ARTICLE IN PRESS

JOURNAL OF
FOOD COMPOSITION
AND ANALYSIS

Journal of Food Composition and Analysis 17 (2004) 331338

www.elsevier.com/locate/jfca

Original Article

Measurement of carbohydrate components and their impact

on energy value of foods
Elizabete W. Menezes*, Alexandra T. de Melo, Giselli H. Lima, Franco M. Lajolo
Departamento de Alimentos e Nutri@ao
* Experimental, Faculdade de Ci#encias Farmac#euticas, Universidade de Sao
* Paulo.
Bl.14 Av. Prof. Lineu Prestes 580, CEP 05508-900, Sao
* Paulo-SP, Brazil
Received 25 June 2003; received in revised form 9 March 2004; accepted 11 March 2004

Abstract
In order to expand data on the Brazilian Food Composition Database Web site (TBCA-USP), the
present work aimed to complement information about carbohydrate content in foods as well as to evaluate
the inuence of different carbohydrates on energy value. The sugars (glucose, fructose and sucrose) and the
indigestible fractions (IF) (total, soluble and insoluble) were quantied in 11 starchy foods usually
consumed by the Brazilian population. The total IF content found was bigger than the total dietary ber
(DF) for all foods. Consequently, when the IF was subtracted on the estimate of available carbohydrate
by difference, it resulted in lower values. Similar results were found in relation to the energy values. The
method applied to the IF analysis is relatively simple and considerably reproducible, representing an
alternative for the analysis of indigestible components. However, this method does not discriminate each
component individually. The total sugar content was lower than 0.5% (fresh wt.) for all foods. The
available carbohydrate content and energy amount were similar in all 11 foods either when
carbohydrates were analyzed individually or by difference subtracting DF in the difference. However,
when the carbohydrates are analyzed individually, they provide information about several carbohydrates
that are differently utilized in the gastrointestinal tract. Efforts must be driven towards a better
differentiation among the different carbohydrate fractions, aiming to insure the quality of carbohydrate
and energy information on food composition databases.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Indigestible fraction; Sugar; Energy; Available carbohydrates

*Corresponding author. Fax: +55-11-3815-4410.

Email-address: wenzelde@usp.br (E.W. Menezes).
0889-1575/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfca.2004.03.018

ARTICLE IN PRESS
332

E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

1. Introduction
The carbohydrates of food have a high degree of acceptability in human diets. Different types
of carbohydrates are associated with several positive physiological effects for human body (FAO/
WHO, 1998; Wahlqvist, 2002). Most food composition databases show total carbohydrates data
measured by difference, which does not specify each carbohydrate component. It also must be
emphasized that the estimate of the food energy value might suffer variation rstly because of the
several components that are part of the carbohydrate fraction, then for the different terminologies
utilized to express its content in food, which can result in ambiguous and misinterpreted
information (Monro and Burlingame, 1996; Burlingame, 1996).
In order to expand food composition databases, initiatives to quantify carbohydrates in an
individual and detailed way have been taken (Menezes and Lajolo, 2000; Rosin et al., 2002; Li
et al., 2002). Li et al. (2002) evaluated the individual sugars, soluble and insoluble dietary ber
(DF) contents of 70 high-consumption American foods. Rosin et al. (2002) have accomplished the
partial characterizations of 11 Brazilian starchy foods. In this work, the total starch, resistant
starch, rapidly and slowly digestible starch, amylose and soluble and insoluble DF contents were
determined in the same 11 common starchy Brazilian foods. The analysis of other carbohydrates,
such as mono- and disaccharides and indigestible fraction (IF) of foods, seems pertinent because
of the lack of national information on those components.
The IF is dened as the part of plant foods that is not digested nor absorbed in the small
intestine. The IF is a substrate for fermentative microora when it reaches the colon, and it is
mainly composed not only of DF, but also of resistant starch and resistant protein (Saura-Calixto
et al., 2000). Other components that are also resistant to digestive enzymes such as polyphenols
and other associated components are included in IFs composition as well. In other words, the IF
is composed of the main substrates for the fermentative microora with important physiological
effects and it is always expected to be greater than DF content.
Hence, this study was carried out to determine the IF and mono- and disaccharides contents in
11 Brazilian starchy foods (used by Rosin et al., 2002). Also, the inuence of different
carbohydrate components on energy value was assessed in these foods.

2. Materials and methods

In the present work, the same 11 foods and conditions previously utilized by Rosin et al. (2002)
were adopted, aiming to complement information on the chemical composition of those foods.
The 11 test foods usually consumed by the Brazilian population were: polished and whole rice
from Supra Alimentos S.A; corn meal, dried peas, dried lentils and dried chick-peas from Hikari
Ind. Com. Ltda; beans from Broto Legal Com. Imp. Exp. Ltda; white spaghetti from Quaker
Brasil Ltda; green corn, white bread and potatoes from local market. A sample of each food
(0.51 kg) was bought in Sao Paulo-SP. The foods were cooked during enough time for
consumption and with open kettle. The time of cooking of each food was the following: polished
and whole rice (28 min), green corn (45 min), corn meal (90 min), white spaghetti (20 min),
potatoes (33 min), peas (30 min), beans (36 min), lentils (21 min) and chick-peas (70 min). After
the removal of a portion from each sample for moisture determination, the samples were dried

ARTICLE IN PRESS
E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

333

under 60
C for 18 h, milled to a particle size of o0.250 mm and stored (at ambient temperature)
until the analysis procedes, that means, for approximately 30 days. The white bread was subjected
to the same process from the drying procedure. The foods identied with an asterisk above were
previously soaked in water under 4
C for 18 h prior to cooking.
The IF was determined as the sum of insoluble IF and soluble IF according to a process
previously described by Saura-Calixto et al. (2000). Samples (300 mg of dry food) were
homogenized in 10 mL of 0.1 M HCl-KCl buffer under controlled conditions and incubated
(40
C, 60 min, pH 1.5) with pepsin (0.1 mL (30 mg/mL), Sigma P-7012) in a water bath with
constant shaking. Then, 9 mL of 0.1 m TRIS-Maleato buffer (pH 6.9) were added and the pH was
checked. a-Amylase was added (1 mL (120 mg/mL), Sigma A-3176), and the samples were
incubated in water bath (37
C, 16 h). The samples were centrifuged (3000 g, 15 min) and the
supernatants were removed. The residues were washed twice with 10 mL of distilled water and all
supernatants were combined. The residues were dried (105
C, overnight) and quantied
gravimetrically as insoluble IF. Supernatants were dialyzed against water (25
C, 48 h) (Dialyses
membrane, 12,00014,000 MWCO, Spectra/Pors Regenerated Cellulose, Spectrums Laboratories Inc., USA). Dialysates were freeze-dried, and the soluble IF was quantied gravimetrically.
Sugars were extracted successively with three portions of boiling 80% (v/v) aqueous ethanol.
The supernatants were combined and the ethanol was evaporated under vacuum. The mono- and
disaccharides contents were analyzed by high-performance liquid chromatography (HPLC-PAD)
(Dionex, Sunnyvale, CA), using a PA1 column (Dionex) in an isocratic run, with 18 mm NaOH
(25 min).
Certain components, such as total starch, resistant starch, DF and the proximate composition
were not tested as part of this study. Results were obtained from previous study (Rosin et al.,
2002) for comparison and energy calculation purpose. The analytical methods used were the
following: total starch (Goni et al., 1997), resistant starch (Goni et al., 1996), DF (Prosky et al.,
1988). The data on protein, lipids and ash were calculated from the Brazilian Food Composition
Database Web site (USP, 1998). The moisture was determined by drying in a vacuum oven for
3.5 h (60
C).
Results were expressed by means of values7standard deviations of three separate determinations. Linear regression and P-value for correlation were estimated using the Minitab-Windows.

3. Results and discussion

3.1. Indigestible fraction (IF) in starchy foods
The IF content of 11 Brazilian starchy foods was determined. The results were compared to
literature and to other DF analyses previously done on those foods.
Table 1 shows the total IF, insoluble IF and soluble IF contents of 11 foods and the literature
data. The total IF (dry wt.) of legumes ranged between 26.40% (lentils) and 36.09% (beans). For
the other foods, the total IF ranged between 7.74% (white spaghetti) and 25.43% (corn).
Comparing such results with the one obtained by Saura-Calixto et al. (2000), the variations
observed are justied by the difference in food type, food process and food preparation. For foods
in which comparison was possible, the linear regression analysis between the total IF results found

ARTICLE IN PRESS
334

E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

Table 1
Indigestible fraction (IF) content of cooked starchy foods, expressed as % of dry and fresh weight
Food

Indigestible fraction

Whole rice
Polished rice
Corn
Corn meal
White spaghetti
White bread
Potatoes
Peas
Beans
Lentils
Chick-peas

Total
(% dry wt.)

Total
(% dry wt.)

Total
(% fresh wt.)

Insoluble
(% fresh wt.)

Soluble
(% fresh wt.)

12.1370.83

11.6170.11
14.0170.82
13.8070.45
12.1870.89
42.3073.58
35.5970.51
29.157.047
28.0370.92

17.6970.11
13.3871.28
25.4370.40
13.4870.27
7.7470.54
11.2470.76
16.7070.49
27.3970.18
36.0971.12
26.407.062
32.0071.55

4.3470.03
4.5870.44
5.9670.09
2.6570.05
2.8070.19
8.4170.57
2.4970.07
9.2070.06
13.5470.42
8.5370.20
13.0670.64

3.3870.12
3.6670.40
4.8570.16
2.1770.07
1.1870.09
3.0170.03
1.9370.05
5.9070.19
10.6770.23
7.4970.11
8.4770.43

0.9670.09
0.9270.07
1.1170.09
0.4870.05
1.6270.12
5.4070.54
0.5670.03
3.3170.16
2.8970.29
1.0370.10
4.5970.21

Mean of values7s.d (standard deviation) (n 3).

 Results obtained by Saura-Calixto et al. (2000).

40
30
25
20
15
y = 0.736x + 4.240
r=0.858, p<0.003, n=9

10
5
0
0

(A)

10

20
30
IF** (% dry wt.)

40

IF*(% dry wt.)

IF*(% dry wt.)

35

40
35
30
25
20
15
10
5
0

y = 0.895x + 3.810
r=0.971, p<0.005, n=6
0

50

(B)

10

20
30
IF** (% dry wt.)

40

Fig. 1. Comparison between the total IF of cooked foods obtained by the present study () and IF results obtained by
Saura-Calixto et al. (2000) (). The following foods were included in the linear regression analysis: Apolished rice,
corn meal, white spaghetti, white bread, potatoes, peas, beans, lentils, chick-peas; Bpolished rice, corn meal, potatoes,
beans, lentils, chick-peas.

(dry wt.) and those from Saura-Calixto et al. (2000) showed a signicant positive correlation
(y 0:736x 4:240; r 0:858; p0.003, n 9). When peas, bread and white spaghetti were
excluded from the correlation, because of the differences above described, an increase in the r
value was found (y 0:895x 3:810; r 0:971; Pp0:005; n 6) (Fig. 1).
For the 11 test foods, the total IF content was bigger than the content of total DF, which
ranged from 13% to 64% of the IF content (Fig. 2).
The linear regression analysis showed a signicant positive correlation between total IF
and total DF (y 1:250x 8:610; n 0:943; Pp0:001; n 11) and between insoluble IF and
insoluble DF (y 1:230x 4:820; r 0:959; Pp0:001; n 11). No correlation was found
between soluble IF and soluble DF. It can be explained by the fact that soluble DF was quantied
by Prosky et al. (1988) method, which uses a high temperature and includes precipitation with
ethanol. On the other hand, Saura-Calixto et al. (2000) observed correlation between their soluble

ARTICLE IN PRESS
E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

335

16

Total dietary fiber

14

14

Indigestible fraction

13

Fresh wt. (%)

12
10

8
6
6

3
1

3
1

hi
ck
-p
ea
s
C

Le
nt
ils

Be
an
s

Pe
as

es
to
ta
Po

hi
te

br
ea
d

he
tti
W

hi
te

or
n
W

sp
ag

M
ea
l

or
n
C

lis
he
d
Po

ho
l

ri
c

ri
ce

Fig. 2. DF and IF contents of 11 cooked starch foods (% fresh wt.).

Table 2
Mono- and disaccharides contents of cooked starchy foods (mg/100 g fresh weight)
Food

Glucose (mg/100 g fresh wt.) Fructose (mg/100 g fresh wt.) Sucrose (mg/100 g fresh wt.) Total

Whole rice
Polished rice
Corn
Corn meal
White spaghetti
White bread
Potatoes
Peas
Beans
Lentils
Chick-peas

35.0270.90
3.0770.34
25.2672.66
38.4271.93
15.5772.17
26.1170.76
36.1372.73
n.d.
3.6270.84
5.7870.01
3.8270.91

13.7370.54
4.6070.01
54.4175.31
17.8670.02
12.1871.26
55.9070.60
23.1372.01
3.7770.44
8.3070.98
13.9770.27
6.7470.62

51.8472.16
9.3971.32
33.0372.64
89.8373.96
39.5375.35
n.d.
125.55710.94
191.19711.08
368.2777.97
145.8075.71
359.60752.17

100.59
17.05
112.70
146.10
67.29
82.01
184.81
194.96
380.19
165.55
370.16

Mean of values7s.d. (standard deviation), (n 3).

n.d.: non-detected.

IF and soluble DF data. In this case, the soluble DF was quantied by the modied method of
Prosky et al. (1988) (Manas et al., 1994), where the precipitation with ethanol is substituted for
dialysis, which is also utilized in IF determination.
The IF analysis represents an alternative for the evaluation of the fraction of foods that is not
digested by the human organism. The methodology proposed by Saura-Calixto et al. (2000) for IF
determination is relatively simple and with good reproducibility; however, it does not individually
discriminate the IF components that have different physiological action.
3.2. Sugar
Data on mono- and disaccharides contents in national foods are scarce. In Table 2 there are
data on the glucose, fructose and sucrose contents of the test foods. The contents (fresh wt.)
ranged between 3.07 and 38.42 mg/100 g for glucose, 3.77 and 55.90 mg/100 g for fructose and 9.39

ARTICLE IN PRESS
336

E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

and 368.27 mg/100 g for sucrose. Because these are starchy foods, with 42.3287.36% (dry wt.) of
total starch (Rosin et al., 2002), the low values of total sugar were already expected (o0.5% fresh
wt.). The results obtained in this work are close to the ones found by Li et al. (2002) for the
majority of foods. It is worth emphasizing that, although the test foods had low sugar content, it
is still important to know its value once they belong to the carbohydrate fraction. In this way,
such information complements data, following the present tendency which is to determine each
component of the carbohydrate fraction and to evaluate foods on a case-by-case (Monro and
Burlingame, 1996; Li et al., 2002). At the same time, this information should be included in food
composition databases to improve the food planning in order to control the glycemic response as
well as the chronic non-infectious diseases (Miller et al., 1995; Jenkins et al., 2002).

3.3. Available carbohydrate and energy value

The food composition databases present a considerable variation among the carbohydrates
content. Such variation is partly due to the terminology applied to express carbohydrates and to
the components that are considered in their estimation. The carbohydrates have a great potential
for being presented in an ambiguous way because of their wide number of components and the
several analytical methods that can be used to analyze them. Monro and Burlingame (1996)
identied more than 80 tagnames to express carbohydrates of food. When the aggregation of
proximate components is done, the carbohydrates are usually referred simply as carbohydrates.
However, there are at least ve commonly used methods for measuring and/or expressing
carbohydrates, each of which is ofcial although would reach different values for the same food.
In the present work, an estimation of the available carbohydrates was made considering the
following components for the calculation: IF (Table 1), glucose, fructose and sucrose (Table 2),
total starch, resistant starch and DF (Rosin et al., 2002) and other macronutrients (USP, 1998).
Table 3 presents the available carbohydrate content estimated by three different ways (A, B
and C) and their respective values of energy (Ae, Be, Ce). Procedure A expresses available
carbohydrate by summation of different fractions (glucose, fructose, sucrose and total starch
minus resistant starch). Procedure B was obtained by difference subtracting total DF (100 g minus
the sum of grams of water, protein, lipids, ash and DF). Procedure C was also obtained by
difference but, in this case, total IF was subtracted instead of DF. It could be noted that there was
no relevant difference between the available carbohydrate content estimated by procedure A,
which is the summation of available carbohydrate, and procedure B, where DF is also
subtracted in the difference. The same was noted in their respective energy values.
However, when the IF was subtracted in the difference on the estimate of the available
carbohydrate (procedure C), lower values were obtained in nine out of the 11 test foods, when
comparing to the estimated values found by procedures A and B. It happened because the IF
includes other components besides RS and the DF. Therefore, the values of available
carbohydrates obtained when subtracting IF are closer to the real ones than when subtracting DF
in the difference. In the case of energy, the same results were observed. The decrease in the
available carbohydrate observed in procedure C leads to a decrease in the energy value (Ce),
comparing to procedures A and B. For example, the legumes showed an overestimation on the
energy value of 16% (procedure Ae) and 20% (procedure Be).

ARTICLE IN PRESS
E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

337

Table 3
Available carbohydrate estimated by different procedures and energy values of cooked starchy foods
Food

Whole rice
Polished rice
Corn
Corn meal
White spaghetti
White bread
Potatoes
Peas
Beans
Lentils
Chick-peas

Available carbohydrate (mg/100 fresh wt.)

Energy (kJ/100 fresh wt.)

Ae

Be

Ce

19.87
29.11
18.98
13.80
30.35
60.32
12.20
19.38
14.47
17.62
22.38

19.80
29.72
16.70
16.66
28.87
58.86
11.54
17.69
18.82
18.04
22.87

16.93
25.82
13.55
15.19
27.15
51.20
10.39
13.19
14.28
14.35
15.52

397
561
402
259
619
1280
234
510
406
448
602

397
573
360
310
598
1255
222
481
481
456
611

351
506
310
285
569
1125
205
406
402
393
490

A=available carbohydrate calculated by summation of analytical values; includes glucose, fructose, sucrose and
starch (total starch-resistant starch).
B=available carbohydrate (by difference) calculated as 100 g minus the sum of grams of water, protein, lipids, ash
and dietary ber.
C=available carbohydrate (by difference) calculated as 100 g minus the sum of grams of water, protein, lipids, ash
and indigestible fraction.
Ae, Be, Ce=energy value calculated as summation of (protein grams  16.7 kJ; lipids grams  37.4 kJ; available
carbohydrate  16.7 kJ; using A, B, C, respectively).

4. Conclusions
Once the IF includes several indigestible components, it presents higher content than the DF.
Consequently, when subtracting the IF in the difference on the estimate of the available
carbohydrates, lower values are obtained than when subtracting DF. Similar results were found in
relation to the energy values. The IF analysis used in this work is relatively simple and has good
reproducibility, and hence it can represent an alternative for the proximate composition
determination, although it does not provide information about each indigestible component.
For the 11 test starchy foods, the total sugar content was low and there was no considerable
difference on the energy value when the available carbohydrates were estimated by the
summation of the different available carbohydrates or by difference subtracting DF.
The analysis of several individual components that constitute the carbohydrate fraction must be
encouraged, aiming to insure the quality of carbohydrate and energy information in food
composition databases.

Acknowledgements
The authors acknowledge with gratitude the nancial support granted by the Conselho
!
Nacional de Desenvolvimento Cient!co e Tecnologico
(CNPq)/Programa Institucional de

ARTICLE IN PRESS
338

E.W. Menezes et al. / Journal of Food Composition and Analysis 17 (2004) 331338

Bolsas de Inicia@*ao Cient!ca (PIBIC), Funda@*ao Coordena@*ao de Aperfei@oamento de

Pessoal de N!vel Superior (CAPES), and Food Science Post Graduation Course, FCF-USP.
References
Burlingame, B., 1996. Proximate methods and modes of expression: variability as a harmonization issue. In: National
Nutrient Databank Conference, 21. Proceedings, Retrieved December 10, 2001 from the World Wide Web: http://
www.nal. usda.gov/fnic/foodomp.
Food and Agriculture Organization/World Health Organization (FAO/WHO), 1998. Carbohydrates in human
nutrition: report of a joint FAO/WHO expert consultation, April 1418, 1997. Food and Nutrition Paper, FAO,
Rome, 140pp.
Goni, I., Garc!a-Diz, L., Manas, E., Saura-Calixto, F., 1996. Analysis of resistant starch: a method for foods and food
products. Food Chemistry 56, 445449.
Goni, I., Garcia-Alonso, A., Saura-Calixto, F., 1997. A starch hydrolysis procedure to estimate glycemic index.
Nutrition Research 17, 427437.
Jenkins, D.J., Kendall, C.W., Augustin, L.S., Franceschi, S., Hamidi, M., Marchie, A., Jenkins, A.L., Axelsen, M.,
2002. Glycemic index: overview of implications in health and disease. American Journal of Clinical Nutrition 76,
266S273S.
Li, B.W., Andrews, K.W., Pehrsson, P.R., 2002. Individual sugars, soluble, and insoluble dietary ber contents of 70
high consumption foods. Journal of Food Composition and Analysis 15, 715723.
Manas, E., Bravo, L., Saura-Calixto, F., 1994. Sources of error in dietary bre analysis. Food Chemistry 50, 331342.
! resistente en alimentos y productos
Menezes, E.W., Lajolo, F.M., 2000. Contenido de bra diet!etica y almidon
! y caracterizacion
! de bra dietetica para su aplicacion
! em
iberoamericanos. Proyecto CYTED XI.6 Obtencion
alimentos para reg!menes especiales, CYTED/CNPq. Docuprint, S*ao Paulo, 121pp, http://www.fcf.usp.br/cytedxi6.
Miller, J.B., Pang, F., Broomhead, L., 1995. The glycemic index of foods containing sugarscomparison of foods with
naturally occurring v added sugars. British Journal of Nutrition 73, 613623.
Monro, J., Burlingame, B., 1996. Carbohydrates and related food compounds: INFOODS tagnames, meanings, and
uses. Journal of Food Composition and Analysis 9, 100118.
Prosky, L., Asp, N-G., Schweizer, T.F., Devries, J.W., Furda, I., 1988. Determination of insoluble, soluble and total
dietary bre in food and foods products. Interlaboratory study. Journal of Association of Ofcial Analytical
Chemists 71, 10171023.
Rosin, P.M., Lajolo, F.M., Menezes, E.W., 2002. Measurement and characterization of dietary starches. Journal of
Food Composition and Analysis 15, 367377.
Saura-Calixto, F., Garc!a-Alonso, A., Goni, I., Bravo, L., 2000. In vitro determination of the indigestible fraction in
foods: an alternative to dietary ber analysis. Journal of Agricultural and Food Chemistry 48 (8), 33423347.
Universidade de S*ao Paulo (USP), 1998. Tabela Brasileira de Composi@*ao de Alimentos-USP. Departamento de
Alimentos e Nutri@*ao Experimental, Faculdade de Ci#encias Farmac#euticas. Retrieved April 2, 2002 from the World
Wide Web: http://www.fcf.usp.br/tabela.
Wahlqvist, M.L., 2002. The place of carbohydrates in newer food formulations: Opportunities for nutritional advances
and their safety. Asia Pacic Journal of Clinical Nutrition 11, S149S154.