Insulinemic and glycemic indexes of six starch-rich foods taken alone and in a mixed meal by type 2 diabetics

Abstract

The glycemic index concept neglects the insulin secretion factor and has not been systematically studied during mixed meals. Six starch-rich foods were tested alone and in an isoglucido-lipido-protidic meal in 18 NIDDs and compared with a glucose challenge. These test meals were randomly assigned using a three factor experiment design. All three tests contained 50 g carbohydrate; mixed meals were adjusted to bring the same amount of fat (20 g), protein (24 g), water (300 mL), and calories (475 kcal) but not the same amount of fiber. Whatever the tested meals, foods elicited a growing glycemic index hierarchy from beans to lentils, rice, spaghetti, potato, and bread (mean range: 0.21 +/- 0.12-92 +/- 0.12, p less than 0.001). Mixing the meals significantly increased the insulinemic indexes (p less than 0.05) and introduced a positive correlation between glycemic and insulinemic indexes (n = 6, r = 0.903; p less than 0.05).
The glycemic index concept remains discriminating, even in the context of an iso-glucido-lipido-protidic meal. Insulinemic indexes do not improve discrimination between foods taken alone in type 2 diabetics: they only discriminate between foods during mixed meals, similarly to glycemic indexes.

Full text

Insulinemic and glycemic indexes of six starch-rich
588 Am I Clin Nuir 1987;45:588-95. Printed in USA. C 1987 American Society for Clinical Nutrition
foods taken alone and in a mixed meal by
type 2 diabetics1-3
Francis RJ Bornet, MD, PhD, Dominique Costagliola, PhD, Saiwa W Rizkalla, MD, PhD,
Anne Blayo, MD, Anne-Marie Fontvieille, BS, RD. Marie-Jo#{235}lle Haardt, MD,
Martine Letanoux, MD, Georges Tchobroutsky, MD, and Gerard Slama, MD
ABSTRACT The glycemic index concept neglects the insulin secretion factor and has not been
systematically studied during mixed meals. Six starch-rich foods were tested alone and in an iso-
glucido-lipido-protidic meal in 18 NIDDs and compared with a glucose challenge. These test meals
were randomly assigned using a three factor experiment design. All three tests contained 50 g car-
bohydrate; mixed meals were adjusted to bring the same amount of fat (20 g), protein (24 g), water
(300 mL), and calories (475 kcal) but not the same amount of fiber. Whatever the tested meals, foods
elicited a growing glycemic index hierarchy from beans to lentils, rice, spaghetti, potato, and bread
(mean range: 0.21 ±0.12-92 ± 0.12, p <0.001). Mixing the meals significantly increased the insu-
linemic indexes (p <0.05) and introduced apositive correlation between glycemic and insulinemic
indexes (n = 6, r=0.903; p <0.05).
The glycemic index concept remains discriminating, even in the context of an iso-glucido-lipido-
protidicmeal. Insulinemic indexes do not improve discrimination between foods taken alone in type
2 diabetics: they only discriminate between foods during mixed meals, similarlyto glycemic indexes.
Am J C/in Nuir 1987;45:588-95.
KEY WORDS Glycemic index, insulinemic index, starch-rich foods, diabetes, mixed meals
Introduction
Emphasis has been put recently on a more
liberal use of carbohydrates in the diet of di-
abetic patients (1-4) so that this type of food
now represents 50-60% ofdaily caloric intake.
In the mid 70s, Crapo et a! published their
first studies on the postprandial plasma glucose
and insulin response to different carbohydrates
(5-8). This opened a new era of research on
metabolic effects of carbohydrate and led to
the introduction by Jenkins et al of the gly-
cemic index concept (9), in which effects of
such foods were classified in relationship to
the response elicited by a glucose challenge.
Even if this index was only intended to “pro-
vide physiological data on the blood glucose
response in man . . . to a range of foods . .
to supplement tables based solely on chemical
analysis” (9) and was, in this regard, real prog-
ress, these indexes could be regarded as insuf-
ficient. As recently pointed out by Coulston
et a! (10) “a major concern is that the main
focus of these studies has been the glycemic
response rather than evaluating both plasma
glucose andinsulin response.” The concept of
glycemic index indeed neglects the insulin se-
cretion factor, which might be ofmajor clinical
significance in the nondiabetic as well as in
the type 2 (noninsulin-treated) diabetic pop-
ulation since peripheral hyperinsulinism could
be a risk factor for atherosclerosis (1 1, 12).
Moreover, most of these studies have not
shown whether the concept ofglycemic index
persists when carbohydrate foods are incor-
porated in a mixed meal, which is the most
common manner in which they are consumed.
1From the Department of Diabetes, H#{244}tel-DieuHos-
pital, the Unite de Recherches Biomath#{233}matiques et Bio-
statistiques, Paris; and the Umt#{233}de Recherches Statis-
tiques, Villejuif, France.
2Supported by a grant from Pierre et Marie Curie Uni-
versity, Paris, France.
3Address reprint requests toG#{233}rardSlama, MD, Service
de Diab#{233}tologie, H#{244}tel-DieuHospital, 1, place du Parvis
Notre-Dame, 75181 Paris Cedex 04, France.
Received April 9, 1986.
Accepted for publication August 5, 1986.
by guest on July 11, 2011www.ajcn.orgDownloaded from

STARCH-RICH FOODS IN TYPE 2 DIABETICS 589
*RefTrivelli (1 3). Normal range 4-6%.
Coulston et a! (10) have suggested that any
study on the subject should meet five criteria:
1) development of any index of metabolic re-
sponse should include both glucose and in-
sulin; 2) response to various foods should be
studied in individuals for whom dietary rec-
ommendations are proposed; 3) studies should
be performed in the context ofa standardized
test meal containing representative portions
of fat, protein, and carbohydrate; 4) interpre-
tation of this research should be directed as
an aid to minimize the postprandial hypergly-
cemia in susceptible individuals; and 5) should
explore the chronic effects of any dietary
change. We shared this outlook when we
planned to explore the acute glycemic and in-
sulinemic effects of six common starch-rich
foods, ie bread, potato, rice, spaghetti, lentils,
and beans taken alone and as part of an iso-
g1ucido-1ipidoprotidic meal by type 2, non-
insulin-treated diabetic subjects.
Subjects and methods
Subjects
Eighteen type 2 diabetic subjects(6 women and 12 men)
participated as informed volunteers in this study, which
received approval from the Hospital Ethical Committee.
All had been diagnosed as type 2 diabetic subjects on the
basis of clinical history, a fasting blood glucose value
 140 mg/100 mL, and a 2 h-postprandial value  190
mg/100 mL; all laboratoiy assays were repeated twice with
a 1 mo interval before the diagnosis was accepted. Six
subjects were treated by diet alone and 12 by oral anti-
diabetic drugs [glibenclamide (Hoechst, Puteaux, France)
and/or metformin (Aron-Medicia, Suresnes, France)J on
an outpatient basis: their clinical characteristics are given
in Table I. Therapy was kept constant for 4 mo before
the study and throughout testing.
Methods
Each subject consumed three test meals on 3 consec-
utive days: starch-rich food taken alone (meal A), the same
food in the same amount taken in the course of a meal
enriched in fat and protein (meal M), and the reference
glucose tolerance test (OGTT), taken as the third meal.
Six starch-rich foods were tested: white bread, spaghetti,
white rice, instant flaked potatoes, dried kidney beans,
and dried lentils. Each food was tested by three different
subjects. The type offood and the order ofthe three meals
(A, M, and OGT) were randomly assigned using a three
factor experiment design.
Constitution ofihe meals (Fig 1 and Table 2)
Every test meal contained 50 g available carbohydrate
(CHO) either as starch (meal A and meal M) or as glucose
(OG1’f). For meals A the fat (0.25-1.30 g/meal), protein
(4.2-20.5 g/meal), and caloric(219-290 kcal/meal) content
TABLE 1
Patient characteristics
No. Age Diabetes
duration Sex
M/F BMI Fasting pLasma
giucoic Fasting plasma
insulin HbA1
Antidiabetic drugs
Glibenclamide Metformin
yr yr kg/rn’ mg/I mL mUI/L % mg/day mg/day
I 68 9 M 27.7 123 31 8.2 12.5 -
2 47 15 F 40.5 186 23 7.7 7.5 1700
359 6 M 28.3 126 17 7.4 -1700
4 62 12 F 29.0 233 21 13.3 12.5 1700
556 1M30.5 142 29 9.1 12.5 1700
6 51 0.2 F 32.8 104 22 6.9 - -
7 32 0.2 M 32.7 151 16 9.0 .- 1700
8 62 20 M 21.8 84 7 6.5 7.5 1700
970 1 F 25.0 124 25 8.3 - -
10 59 11 M 27.5 126 20 8.2 - -
11 52 1 M 26.8 103 17 7.1 - -
12 60 13 M 22.1 198 10 10.3 12.5 850
13 52 9 M 21.0 102 16 6.6 --
14 65 25 F 30.7 114 11 8.3 --
15 62 19 M 21.9 111 18 8.2 12.5 -
16 61 2.5 M27.5 133 26 7.5 -1700
17 40 4 F 28.0 126 17 8.5 12.5 1700
18 73 10 M 27.7 163 19 7.9 12.5 -
Mean 57 8.8 12/6 27.9 136 19.2 8.3
SEM 2 1.8 1.1 9 1.5 0.5
by guest on July 11, 2011www.ajcn.orgDownloaded from

MEAL M
590 BORNET ET AL
(O,3g- I,3g)
(4,2g-20,5g)
CHO
50g
MEAL A
FIG 1. Analytical composition of meals in terms of
carbohydrate (CHO), fat, and protein.
depended on the nature of the food tested; 250 mL de-
caffeinated coffee was allowed for all meals A.
Meals M were made ofthe same amounts ofthe same
tested foods as meals A with the addition of butter and
Comt#{233}cheese in order to obtain constant CHO (50 g,
43%), fat (20 g, 37%), protein (24 g, 20%), energy (475
kcal), and total-water intake (300 mL). A variable quantity
of decaffeinated coffee was allowed in order to meet this
300 mL requirement, which ranged from zero for the po-
tato meal to 240 mL for the bread meal according to in-
trinsic water content of foods as described in reference
tables (14, 15). As indicated in Table 2, fiber content of
test meals differed from one starch-rich food to another
(meals A) and was not adjusted to be constant.
The OGT contained 50 g monodehydrated 1)-glucose
diluted in 250 mL water and was taken in a few seconds;
meals A and M were taken in 15 mm.
Test meals were taken at 0800 h after a 12 h fasting
period with patients taking their usual oral drugs when
indicated.
Processing of the foods
We used white bread (french baguette, industrially pro-
duced), long-grain white rice from Surinam, industrially
produced durum wheat flour spaghetti, dned kidney beans,
and green lentils. Rice and spaghetti were cooked in 2%
salted boiling water for 25 and 20 mm, respectively. Beans
were soaked 12 h in 2% salted water then cooked in this
water in a pressure cooker for 25 mm after rise in pressure.
Lentils were cooked without prior soaking in the same
manner for 40 mm. At the end of the cooking time no
residual water was left thus loss of solubiized starch was
avoided.
All foods were cooked in the same batch; each portion
was frozen at -20#{176}Cand thawed in a microwave oven
when needed.
Potato meals consisted of potato flakes extemporane-
ously prepared with 270 mL warm water.
Blood sampling
Blood samples were drawn 30 mm before and every
30 mm for 180 mm after the start of the meal. Samples
were immediately centrifuged and frozen at -20#{176}Cfor
later assay.
Plasma glucose was assayed using a glucose-oxidase
method (Beckman Autoanalyzer II, Beckman, Fullerton,
CA; intra-assay reproducibility is 2%). Plasma insulin was
tested by a radio immunoassay (Anti-insulin antibody,
Novo Industri, Copenhagen, Denmark) using a charcoal
separation (intra-assay reproducibility is 6%).
The mean glycemic index (0!), calculated for each pa-
tient and each test meal, is the ratio between the incre-
mental area under the 3 h glycemic-response curve to a
food and the incremental area under the 3 h glycemic-
response curve to glucose with the result multiplied by
100. The insulin index was calculated the same way from
the respective insulin curves.
Statistical analysis
We designed a three way experiment design (6 X 3 X 2)
with one case observation per cell. The analysis of variance
allowed us to test the effect of the following factors: type
of starch-rich food, type of meal (alone or mixed meal),
interaction between type of food and type of meal, and
subjects(nested within type offood factor). Multiple corn-
parisons were made using the method of Neuman and
Keuls (16). A possible association between glycemic and
insulin indexes was tested using a correlation analysis. Re-
suits are given as mean ±SEM.
Results
Figures 2a and 2b show the mean curves of
incremental plasma glucose and insulin van-
ations for each ofthe foods tested (meal A and
meal M) and the corresponding OGTT. The
comparisons between the results observed are
best characterized by glycemic and insulin in-
dexes.
Figure 3 shows the mean results of the 0!
for meal A and meal M. For meal A the GI
was very different from one type of food to
another with decreasing from bread (95
±15%) to potato (74 ± 12), spaghetti (64
±15), rice (56 ±2), lentils (30 ± 15), and beans
(23 ±1). For meal M GIs tended to be, on
average, 20% lower than for meal A and in
the same order. The differences between foods
by guest on July 11, 2011www.ajcn.orgDownloaded from

.ojjee
INGESTED CHO PROTEIN FAT CRUDE WATER ENERGY TOTAL MEAL
FOODS WEIGHT’ FIBER VALUE WATER ENERGY
TESTED (g) (g) (g) (g) (g) (g) (kcal) (g) (kcal)
*
WHITE
BREADt
Coffee
Chees4
Butte4
98 50
240 -
50 0.75
250
Coffee
i-i-- 250
32(1
(50)
8 1.17 0.88 37.50
- - - 240
15 15 -19.30
-3.32 -0.60
- - - 250
.4
POTA TOt
Coffee
Cheese
Butter
242
I98
30
a)
‘Ii
50 4.15 0.30 0.70 273.50
65 0.97 19.50 19.50 -2;.OO
250
Coffee 250
221)
RICEt (62)
Coffee Ito
Cheese 65
Butter -
219
257
Coffee
50 4.2 0.25 0.18 165
- - - - hO
0.97 19.50 19.50 -25.00
- - - - 250
219
257
SPAGHE1TIt
Coffee
Ctwese
Butter
Coffee
iii
25(1
210
(67)
130
so
250
2O)
(80)
1110
20
I5
50 8.6 0.8 0.20 150
- - - - 130
0.75 15 15 -19.25
- - 4.15 -0.78
- - - - 250
BEANSt
Coffr
Cheese
Butter
241
198
37.60
Coffee 250
iii-
50 17.3 1.3 3.25 130
- - - - 160
0.30 6 6 -7.20
- - 12.45 -2.30
- - - - 250
LENTIISt
Coffee
Cheec
Butter
280
79.20
112.80
225
(83)
14U
12
18
a)
a) ca
50 20.5 0.9 3.25 153 290
- - - - 140 -
0.18 3.60 3.60 -4.60 47.50
- - 14.90 -2.80 135.50
STARCH-RICH FOODS IN TYPE 2 DIABETICS
TABLE 2
Meal Composition
*The figures represent cooked foods. Dry weight is given in brackets.
tAccording to the Scientific Ciba Geigy Table (14).
tAccording to Ostrowsky’s Table (15).
591
were significant (p <0.001), but the differences
between meal A and meal M were not. mdi-
vidual comparisons showed that the bread GI
was significantly different from that of spagh-
etti (p <0.001), rice (p <0.01), lentils (p
<0.001), and beans (p <0.001). Potato 0!
was significantly different from that of lentils
(p <0.01) and beans (p <0.01); rice 0! was
significantly different from that of beans (p
<0.05).
Figure 4shows the mean insulinemic index
result for each food tested alone and during a
mixed meal. For meal A the insulin indexes
were almost identical, ‘95%, for every food
tested except bread (178 ± 3 1%). However,
the difference was not significant.
Mixing the meals significantly increased in-
sulin secretion (p <0.05) but also reintroduced
the hierarchy observed for the glycemic in-
dexes: insulin index for bread was 278 ± 18%;
for potato 274 ±164; for spaghetti 172 ±38;
forrice 126 ± 16; forlentils 103 ± 37; and for
beans69± 13.
Figure 5 shows the correlation between the
mean glycemic indexes and mean insulinemic
indexes of meals A and M: a significant car-
relation was only found for meal M (r =0.903,
p <0.05).
by guest on July 11, 2011www.ajcn.orgDownloaded from

150
0
0
I-
C)
E
w
(I)
0
C.)
0
(0
E
0.
C
0
E
C)
0
C
BEAD
SPA4
Time (minutes)
FIG 2A and 2B. Mean incremental plasma glucose and insulin variations observed during 180 mm following OGTT
(#{149}),meal A (food taken alone)(O), and meal M (mixed mea1)(i) in three type 2 diabetic patients for each food tested
(Fig 2a: bread, rice, and spaghetti; Fig 2b: potato, lentils, and beans). Mean ± SEM.
Time (minutes)
592 BORNET ET AL
iJ
 
C
#{149}inrJ
E_______
Discussion
Glycemic indexes have been proposed as a
way of classifying CHO containing foods ac-
cording to blood glucose responses. The
methodology used to calculate such indexes
has varied in many ways: the populations
tested have been normal and/or type 2 diabetic
subjects; they have been treated either by diet
alone, diet and oral antidiabetic agents, or by
insulin; the reference carbohydrate chosen also
has varied in terms of quality (glucose or white
FIG 3. Mean glycemic index calculated for the six foods
tested alone (meal A; black bars) and in a mixed meal
(meal M, white bars) in type 2 diabetic patients. n= 3;
mean ±SEM.
bread) (9, 17) and quantity (25 or 50 g) (9,
18). Moreover, and as pointed out by Coulston
et al (10), the concept of glycemic index ne-
glects the insulin secretion aspect and has not
yet been systematically studied in the usual
way in which foods are consumed, which is in
mixed meals. Even if they did not calculate
glycemic and insulinemic indexes, Coulston
et al (19, 20), while studying combinations of
j1jJ
FIG 4. Mean insulinemic index calculated for the six
foods tested alone (meal A; black bars) and in a mixed
meal (meal M, white bars) in type 2 diabetics. n= 3; mean
±SEM.
by guest on July 11, 2011www.ajcn.orgDownloaded from

MEAN GLYCEMUC INDEX (%)
STARCH-RICH FOODS IN TYPE 2 DIABETICS 593
><
w
z
Cu)
w
z
-J
U)
z
z
Ui
FIG 5. Correlation between the mean glycemic indexes
(x axis) and insulinemic indexes(y axis) observed for food
tested alone [meal A (#{149})jor in a mixed meal [meal M
(0)]. Each point represents the mean ±SEM ofthree test
meals. For meal A: r=0.675, NS; for meal B: r= 0.903,
p<0.05; y =3.05x +0.33; Br =bread, P = potato, R
=rice, S =spaghetti, Be =beans, and L =lentils.
foods, showed that substitution ofone starchy
food for another in the course ofa mixed meal
elicited different glucose responses in glucose-
intolerant subjects and different insulin re-
sponses in normal subjects. These results have
not been observed in type 2 diabetic subjects
(21-23). In these lafter studies the lack of in-
fluence may be due to the type ofmeals tested:
amixing ofCHO foods with different GIs but
with similar mean calculated GIs (22).
This study has attempted to calculate for
type 2 diabetic subjects the glycemic and in-
sulinemic indexes of CHOs taken alone and
during mixed meals.
Data concerning glycemic indexes
Our results are in good accordance with
those of other authors (9, 24, 25) who stressed
that foods that contain complex carbohydrates
as well as foods that contain simple sugars (26,
27) are, under acute conditions, sometimes
highly hyperglycemic; some foods may have
one fourth the glycemic effect of others. The
differences observed may be due to various
factors. For example it is well known that
mixing meals reduces glycemic responses (28,
29). We observed a trend towards a reduction
of 15-20%, which was not statistically sig-
nificant, even though fat and protein were at
levels of 37 and 20%, respectively. The ob-
served lack ofsignificant effect ofmixed meals
in our study might be due to the small number
of subjects tested for each food. In our study,
hierarchy of the GIs observed for foods taken
alone persists with our mixed meals, which
confirms the conclusions ofWolever et a! (30)
and demonstrates that the intrinsic nature of
the starchy food seems to be a predominant
factor. The differences observed cannot be due
to the fat, protein, and water content, as these
three factors were kept constant throughout
the mixed meals. When fat is added to a CHO
meal (31), it reduces glycemic response prob-
ably by slowing gastric emptying rather than
by modifying starch availability (32).
The GI differences observed may be par-
tially due to fiber content of the meals, as the
foods with the high-fiber content (beans, len-
tils; Table 2) have the lower 0!. However, de-
spite comparable fiber content, GIs may vary
considerably. The variable susceptibility of
starches or starchy foods to amylase has also
been shown in vitro (33-36). Furthermore,
starch legumes have a high-amylose content
(37), and the high-amylose content of starch
is a determinant factor ofsusceptibility to hy-
drolysis (31). Rice with high-amylose content
elicits a reduced glycemic response in corn-
parison with low-amylose rice (38).
In addition mechanical barriers such as
protein matrix (ie, gluten) that encapsulates
gelatinized starch granules limit access to am-
ylase and reduce the starch availability (39).
The high-gluten content of spaghetti may ex-
plain its lower 0! than that of another wheat
product such as bread.
Data concerning the insulinemic indexes
Our study has shown that, with the excep-
tion of bread, all the foods tested have ap-
proximately the same insulinogenic index
(-95%) when taken alone. In opposition with
our findings and in a comparable work on
foods ingested alone in impaired glucose to!-
erance patients (7) and in type 2diabetics (8),
by guest on July 11, 2011www.ajcn.orgDownloaded from

594 BORNET ET AL
Crapo et al showed that insulin secretion and,
therefore, insulinemic indexes vary from one
food to another in the same direction as the
glycemic response. In addition (as can be
extrapolated from their data), insulin indexes
were comparable in normal (6), impaired glu-
cose tolerance (7), and type 2diabetic (8) sub-
jects. The differences observed with our study
might be due to differences in severity of di-
abetes, endogenous insulin secretion capaci-
ties, and/or the use of oral drugs in our work.
The results observed in our study with bread
meals differ consistently from those of other
authors for glycemic response [95% vs 85% for
Crapo et a! (8) and 69% for Jenkins et al (9)]
and even more so for insulinemic response
[173% in our study vs ‘-85% for Crapo et a!
(8)]. These discrepancies may be due to dif-
ferences ofwheat origin and bread processing.
Whatever the cause, this invalidates the choice
of bread as an internationally accepted stan-
dard for 0! calculation. This is also true for
many other foods when discrepancies have
been found. Glucose, although not a real food,
is the best-known reference CHO.
During the mixed meals, insulin secretion
increased and the same hierarchy for insulin-
emic indexes as for GIs appeared with a pos-
itive correlation between both. The addition
ofproteins to CHOs potentiates insulin secre-
tion in both healthy (40) and diabetic subjects
(41, 42) (Fig 5). In mixed meals when protein
content is kept constant, the insulinemic index
correlates with the glycemic index (Fig 5).
Thus the synergistic effect of proteins and
CHOs is all the greater when the glycemic in-
dcx is high. This phenomenon seems to re-
main valid for some foods taken alone when
they are themselves rich in protein (lentils and
beans).
At this point we may ponder on the rele-
vance of an insulinemic index for choosing a
CHO food. As Coulston et a!, we postulated
that such indexes may have been of help in
classifying CHO foods. In fact our results show
that, at least in the population studied and with
the six foods tested, insulinemic indexes do
not seem to contribute any supplementary
data for classification: there is no goodfood in
terms of GI which becomes bad in terms of
insulinemic index (as far as a high insulinemic
index can be regarded as pejorative) and vice
versa. Studies on insulinemic indexes may
have some importance in the understanding
of physiological events in nutrition.
The conclusions of our study, which is one
of the first to compare simultaneously gly-
cemic and insulinemic indexes offoods taken
alone and in a mixed meal with constant CHO,
fat, protein, and water content, are the fol-
lowing:
1) The GI concept remains discriminating
even in the context of a mixed meal in type 2
diabetics, which validates the use of GI for
choosing foods even in mixed meals.
2) The insulinemic index does not bring
greater discrimination between CHO-foods
but remains ofmterest in physiological studies.
3) Bread should not be taken as the refer-
ence food for calculating 0! ifwe want to have
results that are ofgeneral significance. fl
We thank Mrs Josette Boillot, Mrs Nelly Desplanque,
Mrs Annie Chevalier, and Mrs Madeleine Bros for their
technical assistance and Mrs Annie Chevalier for her
graphical skill.
References
1. American Diabetes Association. Principles of nutrition
and dietary recommendations for individuals with di-
abetes mdllitus. Diabetes 1979;28:1027-30.
2. Mann JI. Diet and diabetes. Diabetologia 1980;18:
89-95.
3. Canadian Diabetes Association. Guidelines for the
nutritional management of diabetes mellitus. J Can
Diet Assoc 198042:110-8.
4. Nutrition Sub-committee of the Medical Advisory
Committee ofthe British Diabetic Association. Dietary
recommendations for diabetics for the 1980s. In:
Cudworth AG, Andreani D, Bodansky Hi, Squaduto
G, eds. Diabetes and its late complications. London:
John Libbey, 1981:156-67.
5. Crapo PA, Reaven 0, Olefsky J. Plasma glucose and
insulin responses to orally administered simple and
complex carbohydrates. Diabetes 1976;25:74 1-7.
6. Crapo PA, Reaven 0, Olefsky J. Postprandial plasma
glucose and insulin responses to different complex
carbohydrates. Diabetes 1977;26:1178-83.
7. Crapo PA, Kolterman OG, Waldeck N, Reaven GM,
Olefsky JM. Postprandial hormonal responses to dif-
ferent types of complex carbohydrate in individuals
with impaired glucose tolerance. Am J Clin Nutr
198033:1723-8.
8. Crapo PA, Insel J, Sperling M, Kolterman 0G. Corn-
parison of serum glucose, insulin, and glucagon re-
sponses to different types ofcomplex carbohydrate in
noninsulin dependent diabetic patients. Am J Clin
Nutr 198 l;34:184-90.
9. Jenkins DJA, Wolever TMS, Taylor RH, et al. Gly-
cemic index of foods: a physiological basis for car-
bohydrate exchange. Am J din Nutr 198 1;34:
362-6.
by guest on July 11, 2011www.ajcn.orgDownloaded from

STARCH-RICH FOODS IN TYPE 2 DIABETICS 595
10. Coulston AM, Hollenbeck CB, Reaven GM. Utility
ofstudies measuring glucose and insulin responses to
various carbohydrate-containing foods. Am J Gin
Nutr 1984;39: 163-5.
1 1.Stout RW. Diabetes and atherosclerosis the role of
insulin. Diabetologia 1979; 16:141-50.
12. Ducimetiere P, Eschwege E, Papoz L, Richard JL,
Claude JR, Rosselin G. Relationship ofplasma insulin
levels to the incidence of myocardial infarction and
coronary disease mortality in middle-aged population.
Diabetologia 1980;19:205-1O.
13. Trivelli LA, Ranney HM, Lai HG. Hemoglobin corn-
ponents in patients with diabetes mellitus. N Engi J
Med 1971;284:353-7.
14. Documenta Geigy. Tables scientifiques. 7 J#{149}Basic:
Ciba-Geigy, 1972.
15. Ostrowsky ZL, Josse M-CH. Les a1iments Tables des
valeurs nutritives. Paris: Jacques Lanore, 1978.
16. Winer BJ. Statistical principles in experimental design.
New York: McGraw Hill, 1971.
17. Jenkins DJA, Wolever TMS, Jenkins AL, et al. The
glycemic index of foods tested in diabetic patients: a
new basis for carbohydrate exchange favouring the
use oflegumes. Diabetologia 1983;24:257-64.
18. lonescu-Tirgoviste C, Popa E, Sintu E, Mihalache N,
Cheta D, Mincu I. Blood glucose and plasma insulin
responses to various carbohydrates in type 2 (non is-
ulin dependent) diabetes. Diabetologia 1983;24:
80-4.
19. Coulston A, Greenfield M, Kraemer F, Tobey T,
Reaven G. Effect of source of dietary carbohydrate
on plasma glucose and insulin responses to test meals
in normal subjects. Am J Clin Nutr 1980;33:1279-
82.
20. Coulston AM, Greenfleld MS, Kraemer FB, Tobey
TA, Reaven GM. Effect of differences in source of
dietary carbohydrate on plasma glucose and insulin
responses to meals in patients with impaired carbo-
hydrate tolerance. Am J Clin Nutr 198 1;34:27 16-20.
2 1.Bantle JP, Lame DC, Castle GW, Thomas JW, Hoog-
werf BJ, Goetz FC. Postprandial glucose and insulin
responses to meals containing different carbohydrates
in normal and diabetic subjects. N Engi JMed
1983;309:7-12.
22. Nuttall FQ, Mooradian AD, DeMarais R, Parker S.
The glycemic effect ofdifferent meals approximately
isocaloric and similar in protein, carbohydrate and
fat content as calculated using the ADA exchange lists.
Diabetes Care 1983;6:432-5.
23. Coulston AM, Hollenbeck CB, Liu GC, et al. Effect
ofsource ofdietary carbohydrate on plasma glucose,
insulin and gastric inhibitory polypeptide responses
to test meals in subjects with noninsulin-dependent
diabetes mellitus. Am J Clin Nutr 1984;40:965-70.
24. Spaethe R, Brinck UC, Sabin J, W#{252}bbensD, Otto H.
Echange des hydrates de carbone d’apr#{232}sle principe
des equivalents biologiques dans Ic r#{233}gimepour
diab#{233}tiques.In: Journ#{233}esAnnuelles de Diab#{233}tologie
de I’H#{244}tel-Dieude Paris. Paris: Flammarion M#{233}decine
Sciences, 1972:253-9.
25. Jenkins DJA, Wolever TMS, Jenkins AL, Lee R,
Wong OS, Josse R. Glycemic response to wheat prod-
ucts. Reduced response to pasta but no effect of fiber.
Diabetes Care 1983;6:155-9.
26. Slama 0, Jean-Joseph P. Goicolea I, et al. Sucrose
taken during a mixed meal has no additional hyper-
glycemic action over isocaloric amounts of starch in
well-controlled diabetics. Lancet 1984;2: 122-5.
27. Bornet F, Haardt MJ, Costagliola D, Blayo A, Slama
G. Sucrose or honey at breakfast have no additional
acute hyperglycemic effect over an isogiucidic amount
of bread in type 2diabetic patients. Diabetologia
1985;28:2 13-7.
28. Mirouze J, Bernard R, Samy C, Satingber A. L’hy-
perglyc#{233}miealimentaire dans le diab#{232}tesucr#{233}(#{233}tude
par enregistrement continu). Path Biol 1965;13:152.
29. Estrich D, Ravnik A, SChIiCrfG, Fukayama 0, Kinsdll
L. Effects ofco-ingestion offat and protein upon car-
bohydrate-induced hyperglycemia Diabetes 1967;16:
232-7.
30. Wolever TM, Nuttall FQ, Lee R, et al. Prediction of
the relative blood glucose response of mixed meals
using the white bread glycemic index. Diabetes Care
1985;8:418-28.
31. CollierG, McLean A, O’Dea K. Effect of co-ingestion
offat on the metabolic responses to slowly and rapidly
absorbed carbohydrates. Diabetologia 1984;26:50-4.
32. Wong OS, Traianedes K, O’Dea K. Factors affecting
the rate ofhydrolysis ofstarch in legumes. Am J Clin
Nutr 1985;42:38-43.
33. Sandstedt RM, Strahan D, Ueda S, Abbot RC. The
digestibility of high-amylose corn starches compared
to that of other starches. The apparent effect on the
ae gene on the susceptibility to amylase action. Cereal
Chem 1962;39: 123-3 1.
34. Jenkins DJA, Wolever TMS, Taylor RH, et al. Rate
of digestion of foods and posiprandial glycaemia in
normal and diabetic subjects. Br Med J 1980;2:14-7.
35. O’Dea K, Snow P, Nestel P. Rate ofstarch hydrolysis
in vitro as a predictor ofmetabolic responses to corn-
plex carbohydrate in vivo. Am J Clin Nutr 198 1;34:
199 1-3.
36. Snow P, O’Dea K. Factors affecting the rate of hy-
drolysis of starch in food. Am J Gin Nutr 198 1;34:
272 1-7.
37. Sosuiski F, Garatt MD, Slinkard AE. Functional
properties of ten legume flours. Can Inst Food Sd
Technol J 1976;9:66-9.
38. GOddard MS, Young 0, Marcus R. The effect of
amylose content on insulin and glucose responses to
ingested rice. Am J am Nutr 1984;39:388-92.
39. Anderson IH, Levine AS, Levitt MD. Incomplete ab-
sorption ofthe carbohydrate in all purpose wheat flour.
N Engi JMed 1981;304:891-2.
40. Rabinowitz D, Merimee TJ, Maffczzoli R, Burgess
JA. Patterns ofhormonal release after glucose, protein
and glucose plus protein. Lancet 1966;2:454-6.
41. Berger S, Vongaraya. Insulin response to ingested
protein in diabetes. Diabetes 1966;15:303-6.
42. Cohn C, Berger S, Norton M. Relationship between
meal size and frequency and plasma insulin response
in man. Diabetes 1968;17:72-5.
by guest on July 11, 2011www.ajcn.orgDownloaded from