotein (HDL) (P <
0.001) cholesterol concentrations but not on very-low-
density lipoprotein (VLDL) triglyceride levels (NS); the
effect was not significantly different for insulin-depen-
dent non-insulin-dependent diabetic patients (ref. 43).
were not significantly modified. There was a signifi-
cant reduction in HDL-chol concentrations in both
IDDM and NIDDM patients with the fiber-rich
high-
carbohydrate diet (Fig. 6).
The beneficial effects of fiber-rich low-fat/high-car-
bohydrate diets have been often ascribed to their
high-
carbohydrate content rather than to fiber. Therefore,
this experiment evaluated the effects of fiber, per se,
on glucose and lipoprotein metabolism. This was done
by comparing the results of the two diets with an
identical carbohydrate quota (53% of calories in both
diets) and differing only for the amount of fiber (16 vs.
54 g/day). This comparison clearly shows that an in-
creased amount of dietary fiber in the diet is able, per
se,
and independent of any change in the other dietary
components, to reduce significantly postprandial and
daily average blood glucose concentrations by 25 and
15%,
respectively, and to reduce LDL-chol and VLDL
triglyceride levels by 25 and 10%, respectively, in
both IDDM and NIDDM patients. Conversely, there is
no significant effect of dietary fiber, per se, on fasting
blood glucose and HDL-chol concentrations.
The results of this study have been confirmed by
other groups. In particular, in a randomized crossover
study performed in IDDM and NIDDM outpatients,
Simpson et al. (20) demonstrated that a high-carbohy-
drate/high-fiber diet can be followed by diabetic pa-
tients in their usual home setting without major feasi-
bility problems and with the same beneficial metabolic
effects that we observed in hospitalized patients. Other
concordant results have been obtained in diabetic
children,
pregnant diabetic women, patients with sec-
ondary failure to oral hypoglycemic drugs, and dia-
betic patients with chronic renal failure (44-47).
1120DIABETES CARE, VOL. 14, NO. 12, DECEMBER 1991
G. RICCARDI AND AA RIVELLESE
TABLE 2
Metabolic effects (increase, decrease, no effect) of low-fat high-carbohydrate diet in which either starch or fiber-rich
foods are mainly represented
High starch
High fiber
Blood
glucose
Increased
Decreased
Plasma
insulin
Increased
No effect
Insulin
sensitivity
Decreased
Increased
VLDL
Increased
No effect
Plasma lipoproteins
LDL HDL
No effect No effect/decreased
Decreased No effect/decreased
VLDL, very-low-density lipoprotein; LDL, low-density lipoprotein; HDL, high-density lipoprotein.
However, there have been other studies in which
the improvement of blood glucose control in diabetic
patients by means of a high-carbohydrate/high-fiber
diet could not be reproduced (48,49). This apparent
discrepancy can be explained by examining the list of
the foods used for the high-fiber diets in the various
experiments. In our study and most other studies with
results concordant with ours, the high dietary fiber
intake was mainly obtained by increasing the con-
sumption of legumes, fruits, and vegetables. Con-
versely, no significant effects of the high-fiber diet
were recorded in studies in which whole-meal cereals
represented the major source of dietary fiber (48). This
difference can be at least partly explained by the dif-
ferent properties of fiber present in cereals, on the one
hand,
and legumes, fruits, and some vegetables on the
other. The latter foods contain mostly water-soluble,
gel-forming fiber able to retard the mixing of food with
the intestinal juice, thus delaying the digestion and
absorption of nutrients; conversely, cereal fiber is
largely non-water-soluble and unable to significantly
influence the rate of food digestion in the small intes-
tine (30,50). In addition, in some studies claiming that
dietary fiber has no effect on blood glucose control,
the measurements of blood glucose concentrations
were only performed in the fasting state, whereas, the
most pronounced influence of dietary fiber on blood
glucose metabolism is exerted in the postprandial pe-
riod (49).
In the evaluation of the metabolic properties of the
high-carbohydrate/high-fiber diet, another aspect that
deserves attention is the effect of this diet on insulin
sensitivity. This has been attempted only in nondia-
betic individuals, and the results seem to indicate that
high-carbohydrate/high-fiber diets can induce improve-
ment of insulin sensitivity (51).
The results of the comparison between a fat-modi-
fied diet and a low-fat/high-carbohydrate diet go in op-
posite directions according to whether fiber-depleted
starchy foods or unprocessed vegetable foods (fiber
rich) are present in the high-carbohydrate diet (Table
2).
In the latter case (contrary to what has been shown
in the previous section for the starch-rich diet), the
low-fat/high-carbohydrate diet has clear metabolic ad-
vantages over the fat-modified diet in both IDDM and
NIDDM patients. It improves blood glucose control
(mainly in the postprandial period), decreases plasma
and LDL-chol concentrations, and prevents the eleva-
tion of the plasma triglyceride levels, usually observed
when dietary carbohydrates are increased. Also, it may
improve peripheral insulin sensitivity.
In relation to the effects of fiber-rich foods on
plasma lipoprotein metabolism, note that their ability
to reduce plasma cholesterol and LDL concentrations
is independent of other dietary modifications. In par-
ticular, reduction in dietary cholesterol and saturated
fat and increased consumption of fiber-rich foods have
an additive effect in lowering plasma cholesterol and
LDL levels. The overall effect of this combined dietary
maneuver is a reduction of plasma cholesterol by as
much as 30% in NIDDM patients (28,29,43).
GLYCEM1C RESPONSE TO CARBOHYDRATE-RICH FOODS
The beneficial effects of low-fat/high-carbohydrate di-
ets with high contents of unprocessed vegetable foods
on blood glucose and lipoprotein metabolism are usu-
ally interpreted as a consequence of the high-fiber
content of these diets. This is only partially true be-
cause other characteristics of these foods, besides fiber
content, are able to influence their glycemic response
and they might contribute to the beneficial metabolic
effects of diets rich in legumes, vegetables, and fruits
(52,53). Among these characteristics, the physical form
of the food certainly plays a major role (54). This has
been clearly shown in a study performed in NIDDM
patients that evaluated the glycemic response to identi-
cal portions of beans processed by different methods
into two physical forms (55). In one form, the integrity
of the cells was maintained, whereas in the other, the
cells were ruptured. Obviously, the composition of the
two portions of beans, including the type and amount
of dietary fiber, was identical, but, nevertheless, the
glycemic response was 50% lower in beans with un-
ruptured cells. These results emphasize the complex
nature of the factors controlling plasma glucose re-
sponse to carbohydrate-rich foods and suggest that the
glycemic and insulinemic response to a food is not
necessarily predictable on the basis of its biochemical
composition.
This concept is now supported by many studies,
including Wolever et al. (56), who tested numerous
DIABETES CARE, VOL. 14, NO. 12, DECEMBER 19911121
DIETARY
FIBER,
CARBOHYDRATE,
AND
LIPID METABOLISM
5
A Blood glucose
(mmol/l)
30
60 90
120 ISO 180
Time (min)
210 240 270 300
FIG.
7. Mean ± SE blood glucose rise above baseline
after 3 standard test meals containing a 50-g carbohy-
drate portion of either bread (O), spaghetti (•), or pota-
toes (•) in non-insulin-dependent diabetic patients (n =
7; ref. 57). *P < 0.05, **P < 0.025, ***P < 0.01 vs. spa-
ghetti.
foods for their blood glucose response in diabetic and
nondiabetic individuals (56).
In the search for carbohydrate-rich foods that, al-
though fiber-depleted, would behave as fiber-rich
foods in relation to their beneficial effects on blood
glucose, we evaluated spaghetti, breads, and potatoes
in NIDDM patients (57). They were given, in random
order and on alternate days, three test meals of identi-
cal composition. Each meal consisted of a standard
serving to which a portion containing 50 g carbohy-
drate of either bread (90 g), spaghetti (65 g), or pota-
toes (285 g) was added. The glycemic response to the
meal containing bread or potatoes was higher by 68
and 48%, respectively, compared with the meal con-
taining spaghetti, thus indicating that, although fiber-
depleted,
the latter has the same beneficial effects on
blood glucose metabolism as fiber-rich foods (Fig. 7).
This enlarges the list of the carbohydrate-rich foods
to be preferentially used in the diet for diabetic pa-
tients.
It should include not only those foods rich in
gel-forming, water-soluble dietary fiber but other foods
with slow rates of digestion and absorption of their
carbohydrate content because of the physical form.
Among them are foods prepared with whole- or partly
milled grains and seeds
(i.e.,
whole-grain bread), in
which starch is present within undisrupted plant struc-
tures.
Therefore, accessibility to hydrolysis by pancre-
atic amylase in the small bowel is limited or, at least,
delayed (58). The list should also include spaghetti and
parboiled rice in which starch accessibility is limited
because the technological process used to prepare
them has modified the physical form of starch (incom-
plete gelatinization delays starch digestion). Influenc-
ing starch digestibility by means of appropriate utiliza-
tion of food technology opens new perspectives on the
preparation of starchy foods that would be particularly
suitable for diabetic patients.
However, long-term studies are needed to verify
whether foods with a low glycemic response in acute
test meal experiments will improve blood glucose con-
trol in the long term. So far, only a few of these studies
are available, and although their results are promising,
further confirmations are required before the evalua-
tion of the glycemic response in acute test meal exper-
iments becomes a routine method for the selection of
the most suitable carbohydrate-containing foods for
diabetic patients (59).
Finally, because the beneficial metabolic effects of
unprocessed vegetable foods are only partially due to
dietary fiber, we cannot expect that purified vegetable
fiber used as a pharmacological agent to treat diabetic
patients could be as effective (and as safe) as a
high-
fiber diet. The data available are not particularly im-
pressive (30).
CONCLUSIONS
Going back to the dietary recommendations for dia-
betic patients—after having reviewed the available evi-
dence on the metabolic effects of clinically significant
dietary manipulations in diabetic patients—we con-
clude that they are wise in proposing a fiber-rich low-
fat/high-carbohydrate diet for the treatment of both
IDDM and NIDDM patients. However, it should be
better clarified that when, for whatever reason, the
patient cannot increase the consumption of unrefined
vegetable foods, he should not be allowed to substitute
these with fiber-depleted starchy foods but rather
should be advised to increase the intake of unsaturated
fat. It is also interesting to contemplate the possibility
that the reduction of saturated fat in the diabetic diet
could be compensated by a balanced increase in both
unprocessed vegetable foods and unsaturated fat to
improve the compliance to the dietary prescription
(60).
A diet based almost exclusively on vegetable,
products does not look feasible to most people in
western countries. Conversely, a moderate increase in
the consumption of fiber-rich foods, especially le-
gumes, vegetables, and fruits (increasing the daily fiber
intake up to 40 g) combined with a preferential use of
unsaturated fat would be more easily accepted. A
1122DIABETES CARE, VOL. 14, NO. 12, DECEMBER 1991
G.
RICCARDI
AND AA RIVELLESE
diet low in saturated fat and cholesterol and rich in
vegetables and olive oil resembles the type of diet
used in the Mediterranean region that is becoming
increasingly popular in many western countries.
The possibility of offering diabetic patients a
nutri-
tional model that is not artificially created in a meta-
bolic kitchen but is a life experience—with strong
historical,
cultural, and gastronomical background
would certainly facilitate compliance for millions of
people. Moreover, this dietary model is not exclusively
for diabetic patients, but is a proposal offered to the
general population, because it represents the best
available diet for the prevention of cardiovascular dis-
ease,
cancer, and many other degenerative diseases
that are becoming epidemic in the western world (61).
Therefore, the diabetic patient will not be separated
from the rest of his family in his dietary prescription
but will share with others a diet useful to preserve and
improve health.
ACKNOWLEDGMENTS
This work was supported in part by Italian National
Research Council Grant 91.00226.PF41—target project
"Prevention and Control of Disease Factors," sub-
project "Nutrition;" by the Ministry of Education and
by the University of Naples.
We gratefully acknowledge Anna Pafundi for secre-
tarial assistance and Rosanna Scala for linguistic con-
sultation.
REFERENCES
1.
Krowlewski
AS,
Warram
JH,
Rand
LI,
Kahn
CR:
Epide-
miologic approach
to the
etiology
of
type
I
diabetes
mellitus
and its
complications.
N
Engl
I Med
26:1390-
98,
1987
2.
Pyorala
K,
Loast
M,
Uusitupa
N:
Diabetes
and
athero-
sclerosis:
an
epidemiological view. Diabetes Metab Rev
3:463-524,
1987
3. Diabetes Drafting Group: Prevalence
of
small vessel
disease and large vessel disease
in
diabetic patients from
14 centres:
the WHO
multinational study
of
vascular
disease
in
diabetics. Diabetologia
28
(Suppl.):715-40,
1985
4.
Rosengren
A,
Welin
L,
Tsipogianni
A,
Wilhelmesen
L:
Impact
of
cardiovascular risk factors
on
coronary heart
disease
and
mortality among middle aged diabetic
men:
a general population study.
Br Med I
299:1127-31,
1989
5. Riccardi
G,
Rivellese
AA,
Mancini
M: The
use
of
diet
to
lower plasma cholesterol levels.
Eur
Heart
) 8
(Suppl.
E):79-85,
1987
6. American Diabetes Association: Position statement:
nu-
tritional recommendations
and
principles
for
individuals
with diabetes mellitus: 1986. Diabetes Care 10:126-32,
1987
7. Diabetes and Nutrition Study Group
of
the EASD: Nutri-
tional recommendations
for
individuals with diabetes
mellitus. Diabetes Nutr Metab 1:145-49,
1988
8. European Atherosclerosis Society: Strategy
for
preven-
tion
of
coronary heart disease:
a
policy statement
of the
European Atherosclerosis Society. Eur Heart) 8:77-88,
1987
9. American Heart Association:
AHA
special report:
rec-
ommendations
for the
treatment
of
hyperlipidemia
in
adults:
a
joint Statement
of the
Nutrition Committee
and
Council
on
Arteriosclerosis
of the
American Heart Asso-
ciation (Abstract).
Arteriosclerosis
4:445-67,
1984
10.
Keys
A:
Coronary heart disease
in
seven countries.
Cir-
culation 41 (Suppl.
1):1-211,
1970
11.
Ferro-Luzzi
A,
Strazzullo
P,
Scaccini
C,
Siani
A,
Sette
S,
Mariani
MA,
Mastranzo
P,
Dougherthy RM, lacono JM,
Mancini
M:
Changing
the
Mediterranean diet: effects
on
blood lipids. Am
)
Clin Nutr 40:1027-37,
1984
12.
Schonfeld
G,
Patsch
W,
Rudel
LL,
Nelson
C,
Epstein
M,
Olson RE: Effect
of
dietary cholesterol
and
fatty acids
on
plasma lipoproteins.) Clin Invest 69:1072-80,
1982
13.
Mensink RP,
de
Groot MJM,
van den
Broeke LT, Severij-
nen Nobels
AP,
Demaker
PNM,
Katan
M:
Effects
of
monounsaturated fatty acids
vs.
complex carbohydrates
on serum lipoproteins
and
apoproteins
in
healthy
men
and women. Metabolism 38:172-78,
1989
14.
Reaven
CM: How
high
the
carbohydrate? Diabetologia
19:409-13,
1980
15.
Garg
A,
Bonanome
A,
Grundy SM, Zhang
Z,
Unger
RH:
Comparison
of a
high carbohydrate diet with
a
high
monounsaturated
fat
diet
in
patients with non-insulin
dependent diabetes mellitus.
N
Engl) Med 319:829-34,
1988
16.
Rivellese AA, Giacco
R,
Genovese S, Patti
L,
Marotta
G,
Pacioni
D,
Annuzzi
G,
Riccardi
G:
Effects
of
changing
amount
of
carbohydrate
in
diet
on
plasma lipoproteins
and apolipoproteins
in
type
II
diabetic patients. Diabetes
Care 13:446-48,
1990
17.
Stout RW: Insulin
and
atheroma: 20-yr perspective.
Dia-
betes
Care 13:631-53,
1990
18.
National Institutes
of
Health: Consensus development
conference
on
diet
and
exercise
in
non-insulin-depen-
dent diabetes mellitus. Diabetes Care 10:639-44,
1987
19.
Abbott WGH, Boyce
VL,
Grundy
SM,
Howard
BV: Ef-
fects
of
replacing saturated
fat
with complex carbohy-
drate
in
diets
of
subjects with NIDDM. Diabetes Care
12:102-107,
1989
20.
Simpson
HCR,
Simpson
RW,
Lousley
S,
Carter
RD,
Geekie
M,
Hockaday
TDR,
Mann
Jl: A
high carbohy-
drate leguminous fibre diet improves
all
aspects
of dia-
betic control. Lancet
1:1-5,
1981
21.
Coulston
AM,
Hollenbeck
CB,
Swislocki
ALM,
Chen
Y-Di,
Reaven GM: Deleterious metabolic effects
of
high
carbohydrate, sucrose-containing diets
in
patients with
non insulin-dependent diabetes mellitus.
Am I Med 82:
213-20,
1987
22.
Brunzell
JD,
Lerner
RL,
Hazzard
WR,
Porte
D Jr,
Bier-
man
EL:
Improved glucose tolerance with high carbohy-
drate feeding
in
mild diabetes.
N
Engl
I Med
284:521
24,
1971
23.
Kolterman
CG,
Greenfield
M,
Reaven
GM,
Saekow
M,
Olefsky JM: Effects
of a
high carbohydrate diet
on
insu-
lin binding
to
adipocytes
and on
insulin action
in
vivo
in man. Diabetes 28:731-36,
1979
24.
Riccardi
G,
Rivellese
AA,
Mancini
M:
Current dietary
DIABETES CARE,
VOL. 14, NO. 12,
DECEMBER
1991
1123
DIETARY FIBER, CARBOHYDRATE, AND LIPID METABOLISM
recommendations for coronary heart disease prevention.
Diabetes Nutr Metab 1:133-37, 1988
25.
Keys A, Menotti A, Karvonen MJ, Aravanis C, Blackburn
H, Buzina