Starch Analysis in Food

Klaus N. Englyst, Geoffrey J. Hudson, and Hans N. Englyst

in
Encyclopedia of Analytical Chemistry
R.A. Meyers (Ed.)
pp. 4246–4262
 John Wiley & Sons Ltd, Chichester, 2000
STARCH ANALYSIS IN FOOD 1

Starch Analysis in Food physiological properties of starch can vary considerably,

and both the botanical source and the effects of food
processing are major determinants of starch digestibility.
Klaus N. Englyst, Geoffrey J. Hudson, and In addition to the nature of the starch itself, the site, rate and
Hans N. Englyst extent of digestion of starch in the human small intestine are
Englyst Carbohydrate Services, Eastleigh, UK influenced by a number of host factors. The rate at which
starch is digested in the human small intestine results in a
wide range of glycemic responses, and this physiological
measurement has been used to rank foods by their glycemic
1 Introduction 1 index. In vitro studies have indicated that glycemic response
2 History 4 and the rate of starch digestion are closely correlated.
Rapidly digestible starch (RDS) and slowly digestible
3 Applications of Starch Measurements 4
starch (SDS) fractions together represent the starch that
3.1 Food Tables and Food Labeling 4
is likely to be digested completely in the human small
3.2 Public Health 5
intestine, with any remaining starch defined as the resistant
4 Methods 5 starch (RS) fraction that is available for fermentation in
4.1 Reagents 5 the large bowel. Measurements of RDS, SDS and RS
4.2 Apparatus 7 can be obtained by one simple procedure. Values for the
4.3 Sample Preparation 8 different starch fractions obtained by the in vitro method
4.4 The Colorimetry Version (Figure 1) 8 described here represent reproducible measurements that
4.5 The High-performance Liquid can be used to classify dietary starch according to its
Chromatography Version (Figure 2) 10 potential digestibility. In addition to these starch fractions,
4.6 Calculations 11 two terms, rapidly available glucose (RAG) and slowly
4.7 Measurement of Resistant Starch available glucose (SAG), are introduced to reflect the rate
1, 2 and 3 (Using Colorime- at which glucose (from both sugars and starch) is likely to
try or High-performance Liquid be absorbed in the small intestine.
Chromatography) 11 The proportions of RAG, SAG, RDS, SDS and RS
5 Quality Control and Troubleshooting 12 in foods can be controlled by food processing. The
5.1 Quality Control 12 implications of altering the rate and extent of starch
5.2 Troubleshooting 13 digestion are potentially of great importance to public
6 Method Development and Validation 13 health. A full understanding of the links between dietary
6.1 Methodological Considerations 13 carbohydrates and health and the underlying mechanisms
6.2 Validation of Rapidly Digestible will come only from the specific measurement of individual
Starch and Slowly Digestible Starch types of dietary carbohydrates.
Values 14
6.3 Validation of Resistant Starch
Values 14 1 INTRODUCTION
7 Comparison with Other Methods 15 Advances in knowledge of the relations between diet and
7.1 Total Starch 15 disease have implicated the importance of dietary car-
7.2 Rate and Extent of Starch Digestion 15 bohydrates. A full understanding of the links between
Abbreviations and Acronyms 15 dietary carbohydrates and health and the underlying
Related Articles 15 mechanisms will come only from the specific measure-
ment of individual types of dietary carbohydrates. Table 1
References 15
shows the overall nutritional classification of plant car-
bohydrates in the human diet. This classification is based
primarily on carbohydrate chemistry, with groupings
Starch is quantitatively an important component of the and sub-divisions related to nutritional and physiological
human diet, being present in grains, tubers and legumes. properties.
Starch has for a long time been considered by many Starch is composed of D-glucose units and exists as
as being slowly but completely digested in the small a mixture of two polymers, amylose and amylopectin.
intestine, resulting in modest glycemic responses and with Amylose is an essentially linear compound with an aver-
no physiological role other than as an energy source. It age degree of polymerization (DP) of 500 with mainly
is now understood that in fact the metabolic fate and a-D-(1 – 4)-glucosidic bonds. It has a helical structure

Encyclopedia of Analytical Chemistry

R.A. Meyers (Ed.) Copyright  John Wiley & Sons Ltd
2 FOOD

Table 1 Classification of the carbohydrates in plant foods

Class/components Comments
Sugars:
Mono- and disaccharides and their alcohols Physiological response depends on identity and rate of release.
Free glucose C glucose from sucrose D FSG
Short-chain carbohydrates:
Maltodextrins Measured as RDS
Resistant short-chain carbohydrates Fermented in the large bowel and may stimulate growth of
(non-digestible oligosaccharides) bifidobacteria
Starch:
RDS RDS C rapidly released FSG D RAG
SDS SDS C slowly released FSG D SAG
RS Escapes digestion in the small intestine
NSP:
Plant cell-wall NSP Encapsulate and slow absorption of other nutrients.
Marker for naturally high-fiber diets for which health benefits have
been shown.
Fermented in the large bowel to different extents
Other NSP Food additives.
Minor components of the human diet.
Fermented in the large bowel to different extents

NSP, nonstarch polysaccharides; FSG, free-sugar glucose.

with a hydrophobic cavity allowing complexation with Heating starch granules in a moist environment
hydrophobic molecules such as lipids and iodine (giving leads to loss of the crystalline structure, swelling and
the characteristic blue – black color). Amylopectin has a leaching of amylose molecules from the granules. This
DP in the thousands with a-D-(1 – 4)-linked chains exten- process, known as gelatinization, usually occurs in the
sively branched through a-D-(1 – 6)-glucosidic linkages temperature range 50 – 70 ° C, or at higher temperatures
(about 5% of the total linkages). The ratio of amylose to in low-moisture conditions. Cooling of gelatinized starch
amylopectin varies considerably between species and has results in gelation and retrogradation, which is the
been manipulated by selective breeding to produce both recrystallization of the starch molecules, particularly
high-amylose and high-amylopectin (waxy) varieties of amylose, into a form with the type B X-ray pattern.
several cereals. Starch enclosed within particulate matter or plant
Starch is the storage polysaccharide of many higher cell walls and starch in the form of intact granules is
plants, especially cereals, legumes and most tubers. In digested slowly, because the physical access of amylolytic
the plant, starch is stored within the cells as partially enzymes is restricted. Disruption of the particulate matter
crystalline granules. An extensive review of the physical and removal of plant cell walls during food processing,
characteristics of starch granules has been given by especially milling, increase the surface area and can
Gallant et al..1/ The size and shape of the granules and damage the structure of starch granules, increasing the
the crystal structure within them are characteristic of the susceptibility to enzymatic degradation. Gelatinization
botanical source. Starch granules have been characterized tends to increase and retrogradation to reduce starch
as type A, B or C according to their X-ray diffraction digestibility. Colonna et al. have provided a detailed
pattern..2,3/ In general, cereal starches have the type account of the influence of food processing on starch
A pattern, starches from tubers such as potatoes tend granules and their digestion..5/
to have the type B pattern, and C-type starch, found The rate at which starch is digested in the human small
in some legumes, is a combination of the A and B intestine results in a wide range of glycemic responses,
patterns. In general, starch granules with B or C type and this physiological measurement has been used to
patterns are more resistant to digestion by pancreatic rank foods by their glycemic index..6/ In vitro studies
amylase, although the degree of resistance varies with have indicated that glycemic response and the rate of
botanical origin..4/ This resistance to hydrolysis affects starch digestion are closely correlated..7 – 15/
the digestibility of starchy foods that are normally Starch that escapes digestion in the small intes-
eaten raw, such as bananas, and of processed foods tine is termed RS, which, by definition, reaches
such as biscuits in which the starch is incompletely the large intestine and is a potential substrate for
gelatinized. microbial fermentation..16/ The main end-products of this
STARCH ANALYSIS IN FOOD 3

fermentation are the gases hydrogen and carbon dioxide, Table 2 Starch fractions and RAG content of some starchy
and methane in about 50% of people, and short-chain foods (grams per 100 g as eaten)
fatty acids, which may be absorbed and utilized..17/
Food TS RDS SDS RS RAG
It is clear from this brief discussion that the metabolic
fate and physiological properties of starch can vary Cereals:
considerably, and that the botanical source and the Pearled barley 17.1 8.0 7.0 2.1 9
effects of food processing are major determinants of Sweetcorn 17.1 15.4 1.4 0.3 18
starch digestibility. In addition to the nature of the starch Bread:
Rye crispbread 59.8 48.8 6.7 4.3 55
itself, the site, rate and extent of digestion of starch in Rye wholemeal 33.8 23.2 7.4 3.2 27
the human small intestine are influenced by a number of Wheat white 41.7 37.4 3.7 0.6 42
other factors. These include the extent of chewing, the Wheat wholemeal 35.0 32.1 1.4 1.5 36
rate of gastric emptying, the transit time of food along the Biscuits:
small intestine, the concentration of pancreatic amylase Digestive 46.5 32.0 12.6 1.9 44
available for breakdown of the starch, the amount Oatmeal 55.9 48.8 6.2 0.9 55
of starch consumed and the influence of other food Breakfast cereals:
components that may inhibit enzymatic hydrolysis..18/ All Bran 22.2 20.6 0.5 1.1 35
Oat bran 45.8 31.2 13.6 1.0 36
The rate and extent of digestion of starch in vivo is Porridge Oats 13.0 9.9 3.1 0.1 11
variable both within and between individuals, which is Puffed Wheat 68.7 62.5 0.0 6.2 70
recognized in the definition of RS as ‘‘the starch (and Rice Krispies 69.8 65.6 1.7 2.5 80
starch degradation products) that, on average, reaches the Shredded Wheat 62.2 48.7 11.9 1.6 55
human large intestine’’..16/ The values for the different Weetabix 57.0 56.8 1.0 0.0 65
starch fractions obtained by the method described here Rice:
represent reproducible in vitro measurements that can be Brown, long grain 23.8 14.6 9.2 0.0 16
Parboiled 27.8 16.6 10.0 1.2 19
used to classify dietary starch according to its potential White, long grain 23.0 17.4 5.6 0.0 19
digestibility. Pasta:
Measurements of RDS, SDS and RS can be obtained Macaroni 26.2 13.4 12.0 0.8 15
by one simple procedure..16,18,19/ RDS and SDS fractions White spaghetti 23.5 13.5 9.0 1.0 15
together represent the starch that is likely to be digested Legumes:
completely in the human small intestine, with any remain- Butter beans 11.4 9.4 0.8 1.2 11
ing starch defined as the RS fraction. Three types of Chickpea 16.4 5.1 8.8 2.5 6
RS have been categorized: physically inaccessible starch Frozen peas 7.2 4.1 1.0 2.1 6
Haricot beans 18.2 4.1 5.8 8.3 5
(RS1), resistant starch granules (RS2) and retrograded Kidney beans 17.0 4.7 9.8 2.5 6
starch (RS3). Pinto beans 16.1 9.3 5.0 1.8 11
In addition to these starch fractions, two terms, RAG Red lentils 15.8 7.3 6.1 2.4 8
and SAG, reflect the rate at which glucose is likely to Tubers:
be absorbed in the small intestine. RAG and SAG may Instant potato 12.7 10.9 1.1 0.8 12
include glucose derived from RDS and SDS, respectively, Potato 16.0 15.2 0.7 0.1 17
Potato crisps 50.0 42.7 2.8 4.5 48
but the glucose may be derived from sugars. FSG is Sweet potato 9.3 7.5 0.8 1.1 11
usually included in the RAG fraction, but may form part Yam 16.8 14.3 0.4 2.1 18
of the SAG fraction if trapped within particulate matter
or plant cell walls. This ambiguity as to what proportions
of RAG and SAG are derived from starch or from FSG starch is probably a combination of starch granules being
complicates the determination of RDS and SDS fractions encapsulated by cell walls (dietary fiber) and not being
in foods containing appreciable amounts of FSG. The use fully gelatinized. Spaghetti, macaroni and pearled barley
of RDS and SDS should therefore be restricted to starchy are examples of foods that have a moderate RAG value
foods that contain only small amounts of FSG, otherwise as the result of a dense structure. For digestive biscuits,
RAG and SAG are more appropriate. a considerable proportion of the starch is measured as
Table 2 shows the proportions of the various fractions SDS, reflecting incomplete gelatinization, but the high
of starch and the RAG values for a range of foods. These dry matter content results in similar RAG values for
fractions are all expressed in units of grams per 100 g digestive biscuits and white bread. Cooked potato has a
of food ‘‘as eaten’’. The legumes have the lowest RAG low RDS value, reflecting the low dry matter content, but
values, which are associated with low levels of free sugar RDS represents a high proportion of the total starch (TS)
and a low proportion of starch measuring as RDS. The content. In general, highly processed breakfast cereals
cause of the slow and incomplete digestion of legume have high RAG values.
4 FOOD

The proportions of RAG, SAG, RDS, SDS and RS Increased interest in dietary fiber (plant cell walls) as
in foods, and thus the expected rate and extent of a potentially protective agent against Western diseases
digestion in the human small intestine, can be controlled of affluence led to the development of the NSP-type
by food processing. The implications of altering the rate measurement techniques for the dietary carbohydrates
and extent of starch digestion are potentially of great that McCance and Lawrence referred to as unavailable.
importance to public health. The identification of specific This type of procedure was pioneered by Southgate and
health benefits related to the ingestion of various types of later developed by Englyst, who identified a fraction of
starch will be possible only when separate measurements RS3 that could not be hydrolyzed without prior dispersion
of these are available. It is therefore important to be able with dimethyl sulfoxide (DMSO) or KOH..21 – 26/ This
to measure nutritionally relevant starch fractions in vitro, starch fraction, subsequently identified as RS3, was
and such methods are the subject of this article. termed RS. Later, in conjunction with ileostomy studies,
two other forms of RS were identified; RS1 and
RS2. The definition of RS includes all three fractions,
2 HISTORY and no starch is included in the measurement of
NSP.
The classification and measurement of dietary carbo- There is growing interest in the rate and extent
hydrates have developed during the 20th century as of starch digestion with respect to the magnitude
knowledge of the nutritional importance of this fraction of of the glycemic response. Jenkins et al..9/ have pio-
the diet has accumulated. The perception of dietary car- neered the glycemic index, which may be used to rank
bohydrates as simply an energy source has been echoed foods by their likely effect on the glycemic response.
in the traditional calculation of ‘‘carbohydrate by differ- In a classical study, Haber et al..27/ demonstrated the
ence’’, as the portion remaining once moisture, protein, effect of particle size, i.e. degree of food process-
fat and ash are accounted for. ing, on the rate of absorption of sugars from apples,
In 1929, McCance and Lawrence highlighted the and several in vitro studies have shown that the rate
general lack of available data for the carbohydrate content of starch digestion varies according to its botanical
of foods and the enormous variation that existed between source and degree of processing..5,10,12,27/ Measure-
some of the published values..20/ McCance and Lawrence ment of RAG, SAG and starch fractions as described
were well aware that the carbohydrate composition of raw here represent tools for studying the mechanisms that
and cooked foods can be very different and they identified underlie the links between dietary carbohydrates and
the limitation of much of the available data, which were health.
nearly all for the composition of raw foods. McCance
and Lawrence were primarily concerned with the dietary
management of diabetes, and they made the distinction 3 APPLICATIONS OF STARCH
between ‘‘available’’ and ‘‘unavailable carbohydrates’’. MEASUREMENTS
They identified available carbohydrates as starch and the
soluble sugars, sucrose, glucose and fructose. 3.1 Food Tables and Food Labeling
McCance and Lawrence compiled a table of the
available carbohydrate content of a range of plant foods, Food tables should supply meaningful, reliable data
which were analyzed in the form in which they were that are suitable for use by dietitians, epidemiologists
normally eaten, i.e. cooked or raw as appropriate..20/ and the food industry. The Food and Agriculture
They used hydrolysis with hot 7.5% (v/v) hydrochloric Organization (FAO) of the United Nations and the
acid under a reflux condenser, followed by neutralization World Health Organization (WHO) have recommended
with sodium hydroxide, and they measured both reducing that compositional data for dietary carbohydrates for use
sugars and pentoses. in food tables and food labeling should be based on
McCance and Lawrence introduced two important chemically identified components, which may be further
principles into the measurement of dietary carbohydrates subdivided by nutritional properties..28/ The classification
for nutritional purposes: (1) the need to analyze foods in of dietary carbohydrates presented in Table 1 conforms
the form in which they are eaten; and (2) the need to make with these criteria. The inclusion in food tables of the
specific chemical measurements of dietary carbohydrates. RDS, SDS and RS subdivisions of starch provides the
The analysis and measurement of dietary carbohydrates tools for investigating the health implications of altering
in later years has been greatly facilitated by the advent the rate and extent of starch digestion. The RAG and
of highly purified enzyme preparations and advances in SAG terms provide information on the rate at which
techniques for the separation and specific measurements glucose from starch and sugars is likely to become
of individual sugars. available for absorption in the human small intestine,
STARCH ANALYSIS IN FOOD 5

and may be used as a guide to the likely magnitude of includes the use of an internal standard. Figures 1 and 2
the glycemic response..19/ Such values do not become summarize the colorimetry and HPLC procedures.
obsolete and can be used in different combinations for The subsidiary procedure for the measurement of FSG
different purposes. The use of these values for food is described, which is required for calculation of RDS and
labeling will depend on confirmation of their importance TS, and this procedure may be used to obtain a rapid
for public health. measurement of TS.
Regarding analytical throughput, a shaking water-bath
3.2 Public Health capable of holding 18 tubes allows the analysis of five test
samples in triplicate. Both the main and the subsidiary
A review of current knowledge suggests that a diet procedures may be completed within a working day.
rich in slowly digested carbohydrates has an overall
protective effect against the non-communicable diseases
4.1 Reagents
of westernization..28,29/ The results of epidemiological
studies have suggested that diets with a high content
4.1.1 Common to the Colorimetry and High-performance
of rapidly digested carbohydrates from refined foods
Liquid Chromatography Procedures
increase the relative risk of developing diabetes mellitus
and coronary heart disease..29 – 32/ There are indications Distilled water or water of equivalent purity should be
from studies with low glycemic index diets that the rate of used throughout the method. All sugars should be dried to
starch digestion is of importance in the glycemic control constant weight under reduced pressure with phosphorus
of people with diabetes..33/ pentoxide before use.
The impact of RS on health remains uncertain. It
has been suggested that the high proportion of butyrate Amyloglucosidase EC 3.2.1.3 (Englyst Carbohydrate
produced during the fermentation of RS in the colon Services Ltd, Cat. No. 61-002).
may benefit the gut epithelium and act as a protective
Amyloglucosidase Solution Amyloglucosidase dilu-
agent against colon cancer. However, caution should be
ted 1 : 7 (v/v) with water.
applied until the health implications of RS fractions have
been assessed, and safe levels of RS intakes have been Benzoic Acid, Saturated Prepare a saturated solution
determined. of benzoic acid at room temperature.

Enzyme Mixture For 18 samples/standards, into each

4 METHODS of six centrifuge tubes weigh 3.0 g of pancreatin and
suspend in 20 mL of water using a vortex-mixer. Add
The main analytical procedure described here quanti- a magnetic stirring bar and mix for 10 min. Centrifuge
fies the various fractions of starch through enzymatic at 1500 g for 10 min, then remove 15 mL of the cloudy
hydrolysis and measurement of the glucose released. supernatant from each tube and combine (90 mL total).
Samples are analyzed ‘‘as eaten’’ and are treated with Add 4 mL of amyloglucosidase and 6 mL of invertase.
protease to disrupt any starch– protein interaction. This Mix well. The enzyme mixture should be prepared
is followed by incubation with amylolytic enzymes under immediately before use.
conditions controlled for temperature, pH, viscosity and
rate of mechanical mixing. Subsamples are taken at 20 min Guar Gum Powder (Sigma, Cat No. G-4129).
[the glucose measured in the portion removed from the
analysis at 20 min after the addition of enzymes (G20)] Heat-stable Amylase EC 3.2.1.1 (Termamyl, Englyst
and 120 min [the glucose measured in the portion removed Carbohydrate Services Ltd, Cat. No. 61-005).
from the analysis at 120 min after the addition of enzymes
(G120)] as measures of the rate and extent of starch Hydrochloric Acid 0.05 mol L 1 .
digestion. Any remaining starch is dispersed before enzy-
Invertase EC 3.2.1.26 (Merck, Cat. No. 39020).
matic hydrolysis to give a value for total glucose (TG). A
value for RS is calculated as the difference between G120 Pancreatin (Sigma, 8 ð USP, Cat. No. P-7545).
and TG. Separate values may be obtained for RS1, RS2
and RS3. Pepsin Powder EC 3.4.23.1 (Sigma, Cat. No. P-7000).
Two versions of the main procedure are described:
one is based on the colorimetric determination of glu- Pepsin – Guar Gum Solution For 18 samples/stand-
cose and the other on the measurement of glucose by ards, add 1 g of pepsin powder to 200 mL of 0.05 mol L 1
high-performance liquid chromatography (HPLC), which hydrochloric acid and mix with a magnetic stirring
6 FOOD

Sample (0.6 − 4 g)

Add 10 mL pepsin−guar gum

30 min at 37°C

Add 10 mL 0.25 mol L−1 sodium acetate

Add 5 glass balls

Add 5 mL enzyme mixture

(pancreatin, amyloglucosidase, invertase)

Incubate with shaking at 37°C

After 20 min remove 0.5 mL

After 120 min remove 0.5 mL

Add 20 mL buffer
Vortex-mix remainder Add 5 glass balls
Place in
20 mL ethanol Vortex-mix vigorously
30 min at 100°C
−1
Cool to 0°C. Add 10 mL 7-mol L KOH 30 min at 100°C
Vortex-mix

Centrifuge Add 0.2 mL invertase

30 min at 0°C with shaking

Vortex-mix. Transfer 1 mL into

Incubate with shaking
10 mL 0.5-mol L−1 acetic acid
30 min at 37°C
Add 0.2 mL amyloglucosidase

30 min at 70°C Transfer 0.5 mL into

10 min at 100°C 10 mL ethanol
Measure glucose released Centrifuge
Add 40 mL water after 20 min (G20)
Centrifuge
Measure glucose released Measure free sugar
Measure total glucose (TG) after 120 min (G120) glucose (FSG)

Figure 1 Flow diagram of the steps in the colorimetry version of the procedure.

bar. Just before use, add 1 g of guar gum and mix Glucose Oxidase Diagnostic Kit [Boehringer glucose
well. The pepsin – guar gum solution should be prepared test combination, Cat. No. 166391].
immediately before use.
Glucose Standard Weigh 50 g of glucose to the
Potassium Hydroxide 7 mol L 1 . nearest 0.1 mg. Make up to 200 mL with sodium acetate
buffer to give a 25 mg mL 1 solution.
Sodium Acetate Buffer 0.1 mol L 1 , pH 5.2 Dissolve
13.6 g of sodium acetate trihydrate in water, add 250 mL
Sodium Acetate 0.25 mol L 1 . Weigh 34 g of sodium
of saturated benzoic acid solution and make up to 1 L
acetate trihydrate and make up to 1 L with water.
with water. Adjust to pH 5.2 with 0.1 mol L 1 acetic acid.
To stabilize and activate enzymes, add 4 mL of 1 mol L 1
calcium chloride to 1 L of buffer. 4.1.3 For the High-performance Liquid Chromatography
Procedure Only
4.1.2 For Colorimetry Only

Acetic Acid 0.5 mol L 1 . Plus 4 mL L 1

of 1 mol L 1 Acetic Acid 1 mol L 1 . Plus 4 mL L 1
of 1 mol L 1

calcium chloride. calcium chloride.

Ethanol 66% (v/v). Ethanol Absolute.

STARCH ANALYSIS IN FOOD 7
Sample (0.6 − 4 g)

Add 5 mL internal standard (arabinose)

Add 10 mL pepsin−guar gum

30 min at 37°C

Add 5 mL 0.5-mol L−1 sodium acetate

Add 5 glass balls

Add 5 mL enzyme mixture

(pancreatin, amyloglucosidase, invertase)

Incubate with shaking at 37 °C

After 20 min remove 0.2 mL

After 120 min remove 0.2 mL

Add 5 mL internal standard
Vortex-mix remainder Add 20 mL water
Place in
4 mL ethanol Add 5 glass balls
30 min at 100°C Vortex-mix vigorously

Cool to 0°C. Add 10 mL 7-mol L−1 KOH

Vortex-mix 30 min at 100 °C

30 min at 0°C with shaking Centrifuge Add 0.2 mL invertase

Vortex-mix.
Take 70 µL into vial
Transfer 0.2 mL into Incubate with shaking
−1
1 mL 1-mol L acetic acid 30 min at 37 °C
Add 40 µL amyloglucosidase Add 1mL water
Vortex-mix Transfer 0.5 mL into
30 min at 70 °C 10 mL ethanol
Centrifuge
10 min at 100 °C
Transfer 70 µL into vial
Add 12 mL ethanol
Add 1 mL water
Centrifuge
Measure glucose released Vortex-mix
Transfer 200 µL into vial
Add 1 mL water after 20 min (G20)
Vortex-mix
Measure glucose released Measure free sugar
after 120 min (G20) glucose (FSG)
Measure total glucose (TG)

Figure 2 Flow diagram of the steps in the HPLC version of the procedure.

Internal Standard Dissolve 40 g of arabinose in water, 4.2 Apparatus

add 500 mL of saturated benzoic acid and make up to 1 L
with water. 4.2.1 Common to the Colorimetry and High-performance
Liquid Chromatography Procedures
Stock Sugar Mixture Dissolve 50 g of glucose and 25 g
Balance Accurate to 0.1 mg.
of fructose in water, add 500 mL of saturated benzoic acid
and make up to 1 L with water. Centrifuge Capable of exerting 1500 g.

Sodium Acetate 0.5 mol L 1 . Weigh 68 g of sodium Centrifuge Tubes Polypropylene (Copolymer) cen-
acetate trihydrate and make up to 1 L with water. trifuge tubes of length 11.5 cm and diameter 3 cm (50-mL
capacity), with screw caps (Alpha Laboratories, Cat. No.
Sodium Hydroxide 50% (w/v). LW 1110).
8 FOOD

Glass Balls (Marbles) Approximately 1.5 cm in may be milled or homogenized. Determine the mois-
diameter, sufficient to provide five balls per sample. ture content of samples as weight loss after overnight
incubation at 104 ° C.
Magnetic Stirrer
Sample weights should be chosen to contain
Mincer Hand-driven, with a plate with 0.9-cm 500 – 600 mg of starch and sugars, which can be estimated
diameter holes. from food tables. Examples of suitable sample weights
are given in Table 3.
Test Tubes Glass or plastic, of 15- and 30-mL capacity,
The procedure should now continue, using either the
preferably with lids, capable of withstanding low-speed
colorimetry or the HPLC version.
centrifugation.
Vortex-mixer 4.4 The Colorimetry Version (Figure 1)
Water-bath Capable of maintaining 100 ° C.
4.4.1 Measurement of Rapidly Available Glucose, Slowly
Water-baths Capable of maintaining temperatures Digestible Starch, Resistant Starch and Total Starch
in the range 35 – 70 ° C. This or another similar bath
must have a linear shaking capacity of not less than Preparation of the Standard and Blank Into each of
160 strokes min 1 and a stroke length of approximately three centrifuge tubes weigh 50 mg of guar gum powder
35 mm. The shaking bath must be fitted with the means and add five glass balls. Pipet 20 mL of glucose standard
to hold all the centrifuge tubes exactly horizontally under into each of two of these and 20 mL of acetate buffer
the water, with the long axis of each tube exactly parallel into the other (blank). Cap and shake well to disperse the
with the direction of movement. Each bath should be of gum. Treat these tubes exactly as the samples after the
sufficient capacity that there is no significant change in addition of sodium acetate.
temperature when a rack containing all the tubes is placed Sample Weight Weigh an appropriate amount of
in it. sample, to the nearest 0.1 mg, into 50-mL tubes. If the
sample is hot, cap the tube and place in a water-bath
4.2.2 For the Colorimetry Procedure Only
at 70 – 80 ° C until ready to analyze. Samples should be
Spectrophotometer analyzed at least in duplicate. Reference samples should
be included in every batch (see section 5).
4.2.3 For the High-performance Liquid Chromatography
Incubation with Pepsin Put 10 mL of the freshly
Procedure Only
prepared pepsin – guar gum solution into each sample
Autoinjector tube, vortex-mix and place the tubes in a water-bath at
37 ° C. After 30 min of incubation, remove the samples
Ion-exchange Column with Guard Column For from the water-bath and add five glass balls and 10 mL
separation of sugars. of 0.25 mol L 1 sodium acetate to each tube. Shake the
Ion-exchange Column For trapping amino acids and tubes gently to disperse the contents and replace the
peptides. sample tubes, standards and blanks in the water-bath at
37 ° C to equilibrate.
Column-switching Mechanism
Enzymatic Hydrolysis of Starch Remove one sample
Gradient Pump
tube from the 37 ° C water-bath and add 5 mL of enzyme
Electrochemical Detector mixture. Immediately cap the tube and mix the contents
gently by inversion before securing the tube horizontally
Computerized Data-handling System
in the 37 ° C shaking water-bath. Start the shaking action
of the water-bath; this is time zero for the incubation
4.3 Sample Preparation
Starch digestibility is greatly influenced by food pro- Table 3 Examples of suitable sample weights
cessing and samples must be analyzed as eaten. Foods
Dry matter (%) Examples Weight (g)
normally eaten dry are analyzed dry, and foods normally
eaten hot and wet are cooked and maintained at 70 – 80 ° C 75 – 100 Starch 0.6
immediately before analysis. Foods with a recognizable 75 – 100 Flours, breakfast cereals 0.8
structure that would normally require chewing (e.g. pasta, 55 – 75 Bread, cakes 1.0 – 2
35 – 55 Beans, pasta, rice 1.5 – 3
rice, maize) are passed through the mincer. For determi- 15 – 35 Canned foods, sauces 3.0 – 4
nation of FSG and the rapid measurement of TS, samples
STARCH ANALYSIS IN FOOD 9

and the shaking action is not interrupted until all the tubes and add five glass balls to each. Pipet 25 mL of 0.1-
G120 portions have been collected (see below). Repeat mol L 1 sodium acetate buffer into each tube, cap and
the addition of enzyme mixture for the rest of the sample vortex-mix vigorously to begin breaking up large parti-
tubes, at 1-min intervals to aid timing of the procedure, cles. Place the rack of tubes into a boiling water-bath.
and place them into the shaking water-bath. Remove After 30 min remove the rack of tubes from the boil-
each tube from the bath at exactly 20 min after addition ing water-bath and vortex-mix vigorously to break up
of the enzyme mixture, transfer 0.5 mL into 20 mL of any remaining particles of sample. Cool to 37 ° C, then
66% ethanol and vortex-mix to stop the hydrolysis; add 0.2 mL of invertase to each tube. Cap the tubes,
this is the G20 portion. Return the sample tube to the immerse horizontally in the shaking water-bath at 37 ° C
shaking water-bath immediately after the sample has and incubate for 30 min.
been taken. After a further 100 min (a total of 120 min Vortex-mix vigorously before transferring 1 mL of each
of incubation), transfer 0.5 mL from the sample tube into sample or standard solution into a test tube containing
20 mL of 66% ethanol and vortex-mix; this is the G120 2 mL of absolute ethanol and vortex-mix. Centrifuge at
portion. 500 g for 5 min, then transfer 1 mL of the supernatant
into 5 mL of water and mix well by inversion. (For
Dispersion of Resistant Starch Having removed
the standard, transfer 1 mL of supernatant into 20 mL
all the tubes from the shaking water-bath, vortex-
of water.) This is the FSG portion to be measured in
mix vigorously to break up any large particles, then
step 4.4.3.
place the rack of tubes into a boiling water-bath for
If TS is to be determined by the main method, the
30 min. Remove the tubes from the bath, vortex-mix
procedure is finished; otherwise, the procedure can be
and place the tubes in ice – water until thoroughly
continued for a rapid measurement of TS.
chilled. Add 10 mL of 7 mol L 1 KOH, cap the tube
and mix the contents by inversion. Immerse the tubes
horizontally in a shaking water-bath containing ice – water Measurement of Total Glucose To the remainder of
for 30 min. the sample solution add 0.1 mL of heat-stable amylase,
cap the tube, vortex-mix vigorously, then place in a boiling
Hydrolysis of Starch to Glucose Remove the samples water-bath for 15 min. Remove the tubes from the bath,
from the waterbath and immediately transfer 1 mL of vortex-mix and then place the rack of tubes in ice – water
the contents into a 50-mL tube containing 10 mL of and leave them to chill thoroughly.
0.5-mol L 1 acetic acid and mix well. Add 0.2 mL of Add 10 mL of 7-mol L 1 KOH, cap the tube and mix
amyloglucosidase solution to each sample. Mix and place the contents by inversion. Immerse the tubes horizontally
in a water-bath at 70 ° C. After 30 min, remove the rack
in a shaking water-bath containing ice – water for 30 min.
of tubes to a boiling water-bath for 10 min. Cool the
Remove the tubes from the water-bath and immedi-
tubes to room temperature, add 40 mL of water, cap
ately transfer 1 mL into a 50-mL tube containing 10 mL
and mix; this is the TG portion to be measured in
of 0.5-mol L 1 acetic acid and mix well. Add 0.2 mL of
step 4.4.3.
amyloglucosidase solution to each sample. Mix and place
in a water-bath at 70 ° C. After 30 min remove the rack of
4.4.2 Measurement of Free-sugar Glucose; Optional tubes to a boiling water-bath for 10 min. Cool the tubes
Rapid Measurement of Total Starch to room temperature, add 40 mL of water, cap and mix;
this is the TG portion.
Preparation of the Standard and Blank Into each of
two 50-mL tubes, pipet 25 mL of glucose standard and
into a third add 25 mL of sodium acetate buffer (blank). 4.4.3 Determination of Glucose Using Glucose Oxidase
Treat these tubes exactly as the samples.

Pretreatment of the Samples Samples may be ana- Centrifuge the samples at 500 g for 5 min to pellet any
lyzed in the form in which they are eaten, as in the main precipitate before determination of glucose.
procedure. Alternatively, samples may be finely divided Pipet 100 µL of water, enzyme blank, samples and
by milling or homogenization before analysis, in which standards in duplicate into test tubes containing 2 mL
case the glass balls may be replaced by a magnetic stirring of the glucose oxidase reagent (prepared according to
bar and mixing of the sample achieved with a magnetic the manufacturer’s instructions) and vortex-mix. Place
stirrer instead of a shaking water-bath. the tubes in a water-bath at 37 ° C for 20 min. Measure
the absorbance of the standards and samples in the
Measurement of Free-sugar Glucose Weigh an appro- spectrophotometer at 510 nm against the reagent blank.
priate amount of sample, to the nearest 0.1 mg, into 50-mL The glucose concentration (grams per 100 g of sample) is
10 FOOD

Table 4 Values of V and C for the different and the shaking action is not interrupted until all the
analytes G120 portions have been collected (see below). Repeat
the addition of enzyme mixture for the rest of the sample
Analyte V (mL)a C (mg mL 1 )
tubes, at 1-min intervals to aid timing of the procedure,
G20 25 20 and place them into the shaking water-bath. Remove each
G120 25 20 tube from the bath at exactly 20 min after addition of the
TG 35.4 14.12 enzyme mixture, transfer 0.2 mL into 4 mL of absolute
FSG 25 7.1
TG (rapid) 35.7 17.5 ethanol and vortex-mix to stop the hydrolysis; this is the
G20 portion to be measured in step 4.5.3. Return the
a C1 mL per 1 g of sample. sample tube to the shaking water-bath immediately after
the sample has been taken. After a further 100 min (a total
given by Equation (1): of 120 min of incubation), transfer 0.2 mL from the sample
tube into 4 mL of absolute ethanol and vortex-mix; this is
A(t)VC the G120 portion to be measured in step 4.5.3.
Glucose.%/ D .1/
A(s)W
Dispersion of Resistant Starch Having removed
where A(t) is the absorbance of the test solution, V is all the tubes from the shaking water-bath, vortex-mix
the total volume of the test solution (milliliters), C is the vigorously to break up any large particles, then place
concentration (milligrams per milliliter) of the standard the rack of tubes in a boiling water-bath for 30 min.
used, A(s) is the absorbance of the standard used and W Remove the tubes from the bath, vortex-mix and place the
is the weight (milligrams) of sample taken for analysis tubes in ice – water until thoroughly chilled. Add 10 mL
(this figure may be corrected for moisture), as illustrated of 7-mol L 1 KOH, cap the tube and mix the contents
in Table 4. by inversion. Immerse the tubes horizontally in a shaking
water-bath containing ice – water for 30 min.
4.5 The High-performance Liquid Chromatography
Version (Figure 2) Hydrolysis of Starch to Glucose Remove the tubes
singly from the ice – water and transfer 0.2 mL of the
4.5.1 Measurement of Rapidly Available Glucose, Slowly contents into a tube containing 1 mL of 1-mol L 1 acetic
Digestible Starch, Resistant Starch and Total Starch acid. To these tubes add 40 µL of amyloglucosidase
solution. Mix and place the tubes in a 70 ° C water-bath for
Sample Weight and Internal Standard and Blanks 30 min followed by 10 min in a boiling water-bath. Cool
Weigh an appropriate amount of sample, to the near- the tubes to room temperature before adding 12 mL of
est 0.1 mg, into 50-mL tubes. If the sample is hot, cap the absolute ethanol; this is the TG portion to be measured
tube and place in a water-bath at 70 – 80 ° C until ready to in step 4.5.3.
analyze. Samples should be analyzed at least in duplicate.
Reference samples should be included in every batch (see 4.5.2 Measurement of Free-sugar Glucose; Optional
section 5). Two tubes, one with 4 mL of stock sugar mix- Rapid Measurement of Total Starch
ture added, are included as enzyme blanks. Add 5 mL of
internal standard to the sample and blank tubes and treat Preparation of Enzyme Blanks No enzyme blank
exactly as the samples. is required for the measurement of FSG. For the rapid
measurement of TS, two tubes, one with 4 mL of the stock
Incubation with Pepsin Add 10 mL of the freshly sugar mixture added, are included as enzyme blanks.
prepared pepsin – guar gum solution to each tube, vortex-
mix and place the tubes in a water-bath at 37 ° C for 30 min. Pretreatment of the Samples Samples may be ana-
Remove the samples from the water-bath and add five lyzed in the form in which they are eaten, as in the main
glass balls and 5 mL of 0.5-mol L 1 sodium acetate to each procedure. Alternatively, samples may be finely divided
tube. Shake the tubes gently to disperse the contents and by milling or homogenization before analysis, in which
replace them in the water-bath at 37 ° C to equilibrate. case the glass balls may be replaced by a magnetic stirring
bar and mixing of the sample achieved with a magnetic
Enzymatic Hydrolysis of Starch Remove one sample stirrer instead of a shaking water-bath.
tube from the 37 ° C water-bath and add 5 mL of enzyme
mixture. Immediately cap the tube and mix the contents Measurement of Free-sugar Glucose Weigh an appro-
gently by inversion before securing the tube horizontally priate amount of sample, to the nearest 0.1 mg, into 50-mL
in the 37 ° C shaking water-bath. Start the shaking action tubes and add five glass balls to each. Add 5 mL of internal
of the water-bath; this is time zero for the incubation standard to each sample tube. Pipet 20 mL of 0.1-mol L 1
STARCH ANALYSIS IN FOOD 11

acetate buffer into each tube, cap and vortex-mix vigor- Table 5 Sequence of elution conditions
ously in order to begin breaking up large particles. Place for the HPLC measurement of sugarsa
the rack of tubes in a boiling water-bath. After 30 min,
Switch [NaOH] Time
remove the rack of tubes from the boiling water-bath and position (mmol L 1 ) (min)
vortex-mix vigorously to break up any remaining particles
of sample. Cool to 37 ° C, then add 0.2 mL of invertase to A 10 0 – 3.5
each tube. Cap the tubes and immerse horizontally in the B 70 3.6 – 14.0
B 200 14.1 – 15.0
shaking water-bath at 37 ° C for 30 min. B 10 15.1 – 20.0
Vortex-mix vigorously before removing 0.2 mL of each
a In switch position A, the flow is from
sample or standard solution into a test tube containing
Aminotrap to guard column to separation
4 mL of absolute ethanol; this is the FSG portion to be column. In switch position B, the flow is
measured in step 4.5.3. from guard column to separation column to
If TS is to be determined by the main method, the Aminotrap. Sample injection is at time 0.1 min.
procedure is finished here; otherwise, the procedure can
be continued for a rapid measurement of TS. and a Dionex Carbopac PA10 guard column. Column
switching and a Dionex Aminotrap may be used to
Dispersion and Hydrolysis of Total Starch To the prevent pancreatic peptides from reaching the analytical
remainder of the sample solution add 0.1 mL of heat- column. The eluents, high-purity water and 200 mol L 1
stable amylase, cap the tube, vortex-mix vigorously, then NaOH [16 mL L 1 of 50% (w/v) NaOH solution in high-
place in a boiling water-bath for 15 min. Remove the tubes purity degassed water] should be degassed. The flow-rate
from the bath, vortex-mix and place the tubes in ice – water is 0.8 mL min 1 and the elution conditions are shown
until thoroughly chilled. Add 10 mL of 7-mol L 1 KOH, in Table 5. Monosaccharide detection may be achieved
cap the tube and mix the contents by inversion. Immerse with a Dionex ED40 electrochemical detector with the
horizontally in a shaking water-bath containing ice – water following pulse potentials, E, and durations, t: E1 0.05 V,
for 30 min. t1 400 ms; E2 0.75 V, t2 200 ms; and E3 0.15 V, t3 400 ms. A
Remove the samples from the shaking water-bath and response time of 1 s is used and the output on the detector
immediately transfer 0.2 mL of the contents into a tube is set at 300 nA.
containing 1 mL of 1-mol L 1 acetic acid and mix well.
Add 40 µL of amyloglucosidase solution. Mix and place
4.6 Calculations
the tubes in a 70 ° C water-bath for 30 min followed by
10 min in a boiling water-bath. Cool the tubes to room Values for RAG, SAG, RDS, SDS, RS and TS are
temperature before adding 12 mL of absolute ethanol; calculated from the measured FSG, G20, G120 and
this is the TG portion to be measured in step 4.5.3. TG values according to Equations (2 – 7). Values for the
starch fractions are expressed as polysaccharides, using
a factor of 0.9 to convert the measured glucose value to
4.5.3 High-performance Liquid Chromatography
polysaccharide.
Analysis of Sugars
Prepare two sugar standards in 50-mL tubes: standard 1, RAG D G20 .2/
1 mL of stock sugar mixture, 19 mL of water and 5 mL SAG D G120 G20 .3/
of internal standard; standard 2, 10 mL of stock sugar
mixture, 10 mL of water and 5 mL of internal standard. RDS D 0.9.G20 FSG/ .4/
Mix well and transfer 0.2 mL from each into a tube SDS D 0.9.G120 G20/ .5/
containing 4 mL of absolute ethanol.
Before HPLC analysis, centrifuge all the ethanolic TS D 0.9.TG FSG/ .6/
fractions for 5 min at 500 g. The amount to be taken for RS D 0.9.TG G120/ .7/
analysis varies according to sugar content: typically 70 µL
for the sugar standards and the G20 and G120 portions,
4.7 Measurement of Resistant Starch 1, 2 and 3 (Using
200 µL for the TG portions and 70 – 120 µL for the FSG
Colorimetry or High-performance Liquid
portions. Place the samples in HPLC vials, add 1 mL of
Chromatography)
deionized water and vortex-mix.
A Dionex AS3500 autoinjector may be used for
4.7.1 Resistant Starch 1, Physically Inaccessible Starch
injection of 20 µL of the diluted ethanolic fractions. Sugar
separation may be achieved with a Dionex Carbopac There are three forms of RS: RS1 is present in foods
PA100 column using a Dionex GP40 gradient pump having a dense or rigid structure, e.g. boiled rice, pasta,
12 FOOD

whole-grain bread, maize and legumes; RS2 is present 5 QUALITY CONTROL AND
in raw foods, e.g. bananas, and raw cereal flours and TROUBLESHOOTING
grains, in foods cooked with very little water, e.g. some
dry-baked biscuits, and in legumes; RS3 is present in 5.1 Quality Control
foods that have been cooked and then cooled, e.g. bread,
For reproducible results to be obtained within and
breakfast cereals and cold potatoes. (Note: for a few high- between laboratories, it is essential that stringent quality
amylose products, e.g. Hylon VII, the boiling step in the control is applied. The procedures described here are
procedure will not gelatinize all of the starch. Any starch based on the measurement of the rate and extent
that remains ungelatinized after this step will therefore of starch digestion in vitro and require standardized
be included as RS3.) enzyme preparations and incubation conditions. A kit
A value for total RS is obtained by the main procedure. is available (Englyst Carbohydrate Services Ltd, Cat.
Values for the individual RS1, RS2 and RS3 fractions No. 61-000), which includes the necessary enzymes and
may be obtained by the parallel analysis of three samples three reference samples for calibration purposes. It is
of the test food, which are prepared and analyzed in suggested that reference samples 1 and 2 are included in
different ways. every batch of samples analyzed. The reference samples
Sample A is prepared in the form in which the food have been selected for their specific starch digestion
would normally be eaten, and is then passed through a profiles.
hand-operated mincer. The analytical sample is weighed Reference sample 1 (white wheat flour; Homepride,
into 50-mL tubes and treated as described for the main UK), with its high SDS content, is used to check the
procedure. Only the G120(A) value is required for efficiency of starch hydrolysis during the amylolytic
calculation of the RS1 fraction but it is recommended that incubation. Values for G120 that are too low (see
G20(A) and TG(A) are measured during the procedure target values below) may indicate that the activity
and the values used for calculation of RAG, SAG and of the amylolytic enzymes has decreased. Reference
total RS. sample 2 (raw potato starch; Kartoffelmel Centralen,
Sample B is homogenized for moist foods or milled for Denmark), with its high RS content, is used to establish
dry foods to disrupt the food matrix before analysis but the optimum stroke speed of the shaking water-bath
ball milling should not be used, as this may damage starch during the amylolytic incubation. If the G120 value
granules. The analytical sample is weighed into 50-mL for the potato starch is too high (see target values in
tubes and treated as described for the main procedure. Table 6), decrease the stroke speed and vice versa, e.g. we
Only the G120(B) value is required for calculation of the observe a linear increase in the G120 value for reference
RS2 fraction. sample 2 from 12.4 to 36.4 g per 100 g between 120 and
Sample C is homogenized or milled as described for 170 strokes min 1 .
sample B. The analytical sample is weighed into 50-mL Reference sample 3 (corn flakes; Kellogg’s, UK),
tubes and treated as described for the main procedure up with its high RDS content, is used to check the
to and including the addition of sodium acetate. At this efficiency of starch and maltose hydrolysis during
stage, the tubes are placed in boiling water for 30 min to the first stage of amylolytic incubation. Values for
gelatinize any native starch granules that would normally G20 that are too low (see target values below)
be gelatinized during cooking. The tubes are cooled may indicate that the amyloglucosidase activity has
to 37 ° C and the remainder of the main procedure is decreased.
followed to obtain the G120(C) value that is required for Target values for the three reference samples are given
calculation of the RS2 and RS3 fractions. The analysis is in Table 6.
continued to obtain a TG(C) value, which is required for The method yields coefficients of variation of 3.7%
calculation of the RS3 fraction. The step in the boiling intraassay and 6.6% interassay for the measurement of
water-bath in the main procedure after the G120(C) RAG, 2.2% and 3.9% for G120 (RAG C SAG) and 1.6%
subsample has been withdrawn may be omitted, instead and 2.5% for TG for the three reference samples.
continuing with the dispersion of RS3 by treatment with
KOH. Table 6 Target values (grams per 100 g) for the reference
Calculations are performed using Equations (8 – 10). samples

Reference sample G20 G120 TG

RS1 D 0.9[G120(B) G120(A)] .8/
Potato starch 4 (š0.5) 26 (š1) 89 (š1)
RS2 D 0.9[G120(C) G120(B)] .9/ Wheat flour 35 (š1) 77 (š1) 78 (š1)
Corn flakes 79 (š1) 81 (š1) 85 (š1)
RS3 D 0.9[TG(C) G120(C)] .10/
STARCH ANALYSIS IN FOOD 13

5.2 Troubleshooting 5.2.3 Troubleshooting for High-performance Liquid

Chromatography Procedure
5.2.1 Troubleshooting for the Common Hydrolysis Steps
Variation Between Replicates This may be caused
Accurate timing during subsampling for G20 and by inaccurate pipeting during the addition of internal
G120 measurements is critical. Pitted glass balls may standard to the samples.
increase the mechanical disruption of the samples,
and should be replaced. Failure to ensure that tubes Spurious Peaks on the Chromatogram Peaks that
are held exactly horizontal in the shaking water-bath significantly interfere with the integration of sugar peaks
and exactly parallel with the direction of movement may be due to amino acids or peptides in the sample
may result in differences in mechanical disruption of that have eluted on to the separator column from the
samples. guard column intended for their retention. Clean the
guard column or alter the timing of the column-switching
Variation in Reference Sample Values Between Batches mechanism.
Use the reference samples to check enzyme activity
levels. Consistently low results may indicate incor-
rect pH during hydrolysis: check the buffer system. 6 METHOD DEVELOPMENT AND
Variation in G120 values between batches for refer- VALIDATION
ence sample 2 may be due to changes in the shaking
speed or the temperature of the water-bath. It is rec- The measurements of starch fractions obtained by the
ommended that the shaking speed and temperature procedures described here characterize starchy foods in
controls are physically fixed once optimized to the target terms that reflect the average rate and extent of starch
values. digestion in the human gut. Owing to the variation in
both the rate and extent of starch digestion that is seen
Variation in Replicate Analysis The sample hetero- both within and between individuals in human studies,
geneity of many foods analyzed ‘‘as eaten’’ can make it the measurement of RS is based on the average of
difficult to take fully representative subsamples, leading measurements observed in vivo. The measurement of
to greater variation in results than for homogenized or RAG, which includes RDS but is not limited to starch,
milled samples. It is recommended that an increased has been shown to be correlated with in vivo glycemic
number of replicates are analyzed for such problem response measurements..19/
samples.
6.1 Methodological Considerations
5.2.2 Troubleshooting for the Colorimetric Procedure This section provides a brief discussion of the develop-
ment of methodology for the measurement of nutrition-
Variation in Replicate Analysis This may be due ally important starch fractions. A full description has been
to inaccurate pipeting when taking subsamples for given in the papers cited below.
measurement of G20, G120 and TG, due to the presence For foods with a well-defined particulate structure, the
of large food particles in the sample and/or the high rate and extent of starch digestion are critically dependent
viscosity of the suspension. The reproducibility of pipeting on the way the food sample is divided and the method
may be improved by cutting off the end of the pipet chosen should reflect the average disruption of food
tip to produce a larger orifice and by the repeated structure achieved by chewing. Comparison of the starch
washing of the pipet tip (after wiping the outside) into digestibility in food samples that were chewed before
the ethanol fraction to ensure complete transfer of the analysis and those minced as described here showed that,
subsample. although both gave similar mean values, the values for
the minced samples had a smaller standard deviation..18/
Variation in Absorbance Measurements Accurate The addition of glass balls to the incubation tubes
pipeting is essential, and may be improved by repeatedly provides further mechanical disruption of particulate
washing the tip (after wiping the outside) into the color matter during the hydrolysis with pancreatic amylase.
reagent to ensure complete transfer of the subsample. This allowed the determination of in vitro RS val-
Ensure that each set of samples is analyzed with suitable ues for polished rice commensurate with the starch
standards and blanks, and that the interval between recovered in ileostomy studies with similar prod-
the addition of sample to the color reagent and the ucts. To prevent excessive grinding of starch gran-
measurement of the absorbance is identical for all the ules by the glass balls, the viscosity of the sam-
samples, standards and blanks. ple mixture was increased by the addition of guar
14 FOOD

gum, which keeps the sample material in suspension. 175

7
The amounts of the enzymes used in this procedure
are calculated to be present in excess so that they 150

Glycemic response (iAUC)

8
are not rate-limiting in the hydrolysis of starch to
glucose. 125
Three approaches are used to disperse the starch
fraction that resists hydrolysis during the incubation with 100 6
pancreatic a-amylase. Physical disruption of particulate 45
3
75
material (RS1) by vigorous vortex-mixing at several
stages is aided by the presence of the glass balls. A 2
50 1
30 min boiling step is used to gelatinize starch granules
(RS2). Retrograded amylose (RS3) is dispersed by 30 min
25
treatment with 2-M KOH, at 0 ° C to avoid the destruction
of sugars that can occur at higher temperatures. These
0
treatments in combination have been shown to totally 0 10 20 30 40 50
disperse starch in all but the hardest whole grains, which RAG intake (g)
may require milling to obtain a reliable value for TS.
Figure 3 Correlation of glycemic response and RAG content
6.2 Validation of Rapidly Digestible Starch and Slowly of test meals. The line is fit by least-squares analysis to show
the correlation between the mean glycemic response, measured
Digestible Starch Values as the incremental area under the curve (iAUC; error bars
indicate š standard error of the mean), and the eight intakes of
In the procedure described here, the analytical time
RAG over all subjects (n D 8). The numerals indicate small and
points of 20 and 120 min for the measurement of RDS large portions, respectively, of the different test meals: 1 and 3,
and SDS are not intended to represent that time of pearled barley; 2 and 6, spaghetti; 4 and 7, corn flakes; 5 and 8,
passage through the human gut. The choice of 20 min white bread.
of incubation for the measurement of RDS is based on
observed rates of hydrolysis of starch in vitro. For foods measurements of RAG and SAG, and the corresponding
in which starch is rapidly digested, such as bread and corn starch fractions, RDS and SDS, have physiological
flakes, the majority of starch is digested by 20 min under significance..19/
the conditions described here, whereas a considerable
amount of starch in foods such as spaghetti and pearled
6.3 Validation of Resistant Starch Values
barley is digested between 20 and 120 min.
A recent study has shown a good correlation between Quantifying the amount of starch entering the human
glycemic index values from the literature and in vitro colon presents a considerable technical problem. Hydro-
RAG values, which include RDS, and RAG is proposed gen excreted in breath has been used as an indicator
as an important food-related determinant of the glycemic of fermentation in the colon, but it is considered to
response..34/ lack the sensitivity required for quantification. Intuba-
More direct evidence for this was obtained when the tion has been used to sample digesta from the ileum,
glycemic response to four starchy foods with different but these studies were restricted to liquidized meals.
proportions of RDS and SDS was investigated. Two Several studies have used human ileostomy subjects as
of the foods, pearled barley and spaghetti, have a low a model to investigate the digestive physiology of the
proportion of starch as RDS, whereas corn flakes and small intestine. Carbohydrate analysis of the ileostomy
white bread contain a high proportion of starch as effluent allows determination of the amount of starch
RDS. The four foods were each fed in two portion and starch digestion products escaping digestion in the
sizes, providing 25 and 50 g of glucose (from RDS, SDS small intestine..35 – 37/ The amount of starch recovered
and FSG) likely to be available for absorption in the in the ileostomy effluent varied between individuals by
small intestine. The eight test meals represented a range š20% around the mean. It is from these studies that
of intake from 11 to 49 g of RAG. There was a very the currently accepted definition of RS is derived ‘‘the
strong correlation (P < 0.001) between RAG intake and sum of starch and starch degradation products that, on
glycemic response, with RAG explaining 70% of the average, reach the human large intestine’’..16/ The in
variance in glycemic response after subject variation was vitro methodology has been tuned to yield values for
accounted for (Figure 3)..19/ RS that match the mean proportion of starch recovered
The results of this study and the good agreement in ileostomy studies for both single foods and mixed
between RAG and glycemic index strongly suggest that meals..14,35 – 38/
STARCH ANALYSIS IN FOOD 15

7 COMPARISON WITH OTHER METHODS ž standardized conditions of amylolytic hydrolysis;

ž development and validation of methods in conjunc-
7.1 Total Starch tion with in vivo studies.
A discussion of the principles of starch analysis has The method described here is based on these principles,
been provided by Southgate..39/ Most methods for the and incorporates in vitro measurements that reflect
measurement of TS are dependent on acid or enzymatic the likely rate and extent of starch digestion in the
hydrolysis of starch and measurement of the released human gut.
glucose. Generally, enzymatic approaches are more
specific and do not result in destruction of sugars.
Samples are usually finely divided and may be boiled
ABBREVIATIONS AND ACRONYMS
to gelatinize starch before enzymatic hydrolysis with a-
amylase and/or amyloglucosidase. However, retrograded
amylose is not accessible to enzymatic hydrolysis before DMSO Dimethyl Sulfoxide
it has been dispersed with DMSO or 2-M KOH..22/ FAO Food and Agriculture Organization
Procedures that do not contain an efficient dispersal step FSG Free-sugar Glucose
will underestimate the starch content of some processed G20 The Glucose Measured in the Portion
foods containing RS3. The values obtained by such Removed from the Analysis at 20 min After
methods will therefore be lower than those obtained the Addition of Enzymes
by the method described here. G120 The Glucose Measured in the Portion
Removed from the Analysis at 120 min After
the Addition of Enzymes
7.2 Rate and Extent of Starch Digestion
HPLC High-performance Liquid Chromatography
The increased knowledge of the different metabolic NSP Nonstarch Polysaccharides
fates of dietary starch has resulted in the development RAG Rapidly Available Glucose
of several methods for assessing the rate of starch RDS Rapidly Digestible Starch
digestion..8 – 13,15,18,22,40 – 43/ RS Resistant Starch
The basis of the methods investigating the rate of starch RS1 Physically Inaccessible Starch
digestion is the measurement of starch hydrolysis after a RS2 Resistant Starch Granules
specified period of incubation with amylolytic enzymes. RS3 Retrograded Starch
Meaningful comparison between methods is difficult, SAG Slowly Available Glucose
owing to differences in sample preparation, enzymes SDS Slowly Digestible Starch
and hydrolysis conditions. However, the ranking between TG Total Glucose
foods is often similar between methods, and most have TS Total Starch
shown strong correlations with glycemic response data. WHO World Health Organization
The RDS and SDS terms described here are at present
the only measures of rate of starch digestion expressed in
grams per 100 g of food. RELATED ARTICLES
The basis of methods for the determination of RS
is the measurement of starch remaining unhydrolyzed
after a defined period of incubation with amylolytic Food (Volume 5)
enzymes. Measurements of total RS must include RS1, Dietary Fiber Analysis as Non-starch Polysaccharides
RS2 and RS3. However, owing to inappropriate sample
preparation and gelatinization of starch during the
procedure, many methods measure only RS3 or RS2 REFERENCES
and RS3 fractions. In addition, few of the methods have
been developed and validated in conjunction with in vivo 1. D.J. Gallant, B. Bouchet, A. Buleon, S. Perez, ‘Physical
studies. Characteristics of Starch Granules and Susceptibility to
Methods designed to investigate the rate and/or the Enzymatic Degradation’, Eur. J. Clin. Nutr., 46, Suppl. 2,
extent of starch digestion in the human gut should S3 – S16 (1992).
incorporate the following principles: 2. J.R. Katz, ‘The Amorphous Part of Starch in Fresh Bread,
and in Fresh Pastes and Solutions of Starch’, Recl. Trav.
ž analysis of foods prepared ‘‘as eaten’’; Chim. Pays Bas, 18, 55 – 59 (1937).
ž reproducible disruption of the physical structure of 3. M.J. Gidley, ‘Factors Affecting the Crystalline Type
food; (A – C) of Native Starches and Model Compounds:
16 FOOD

A Rationalisation of Observed Effects in Terms of 17. J.H. Cummings, H.N. Englyst, ‘Fermentation in the
Polymorphic Structures’, Carbohydr. Res., 161, 301 – 304 Human Large Intestine and the Available Substrates’,
(1987). Am. J. Clin. Nutr., 45, 1243 – 1255 (1987).
4. H. Fuwa, T. Takaya, Y. Sugimoto, ‘Degradation of Vari- 18. H.N. Englyst, S.M. Kingman, J.H. Cummings, ‘Classifica-
ous Starch Granules by Amylases’, in Mechanisms of Sac- tion and Measurement of Nutritionally Important Starch
charide Polymerization/Depolymerization, ed. J.J. Mar- Fractions’, Eur. J. Clin. Nutr., 46, S33 – S50 (1992).
shall, Academic Press, New York, 73 – 100, 1980. 19. K.N. Englyst, H.N. Englyst, G.J. Hudson, T.J. Cole, J.H.
5. P. Colonna, V. Leloup, A. Buleon, ‘Limiting Factors of Cummings, ‘Rapidly Available Glucose in Foods. A Mea-
Starch Hydrolysis’, Eur. J. Clin. Nutr., 46, Suppl. 2, surement in Vitro that Reflects the Glycemic Response’,
S17 – S32 (1992). Am. J. Clin. Nutr., 69, 448 – 454 (1999).
6. D.J.A. Jenkins, T.M.S. Wolever, R.H. Taylor, H. Barker, 20. R.A. McCance, R.D. Lawrence, The Carbohydrate Con-
H. Fielen, J.M. Baldwin, A.C. Bowling, H.C. Newman, tent of Foods, MRC Special Report, HMSO, London,
A.L. Jenkins, D.V. Goff, ‘Glycemic Index of Foods: A 1929.
Physiological Basis for Carbohydrate Exchange’, Am. J. 21. D.A.T. Southgate, ‘Determination of Carbohydrates in
Clin. Nutr., 34, 362 – 366 (1981). Foods. II. Unavailable Carbohydrates’, J. Sci. Food Agric.,
7. K. O’Dea, P.J. Nestel, L. Antonoff, ‘Physical Factors 20, 331 – 335 (1969).
Influencing Postprandial Glucose and Insulin Responses 22. H.N. Englyst, H.S. Wiggins, J.H. Cummings, ‘Determina-
to Starch’, Am. J. Clin. Nutr., 33, 760 – 765 (1980). tion of the Non-starch Polysaccharides in Plant Foods by
8. K. O’Dea, P. Snow, P. Nestel, ‘Rate of Starch Hydrolysis Gas – Liquid Chromatography of Constituent Sugars as
in Vitro as a Predictor of Metabolic Responses to Alditol Acetates’, Analyst, 107, 307 – 318 (1982).
Complex Carbohydrate in Vivo’, Am. J. Clin. Nutr., 34, 23. H.N. Englyst, M.E. Quigley, G.J. Hudson, J.H. Cum-
1991 – 1993 (1981). mings, ‘Determination of Dietary Fiber as Non-starch
9. D.J.A. Jenkins, T.M.S. Wolever, R.H. Taylor, H. Ghafari, Polysaccharides by Gas – Liquid Chromatography’, Ana-
A.L. Jenkins, H. Barker, M.J.A. Jenkins, ‘Rate of Diges- lyst, 117, 1707 – 1714 (1992).
tion of Foods and Postprandial Glycaemia in Normal and 24. M.E. Quigley, H.N. Englyst, ‘Determination of Neutral
Diabetic Subjects’, Br. Med. J., 2, 14 – 17 (1980). Sugars and Hexosamines by High-performance Liquid
10. K.W. Heaton, S.N. Marcus, P.M. Emmett, C.H. Bolton, Chromatography with Pulsed Amperometric Detection’,
‘Particle Size of Wheat, Maize and Oat Test Meals: Analyst, 117, 1715 – 1718 (1992).
Effects on Plasma Glucose and Insulin Responses and 25. H.N. Englyst, M.E. Quigley, G.J. Hudson, ‘Determina-
on the Rate of Starch Digestion in Vitro’, Am. J. Clin. tion of Dietary Fiber as Non-starch Polysaccharides with
Nutr., 47, 675 – 682 (1988). Gas – Liquid Chromatographic, High-performance Liquid
11. J. Holm, I. Lundquist, I. Bjork, A.C. Eliasson, N.-G. Asp, Chromatographic or Spectrophotometric Measurement
‘Degree of Starch Gelatinization, Digestion Rate of of Constituent Sugars’, Analyst, 119, 1497 – 1509 (1994).
Starch in Vitro, and Metabolic Response in Rats’, Am. J. 26. M.E. Quigley, H.N. Englyst, ‘Determination of the Uro-
Clin. Nutr., 47, 1010 – 1016 (1988). nic Acid Constituents of Non-starch Polysaccharides by
12. F. Bornet, A. Fontvieille, S. Rizkalla, P. Colonna, High-performance Liquid Chromatography with Pulsed
A. Blayo, C. Mercier, G. Slama, ‘Insulin and Glycemic Amperometric Detection’, Analyst, 119, 1511 – 1518
Responses in Healthy Humans to Native Starches Pro- (1994).
cessed in Different Ways, Correlation with in Vitro 27. G.B. Haber, K.W. Heaton, D. Murphy, L.F. Burroughs,
a-Amylase Hydrolysis’, Am. J. Clin. Nutr., 50, 315 – 323 ‘Depletion and Disruption of Dietary Fiber. Effects on
(1989). Satiety, Plasma Glucose and Serum Insulin’, Lancet, i,
13. Y. Granfeldt, I. Bjork, A. Drews, J. Tovar, ‘An in Vitro 679 – 682 (1977).
Procedure Based on Chewing to Predict Metabolic 28. FAO/WHO, ‘Carbohydrates in Human Nutrition. Report
Response to Starch in Cereal and Legume Products’, of a Joint FAO/WHO Expert Consultation, Rome, 14 – 18
Eur. J. Clin. Nutr., 46, 649 – 660 (1992). April 1997’, FAO Food and Nutrition Paper 66, FAO,
14. H.N. Englyst, S.M. Kingman, G.J. Hudson, J.H. Cum- Rome, 1998.
mings, ‘Measurement of Resistant Starch in Vitro and 29. W.C. Willett, ‘The Dietary Pyramid: Does the Foundation
in Vivo’, Br. J. Nutr., 75, 749 – 755 (1996). Need Repair’, Am. J. Clin. Nutr., 68, 218 – 219 (1998).
15. I. Goni, A. Garcia-Alonso, F. Saura-Calixto, ‘A Starch 30. J. Salméron, A. Ascherio, E.B. Rimm, G.A. Colditz,
Hydrolysis Procedure to Estimate Glycemic Index’, Nutr. D. Spiegelman, D.J. Jenkins, M.J. Stampfer, A.L. Wing,
Res., 17, 427 – 437 (1997). W.C. Willett, ‘Dietary Fiber, Glycemic Load, and Risk of
16. H.N. Englyst, S.M. Kingman, ‘Dietary Fiber and Resis- NIDDM in Men’, Diabetes Care, 20, 545 – 550 (1997).
tant Starch. A Nutritional Classification of Plant 31. J. Salméron, J.E. Manson, M.J. Stampfer, G.A. Colditz,
Polysaccharides’, in Dietary Fiber, eds. D. Kritchevsky, A.L. Wing, W.C. Willett, ‘Dietary Fiber, Glycemic Load,
C. Bonfield, J.W. Anderson, Plenum Press, New York, and the Risk of NIDDM in Women’, J. Am. Med. Assoc.,
49 – 65, 1990. 277, 472 – 477 (1997).
STARCH ANALYSIS IN FOOD 17
32. D.R. Jacobs, K.A. Meyer, L.H. Kushi, A.R. Folsom, 38. K.R. Silvester, H.N. Englyst, J.H. Cummings, ‘Ileal Re-
‘Whole Grain Intake May Reduce the Risk of Ischemic covery of Starch from Whole Diets Containing Resistant
Heart Disease Death in Postmenopausal Women: The Starch Measured in Vitro and Fermentation of Ileal
Iowa Women’s Health Study’, Am. J. Clin. Nutr., 68, Effluent’, Am. J. Clin. Nutr., 62, 403 – 411 (1995).
248 – 257 (1998). 39. D.A.T. Southgate, Determination of Food Carbohydrates,
33. J. Brand-Miller, ‘Importance of Glycemic Index in Dia- 2nd edition, Elsevier Applied Science, London, 1991.
betes’, Am. J. Clin. Nutr., 59, Suppl, S747 – S752 (1994). 40. I. Bjork, M. Nyman, B. Pedersen, M. Siljestrom, N.-
34. H.N. Englyst, J. Veenstra, G.J. Hudson, ‘Measurement G. Asp, B.O. Eggum, ‘On the Digestibility of Starch in
of Rapidly Available Glucose (RAG) in Plant Foods: A Wheat Bread – Studies in Vitro and in Vivo’, J. Cereal
Potential in Vitro Predictor of the Glycaemic Response’, Sci., 4, 1 – 11 (1986).
Br. J. Nutr., 75, 327 – 337 (1996). 41. C.S. Berry, ‘Resistant Starch: Formation and Measure-
35. H.N. Englyst, J.H. Cummings, ‘Digestion of the Polysac- ment of Starch that Survives Exhaustive Digestion
charides of Some Cereal Foods in the Human Small with Amylolytic Enzymes During the Determination of
Intestine’, Am. J. Clin. Nutr., 42, 778 – 787 (1985). Dietary Fiber’, J. Cereal Sci., 4, 301 – 314 (1986).
36. H.N. Englyst, J.H. Cummings, ‘Digestion of the Carbo- 42. J.G. Muir, K. O’Dea, ‘Measurement of Resistant Starch:
hydrates of Banana (Musa paradisiaca sapientum) in the Factors Affecting the Amount of Starch Escaping Diges-
Human Small Intestine’, Am. J. Clin. Nutr., 44, 42 – 50 tion in Vitro’, Am. J. Clin. Nutr., 56, 123 – 127 (1992).
(1986). 43. N. Faisant, D.J. Gallant, B. Bouchet, M. Champ, ‘Banana
37. H.N. Englyst, J.H. Cummings, ‘Digestion of the Polysac- Starch Breakdown in the Human Small Intestine Studied
charides of Potato in the Small Intestine of Man’, Am. J. by Electron Microscopy’, Eur. J. Clin. Nutr., 49, 98 – 104
Clin. Nutr., 45, 423 – 431 (1987). (1995).