Available online at www.sciencedirect.com
ScienceDirect
Rice Science, 2017, 24(5): 274í282
Estimating Glycemic Index of Rice-Based Mixed Meals by
Using Predicted and Adjusted Formulae
Nur Maziah Hanum OSMAN1, Barakatun-Nisak MOHD-YUSOF1, 2, Amin ISMAIL1, 2
(1Department of Nutrition and Dietetics, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang 43400,
Selangor, Malaysia; 2Research Centre of Excellent for Nutrition and Non-communicable Diseases, Faculty of Medicine and Health
Sciences, Universiti Putra Malaysia, Serdang 43400, Selangor, Malaysia)
Abstract: The estimation of glycemic index (GI) of rice-based mixed meal either by using predicted GI
(GIpred) or adjusted GI (GIadj) formula is unclear. This study aimed to determine the glycemic response of
rice in rice alone or mixed meals and to identify the appropriate formula for estimating the GI of rice-based
mixed meals. The glycemic responses produced by the rice alone (red rice, fragrant white rice or
parboiled rice) and the rice-based mixed meals (fried red rice, fried fragrant white rice or fried parboiled
rice) which provided 25 g available carbohydrate were assessed in 11 healthy individuals. To determine
the measured GI (GImeasured) of rice alone and rice-based mixed meals, participants underwent three
repeated tests of a reference food (Glucolin®). Tests were performed in random order on nine separate
visits after an overnight fasting for at least 8 h. Capillary glucose at baseline (0 min), 15, 30, 45, 60, 90
and 120 min from starting the meals was assessed and used to determine the incremental area under the
curve (iAUC120). The agreement between GImeasured and the estimation formulae (GIpred or GIadj) were
determined using Bland-Altman analysis. The iAUC120 after consuming rice alone was significantly higher
than the rice-based mixed meals except for fried fragrant rice, which was comparable to the rice alone
(P > 0.05). The GImeasured values of rice were categorized as medium (61 for parboiled rice, 67 for fragrant
white rice, and 68 for red rice). GIpred (r = 0.40, P < 0.01) and GIadj (r = 0.41, P < 0.01) were significantly
correlated with iAUC120. The agreement between GImeasured and GIadj is apparent suggesting the
usefulness of GIadj in estimating meal GI of rice-based mixed meals.
Key words: glycemic index; mixed meal; glycemic response; rice
The glycemic index (GI) is a method used to classify
dietary carbohydrates based on their effect on
postprandial blood glucose levels (Wolever, 2013).
Low GI foods have been shown to reduce the risk of
chronic diseases, in particular, type 2 diabetes mellitus
(Bhupathiraju et al, 2014). The GI concept has also
been used in the management of type 2 diabetes mellitus
to optimise glycemic control (Chen et al, 2015).
Rice (Oryza sativa L.) is the staple food for more
than half of the world’s population mainly from the
Asian region. Increased white rice consumption has
been associated with increased risk of type 2 diabetes,
especially among Asian populations (Hu et al, 2012).
In Malaysia, more than 97% of its adult population
eats rice twice daily mostly in the form of white rice
(Abdul Karim et al, 2008). Many types of rice are
available including fragrant white rice, parboiled rice
and red rice. GI value of rice shows a wide variation
ranging from 48 to 92 with an average of 64, depending
on the type of rice (Atkinson et al, 2008). The variation
in GI value from country to country is probably due to
the botanical effects (Foster-Powell et al, 2002). Hence,
it is suggested that each country determines the GI
values of its common types of rice (Singh et al, 2010).
The GI concept is a property of individual
carbohydrate and its utility in the context of mixed
Received: 30 October 2016; Accepted: 15 June 2017
Corresponding author: Barakatun-Nisak MOHD-YUSOF (bnisak@upm.edu.my)
Copyright © 2017, China National Rice Research Institute. Hosting by Elsevier B.V.
B V This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer review under responsibility of China National Rice Research Institute
http://dx.doi.org/
http://dx.doi.org/10.1016/j.rsci.2017.06.001
Nur Maziah Hanum OSMAN, et al. Estimating Glycemic Index of Rice-Based Mixed Meals
meal is controversial. The meal GI should be
estimated using appropriate formulae rather than
directly measured. And it can be estimated using
predicted (GIpred) or adjusted (GIadj) formula (Wolever,
2013). GIpred considers the carbohydrate content of the
meal, while GIadj takes into account not only the
carbohydrates but also fats and proteins when
estimating the meal GI (Wolever, 2013). Adding fat
and protein to carbohydrate may reduce its GI value
which does not reflect the actual effect of the food’s
GI (Wolever, 2013). Studies have evaluated the use of
GIpred and GIadj. However, results are not consistent.
Some studies observed that the GIpred overestimates
the GI value of the meals (Dodd et al, 2011; Hatonen
et al, 2011). Other studies found that both formulae
are useful in estimating the meal GI (Robert and
Ismail, 2012; Sun et al, 2014).
The GI values of commonly consumed rice in
Malaysia were previously determined (Yusof et al,
2005). However, its applicability in the context of
mixed meals is unclear. This study aimed to evaluate
the glycemic response of rice-based diets either in the
form of rice alone or mixed meal, and to identify the
appropriate formula to estimate the GI of the
rice-based mixed meal. We hypothesised that the GI
value of rice alone could be applied in mixed meals.
MATERIALS AND METHODS
Study design and subjects
This is an experimental crossover study conducted at
Universiti Putra Malaysia. The Institution’s Committee
on Human Studies approved the study protocol and
participants provided their informed consent before
enrolment. Eligible participants were healthy men and
women between age 20 and 30 years, normal body
mass index (BMI, 19.0–24.9 kg/m2) and normal fasting
glucose level (less than 5.6 mmol/L). Participants with
chronic diseases, smokers, pregnancy or lactation and
275
the used of medication that affects glucose metabolism
(steroid) were excluded. A total of 63 individuals were
screened for eligibility with 13 of them met our
inclusion criteria. All of them enrolled, from which 11
participants completed all study visits. Two participants
were unable to complete the study due to medical (n =
1) and personal (n = 1) reasons.
Test meals
The types of rice tested were fragrant white rice
(Super fragrant AAA, Thailand), red rice (Jasmine
Nutri Rice, Thailand) and parboiled rice (Faiza
Basmati, Malaysia). As part of the mixed meal study,
these three types of rice were stir-fried individually by
adding fat and protein (fried fragrant white rice, fried
red rice, and fried parboiled rice). Fried rice is usually
consumed during breakfast in Malaysia. Protein and
fat sources were kept constant in all the tested meals.
The rice to water ratio followed the cooking
instruction from the packaging (Table 1). Nutrition
information of the test rice is shown in Table 1.
Cooking procedure and duration were consistent for
each type of rice. Rice was cooked well using an
electronic cooker and cooking time was set as 35 min.
For the preparation of the mixed meal, rice was
stir-fried using a frying pan. Fried rice was cooked
according to a general preparation for fried rice. The
ingredients, besides rice, included oil, garlic, onion,
oyster sauce, mixed vegetables and egg. All the meals
were prepared on the morning of the test day and
served warm. All of the meals were served with 220
mL water. A glucose solution was diluted with 220
mL water (Glucolin®, Boots Company, Nottingham,
United Kingdom) and used as a reference food. A
smaller portion (25 g) rather than 50 g available
carbohydrate was used to ensure that participants
completed the test meals within the allocated time.
The proximate analysis of rice and the analyses of
total dietary fibre content of rice alone were performed
by a certified private laboratory. The methods for the
Table 1. Nutrient composition and rice to water ratio of test rice.
Food
Weight (g)
Energy (kcal)
Carbohydrate (g)
Fat (g)
Red rice
84
114
25.00 (88.5)
0.16 (0.6)
Fragrant white rice
77
110
25.00 (91.5)
0.15 (0.6)
Parboiled rice
110
113
25.00 (89.0)
0.22 (0.8)
Fried red rice a
135
247
25.20 (55.6)
13.24 (29.2)
Fried fragrant white rice a
129
243
25.10 (56.5)
13.20 (29.7)
Fried parboiled rice a
157
246
25.10 (55.8)
13.20 (29.3)
a
Estimated using Nutritionist ProTM (First Data Bank Inc, Washington, USA).
Values in the parentheses are the contributive percentage for energy.
Protein (g)
Fiber (g)
Rice to water ratio
3.08 (10.9)
2.16 (7.9)
2.86 (10.2)
6.90 (15.2)
6.10 (13.8)
6.70 (14.9)
1.50
0.53
0.10
1.07
0.32
0.09
1:2
1:1
1:1
–
–
–
276
Rice Science, Vol. 24, No. 5, 2017
determination of proximate analysis were based on
AOAC (1998). Total dietary fibre was analysed based
on the methods of Asp (2001) and Prosky et al (1985).
Available carbohydrate was calculated by subtracting
total dietary fibre from the total carbohydrate (Australian
Standard, 2007). Meanwhile, nutrient analyses for rice
based-mixed meals were calculated using Nutritionist
ProTM (First Data Bank Inc, Washington, USA).
following formula (Wolever, 2013):
Study procedures
Precision of measured glycemic index
GImeasured is largely determined by within-subject
variation (Australian Standard, 2007). The coefficient
variation (CV = 100 × SD / Mean) of the iAUC after
the repeated tests of the reference food was calculated
to ensure the precision of the GImeasured. If the mean
CV is higher than 30%, one outlying result for the
reference test in each subject is deleted (International
Standard Organisation, 2010).
Participants underwent the study procedures on nine
separate visits: three tests of rice alone, three tests of
rice in mixed meals and three repeated tests of glucose
(Glucolin®, the Boots Company, Nottingham, United
Kingdom). The sequence of the nine meals was
randomly assigned with a washout period of at least
three days between visits. The tests were conducted
over nine weeks. Participants were asked to maintain
their usual food intake and physical activity pattern
throughout the study period. A 24-h dietary recall was
performed by a research dietitian at each visit to
confirm the consistency of dietary patterns.
Visits were conducted in the morning after an
overnight fasting for at least 8 h. During the study
visit, capillary fasting glucose was measured twice
with 5 min apart and the average was taken as the
tested value at 0 min. The average fasting glucose was
used as the incremental area under the curve (iAUC120)
calculation. Participants were required to consume the
test meals at a comfortable pace within 12 to 14 min,
and they remained seated throughout the duration of
the study. Blood samples were collected at 15, 30, 45,
60, 90 and 120 min from the start of the meal.
Capillary blood samples were obtained using finger
pricks with the used of the lancet (Accucheck Safe
TProTM Indianapolis, North America) and collected
into Microtainer containing NaFl, EDTA (Becton,
Dickinson, and Company, New Jersey, USA). Blood
glucose was measured using the enzyme-coupled
system by glucose oxidase method (YSI STAT 2300,
Yellow Springs, OH, USA).
Data analysis
Measured glycemic index value (GImeasured)
GImeasured was calculated based on the method
described by Australian Standard (2007). The blood
glucose responses for every point of time over 2 h
were used to derive the incremental area under the
curve (iAUC120) ignoring the area beneath the baseline
level. GImeasured value was calculated using the
GImeasured = (iAUCrice / iAUCglucose) × 100
GImeasured value for each meal was calculated as the
average of all individual participants’ GI value.
Participants who had a GI value more than mean ±
2SD were excluded from the analyses and known as
outliers. Final GI value was recalculated.
Glycemic index estimation of rice-based mixed meals
Estimation of GI was based on available carbohydrate
present which was mixed vegetables (3 g) consisting
of corn, peas and carrot. The GI used were as corn
(GI = 60), carrot (GI = 16) and peas (GI = 22) (Sydney
University Glycemic Index Research Service, 2014).
Another 22 g available carbohydrate was contributed
by rice. The GI value of rice in mixed meals was
estimated using predicted (GIpred) and adjusted (GIadj)
formulae. GIpred was calculated by GImeasured value for
rice alone using the formula reported by Dodd et al
(2011). The calculation is as follows:
GIpred = [(GI of fooda × Available carbohydrate of
fooda) + (GI of foodb × Available carbohydrate of
foodb)] + x + x ….. / Total available carbohydrate of
meal
where x indicates the contribution of GI for each
food and available carbohydrate component in mixed
meals.
GIadj was calculated based on GIpred. The reduction
in GIadj was based on the amounts of protein and fat in
the meal (Wolever, 2013). The adjustment factors for
protein and fat were 1.45 and 0.29, respectively
(Wolever, 2013). The adjustment factors for fat and
protein were based on the mean percentage reduction
in iAUC/g factor.
Statistical analysis
Data was analysed using SPSS (version 22.0 Inc.,
Chicago, USA) and criterion for significance was set
at P < 0.05. Shapiro-Wilks test confirmed the normality.
The blood glucose response, iAUC and GI values,
Nur Maziah Hanum OSMAN, et al. Estimating Glycemic Index of Rice-Based Mixed Meals
277
Table 2. Coefficient variation (CV) calculation for iAUC of three repeated test of glucose for each participant.
Participant ID
Glucose iAUC 1 (mmol·min/L)
Glucose iAUC 2 (mmol·min/L)
Glucose iAUC 3 (mmol·min/L)
CV (%)
1
117
127
144
10.55
2
141
222
162
24.02
3
270
205
168
24.09
4
176
218
167
14.55
5
268
260
324
12.27
6
215
137
208
23.11
7
117
177 a
90
18.44
8
171
114
117
23.93
9
214
210
170
12.28
10
179
272
183
24.87
11
132
234
187
27.69
Average CV b
22.23 (20.23)
a
The value is excluded to obtain individual of CV < 30%. b The value in the parenthesis is the average CV after removing one outlying result in one
participant. Glucose iAUC 1, 2 and 3 refer to the three repeated test of glucose (Glucolin®, the Boots Company, Nottingham, United Kingdom).
were analysed using repeated measures analysis of
variance (ANOVA) with the least significant difference
(LSD) posthoc test. Pearson correlation determined
the relationship between estimated formulae (GIadj and
GIpred) and iAUC120 with the r = 0.50 as strong, r =
0.30 as moderate and r = 0.10 as weak relationship
(Cohen, 1977). The Bland-Altman analysis was used
to detect an agreement between GImeasured and the two
formulae (GIpred and GIadj). The critical value for this
procedure is termed limit of agreement, defined as the
mean ± 2SD of the differences between the two
formulae. The limit of agreement represents the range
within 95% of the differences lie.
*
*
*
*
*
*
*
RESULTS
A total of 11 healthy participants (mean age of 23.0 ±
2.2 years) completed the study. The mean CV for each
of the participant is shown in Table 2. The mean CV
for this study was 22.23%. One participant had a
CV > 30%, and therefore, one outlying result of that
particular participant was removed providing a final
CV of 20.23% (Table 2). No participants had a GI
value exceeding mean ± 2SD.
All participants had a normal BMI (21.0 ± 2.4
kg/m2) and fasting blood glucose (4.8 ± 0.3 mmol/L).
There was no day-to-day variation within the subjects
based on no differences in energy intake, BMI and
physical activity level throughout the study (P > 0.05).
Baseline fasting blood glucose before consuming the
test meals was comparable (Fig. 1). Blood glucose
levels increased significantly after consuming all the
test meals at 15 and 30 min (P < 0.05) (Fig. 1). Blood
glucose levels peaked at 30 min and the reference
food had the highest peak as compared to the rice
*
*
*
*
*
*
*
*
*
*
*
*
Fig. 1. Glycemic response of rice alone and mixed meals for red
rice, fragrant rice and parboiled rice (Mean ± SD, n = 11).
* indicates significant difference between rice and glucose (P <
0.05).
278
Rice Science, Vol. 24, No. 5, 2017
Fig. 2. iAUC120 comparison between rice alone and mixed meals.
Values are mean ± SD (n = 11).
Different letters above the bars mean significant difference at the
0.05 level.
alone and the rice-based mixed meals (P < 0.05). Only
the fragrant white rice peaked at 45 min. Blood
glucose levels after consuming fried red rice and fried
parboiled rice were significantly lower than the red
rice and the parboiled rice at 60 min (P < 0.05),
respectively. At 120 min, the blood glucose level
dropped below the fasting values for the reference
food and the fragrant white rice.
The iAUC120 reflects changes in blood glucose
levels over the 2 h after consuming different test
meals (Fig. 2). The iAUC120 of the test meals ranged
from 72.2 to 189.7 mmol·min/L with the reference
food having the highest iAUC120 (P < 0.05). Within
the rice-based mixed meals, fried fragrant white rice
had a higher iAUC120 compared to the fried parboiled
rice and fried red rice (P < 0.05). Adding fat and
protein to the rice-based mixed meals reduced the
glycemic responses by 25% to 39% (Table 3). In
general, rice-based mixed meals showed a significantly
lower incremental peak of glucose when compared to
the rice alone (P < 0.05) (Table 4).
The mean values of GImeasured (Mean ± SE) in
descending order are red rice (68 ± 8), fragrant white
rice (67 ± 7) and parboiled rice (61 ± 8). These values
were significantly lower than the reference food (GI =
100) (P < 0.05) (Table 4).
Fried red rice showed the highest estimated meal GI
(GIadj = 57 and GIpred = 62) followed by fried fragrant
white rice (GIadj = 55 and GIpred = 61) and fried
parboiled rice (GIadj = 51 and GIpred = 56). GIadj (r =
0.46, P < 0.01) and GIpred (r = 0.41, P < 0.01) were
moderately correlated with the iAUC120 (P < 0.01). The
Bland-Altman analysis showed a fixed bias between
GImeasured and GIpred which was not seen between
GImeasured and GIadj. The use of GIadj showed a good
level of agreement as most of the data evaluated were
within the limit of agreement (Fig. 3). The limits of
agreement for GImeasured and GIadj were between 37.62
(upper limit) and -46.65 (lower limit) with the mean
being -4.517 (Fig. 3).
Table 3. Adjustment of calculated meal glycemic index (GI) for differences in fat, protein and carbohydrate content in mixed meals.
Test material
Protein (g)
Fat (g)
GIpred a
Proteinadj b
Fatadj c
GIadj d
Red rice
3.60
0.20
–
1.00
1.00
Fried red rice
6.90
13.20
62
0.95
0.96
57
Fragrant white rice
2.16
0.15
–
1.00
1.00
Fried fragrant white rice
6.10
13.20
61
0.94
0.96
55
Parboiled rice
2.86
0.22
–
1.00
1.00
Fried parboiled
6.70
13.20
56
0.94
0.96
51
a
Predicted meal GI as described by Dodd et al (2011). b Proteinadj = 1 – 1.45 × (F – R) /100, where F is fried rice (mixed meal) protein and R is
rice (alone) protein. The factor 1.45 is the mean percentage of reduction in AUC/g protein from Moghaddam et al (2006) as described by Wolever
(2013). c Fatadj = 1 – 0.29 × (E – P) /100, where E is fried rice (mixed meal) fat and P is rice (alone) fat. The factor 0.29 is the mean percentage of
reduction in AUC/g fat from Moghaddam et al (2006) and Lan-Pidnainy and Wolever (2010) as described by Wolever (2013). d GIadj = GIpred ×
Proteinadj × Fatadj.
Table 4. Fasting blood glucose values, peak blood glucose values, estimated glycemic index (GI) and measured GI (Mean ± SE, n = 11).
Test material
Fasting blood glucose (mmol/L)
Peak blood glucose value (mmol/L)
GIpred a
Glucose
5.00 ± 0.21
8.86 ± 0.14
100
Red rice
4.88 ± 0.13
7.08 ± 0.32
–
Fried red rice
4.83 ± 0.09
6.36 ± 0.25
62
Fragrant white rice
5.05 ± 0.09
6.95 ± 0.19
–
Fried fragrant white rice
5.22 ± 0.19
6.86 ± 0.20
61
Parboiled rice
4.96 ± 0.12
6.87 ± 0.23
–
Fried parboiled rice
4.91 ± 0.10
6.68 ± 0.25
56
a
GIpred and GIadj values were calculated via equations from Dodd et al (2011) and Wolever (2013), respectively.
GIadj a
GImeasured
100
–
57
–
55
–
51
100 ± 0
68 ± 8
41 ± 4
67 ± 7
50 ± 7
61 ± 8
41 ± 4
Nur Maziah Hanum OSMAN, et al. Estimating Glycemic Index of Rice-Based Mixed Meals
Fig. 3. Limit of agreement between GImeasured and GIadj.
Difference in GImeasured and GIadj was calculated by substracting
GImeasured from GIadj. Mean of GImeasured and GIadj was calculated by
averaging GImeasured and GIadj. The dotted line represents the limit of
agreement between GImeasured and GIadj. Middle line represents a mean
difference between GImeasured and GIadj.
DISCUSSION
This study aimed to evaluate the glycemic response of
rice-based meals either in the form of rice alone or
mixed meals and to identify the appropriate formula to
estimate the rice-based mixed meal GI. We observed
that co-ingestion of rice with fat and protein in mixed
meals reduced the glycemic response compared to rice
alone. This observation was also in line with the study
by Sun et al (2014), which indicated that the additions
of proteins, fats, and vegetables to white rice produce
a lower glycemic and insulinemic responses than rice
alone (Sun et al, 2014), and the extent of reduction in
glycemic response after adding fat and protein is
nearly 50%. We observed less reduction in glycemic
response (25%–39%) of the mixed meals. This could
be explained by the amount of carbohydrates (50 g
available carbohydrate) in their study, which was
twice of that in this study (25 g available carbohydrate).
Fat and protein reduce glycemic response by
delaying gastric emptying and stimulating insulin
secretions (Ryan et al, 2013). The independent effect
of fat and protein on glycemic response cannot be
entirely ruled out. This is because we used the same
type and amount of fat (13 g) and protein (6 g) to the
meals, which controlled their effects on glycemic
response. A number of carbohydrates were also
consistent for all the tested meals. Despite having the
same amount of carbohydrates in all the rice-based
mixed meals, various types of rice produced different
279
glycemic responses. This implicates that modifying
the types of carbohydrates is equally important in
reducing glycemic response especially for people with
diabetes (Ajala et al, 2013). This finding challenges
the current thought of diabetes nutrition recommendations,
which only emphasise on the amount of carbohydrate
rather than the type as the mainstay of dietary therapy.
White rice consumption has been associated with
increased diabetes risk. In a meta-analysis of
prospective cohort studies, an increase in a daily
serving of white rice alone has been associated with
11% increase in diabetes risk (Hu et al, 2012). It is
interesting to note that replacing 50 g/day of white
rice to the same quantity of brown rice intake has been
associated with a 16% lower risk of diabetes. Hence,
improving the carbohydrate quality may provide
feasible initiative for diabetes prevention (Sun et al,
2010). Although the acute glucose-lowering effect of
dietary fat and protein to rice-based mixed meals has
been consistently observed (Robert and Ismail, 2012;
Sun et al, 2014), the degree of reduction to influence
the diabetes risk reduction warrants further investigation.
Estimation of meal GI in mixed meal using
predicted or adjusted formula provided contradictory
findings (Dodd et al, 2011; Hatonen et al, 2011;
Robert and Ismail, 2012; Sun et al, 2014). In our study,
we used predicted formula to estimate the meal GI
(GIpred). We also estimated the meal GI (GIadj) by
making a necessary adjustment for fat and protein.
The fat and protein contents were considered in the
adjusted formula. Therefore, the meal GI can be
estimated reliably (Wolever, 2013). In this study, both
formulae were closely correlated to the glycemic
response they elicited. However, using the BlandAltman analysis, the agreement between GIadj and
GImeasured is more apparent than GIpred in estimating the
meal GI of rice-based mixed meals. The use of
Pearson correlation to determine the appropriate
formula may be misleading. This is because Pearson
correlation only estimates a linear correlation and it
does not provide information on the agreement
between measured and estimated GI values (Giavarina,
2015).
Flint et al (2004) was unable to detect the ability of
GIpred to estimate the GI of the European breakfast
meals. However, another study using the GIpred
formula can reliably estimate the GI of breakfast
mixed meals when performed using a proper design
without controlling a number of carbohydrates (Wolever
et al, 2006). The non-significant findings in the earlier
280
study could be related to the misclassification of the
foods’ GI and the fixed amount of carbohydrates
(Flint et al, 2004). In this study, incorrect GI
classification was unlikely because we determined the
GI of individual rice using a standardised procedure.
However, we fixed the amount of carbohydrates
content at 25 g of each meal. This approach may
reduce the strength of the association between
carbohydrate content and glycemic responses to the
meal. Besides, we controlled the cooking procedures
for all the meals, which may also hinder a consistent
result. The amount of water and time of cooking
influences GI of the rice. A longer cooking time and
enough water allow higher starch gelatinization,
which influences digestibility and glycemic responses
(Kaur et al, 2016).
In this study, the use of GIpred for the rice-based
mixed meals may not appropriate as it overestimated
the measured GI between 11 and 21 units.
Nevertheless, the GIpred formula has been used in most
of the previous studies, where rice is the staple diets
(Nisak et al, 2010; Loh et al, 2013; Shyam et al, 2013).
Hence, it may overestimate the dietary GI that may
affect the study outcomes. On the other hand, it is
worth noting that dietary intakes were not in the
controlled condition in these studies (Nisak et al, 2010;
Loh et al, 2013; Shyam et al, 2013). The ad libitum
conditions may increase the likelihood to detect the
strong correlation between estimated GI and glycemic
response (Wolever, 2013). Therefore, further investigation
is warranted to confirm the applicability of GIadj in
applied research.
The GI values of three types of rice were compared
with previously published data (Table 5). GI value of
parboiled rice is comparable with the previous studies
(Foster-Powell et al, 2002; Yusof et al, 2005;
Atkinson et al, 2008; Sydney University Glycemic
Index Research Service, 2014). However, the GI
values of red rice and fragrant white rice are not
comparable to the other studies (Foster-Powell et al,
2002; Yusof et al, 2005; Atkinson et al, 2008; Sydney
University Glycemic Index Research Service, 2014).
All the rice materials tested are classified into
moderate GI rice. This cutoff uses < 55 as low, 55–70
as medium and > 70 as high GI (Australian Standard,
2007; International Standard Organisation, 2010).
This variation in GI values for fragrant white rice and
red rice could be influenced by many factors including
volume of water added to the rice which determines
the starch gelatinization (Foster-Powell et al, 2002). In
Rice Science, Vol. 24, No. 5, 2017
Table 5. Glycemic index values of three different types of rice from
various studies and in this study.
Other studies a
This study
Test rice
Number
Range
Mean ± SE
Red rice
2
76–99
68 ± 8
Fragrant white rice
4
79–84
67 ± 7
Parboiled rice
19
38–87
61 ± 8
a
Data obtained from Foster-Powell et al (2002), Sydney University
Glycemic Index Research Service (2014), Yusof et al (2005) and
Atkinson et al (2008).
this study, more water is added to the red rice (1:2)
than to the parboiled rice (1:1) and fragrant white rice
(1:1). Increased starch gelatinization has been shown
to directly influence starch digestibility which in turn
increases the glycemic responses (Kaur et al, 2016).
Gelatinization enhances the susceptibility of starch to
digestive enzymes, as the swelling of starch granules
enhances the accessibility of enzymes to penetrate into
the granules (Syahariza et al, 2013). Fragrant white
rice has a low amylose content (Miller et al, 1992).
Amylose consists of a linear molecule while
amylopectin consists of the high-branched molecule.
Low amylose ratio produces higher GI value because
of the higher rate of digestion. There is less hydrogen
bonding between glucose units in amylose molecule
than in amylopectin, thus amylose molecule showing
high exposure to enzymatic digestion (Behall and
Scholfield, 2005).
The study has several limitations. As demonstrated
in this study, a total of 25 g available carbohydrates
were used to estimate GIadj which may be
inappropriate as the original articles. The use of
adjusted formula may cause inaccuracy when applied
to a study using 25 g available carbohydrate. This
study was primarily conducted to determine the
appropriate formula in estimating meal GI of the
rice-based mixed meal. It is evident from the study
that the adjusted or predicted formula can moderately
estimate the meal GI in an experimental condition.
Therefore, both of formulae must be further tested on
various health outcomes and in a long-term trial. We
were unable to separate the independent effects of fat
and protein to rice-based mixed meals from this study,
and therefore, further work is required to understand
their interaction following consumption of rice-based
mixed meals. As this study did not measure the insulin
and gut hormones parameters, another study may be
needed to systematically rule out the mechanism of
action of fat and protein on rice-based mixed meals.
The reliability of adjusted formula to other staple
carbohydrates warrants further investigation. In
Nur Maziah Hanum OSMAN, et al. Estimating Glycemic Index of Rice-Based Mixed Meals
conclusion, the GI of rice alone can be used to
estimate the GI of the rice-based mixed meal. The
Bland-Altman analysis showed that adjusted formula
(GIadj) was more appropriate than predicted formula
(GIpred) to estimate the GI of the rice-based mixed
meal. This study may support the clinical utility of GI
concept in the context of rice-based mixed meals.
ACKNOWLEDGEMENTS
We are grateful to all study participants for their
contributions and Nurul Ain SAIPUDIN and Mohd Faez
BACHOK for excellent technical support. This study
was funded by Ministry of Higher Education
(Fundamental Research Grant Scheme (FRGS
5524213) and Universiti Putra Malaysia.
REFERENCES
Abdul Karim N, Mohd Yusof S, Hashim J K, Din M, Haslinda S,
Harun Z, Saleh Hudin R, Salim F, Ngadikin M, Norazlin S. 2008.
Food consumption patterns: Findings from the Malaysian adult
nutrition survey (MANS). Mal J Nutr, 14(1): 25–39.
Ajala O, English P, Pinkney J. 2013. Systematic review and
meta-analysis of different dietary approaches to the management
of Type 2 diabetes. Am J Clin Nutr, 97(3): 505–516.
Asp N G. 2001. Enzymatic gravimetric methods. In: Spiller G A.
CRC Handbook of Dietary Fiber in Human Nutrition. 3rd edn.
Boca Raton, London, New York, Washington D C: CRC Press.
Association of Official Analytical Chemists (AOAC). 1998.
Official Methods of Analysis of the AOAC International. 16th
edn. Gaithersburg, MD, USA: AOAC International.
Atkinson F S, Foster-Powell K, Brand-Miller J C. 2008.
International tables of glycemic index and glycemic load values:
2008. Diab Care, 31(12): 2281–2283.
Australian Standard. 2007. Glycemic Index of Food. Sydney:
Standards Australia.
Behall K M, Scholfield D J. 2005. Food amylose content affects
postprandial glucose and insulin responses. Cereal Chem, 82(6):
654–659.
Bhupathiraju S N, Tobias D K, Malik V S, Pan A, Hruby A,
Manson J E, Willett W C, Hu F B. 2014. Glycemic index,
glycemic load, and risk of Type 2 diabetes: Results from 3 large
US cohorts and an updated meta-analysis. Am J Clin Nutr, 100:
218–232.
Chen L, Pei J H, Kuang J, Chen H M, Chen Z, Li Z W, Yang H Z.
2015. Effect of lifestyle intervention in patients with Type 2
diabetes: A meta-analysis. Metabolism, 64(2): 338–347.
Cohen J. 1977. Statistical Power Analysis for the Behavioral
Sciences. New York, USA: Lawrence Erlbaum Associates.
Dodd H, Williams S, Brown R, Venn B. 2011. Calculating meal
glycemic index by using measured and published food values
281
compared with directly measured meal glycemic index. Am J
Clin Nutr, 94(4): 992–996.
Flint A, Møller B K, Raben A, Pedersen D, Tetens I, Holst J J,
Astrup A. 2004. The use of glycaemic index tables to predict
glycaemic index of composite breakfast meals. Brit J Nutr, 91(6):
979–989.
Foster-Powell K, Holt S H, Brand-Miller J C. 2002. International
table of glycemic index and glycemic load values: 2002. Am J
Clin Nutr, 76(1): 5–56.
Giavarina D. 2015. Understanding bland altman analysis. Biochem
Med, 25(2): 141–151.
Hatonen K A, Virtamo J, Eriksson J G, Sinkko H K, Sundvall J E,
Valsta L M. 2011. Protein and fat modify the glycaemic and
insulinemic responses to a mashed potato-based meal. Brit J
Nutr, 106(2): 248–253.
Hu E A, Pan A, Malik V, Sun Q. 2012. White rice consumption
and risk of Type 2 diabetes: Meta-analysis and systematic
review. Brit Med J, 344: e1454.
International Organization Standard. 2010. ISO26642–2010: Food
Products: Determination of the Glycaemic Index (GI) and
Recommendation for Food Classification. Switzerland:
International Organization for Standardization.
Kaur B, Ranawana V, Henry J. 2016. The glycaemic index of rice
and rice products: A review, and table of GI values. Crit Rev
Food Sci Nutr, 56(2): 215–236.
Loh B I, Sathyasurya D R, Jan Mohamed H J. 2013. Plasma
adiponectin concentrations are associated with dietary glycemic
index in Malaysian patients with Type 2 diabetes. Asia Pac J
Clin Nutr, 22(2): 241–248.
Miller J B, Pang E, Bramall L. 1992. Rice: A high or low glycemic
index food? Am J Clin Nutr, 56(6): 1034–1036.
Nisak M Y B, Abd Talib R, Norimah A K, Gilbertson H, Azmi K
N. 2010. Improvement of dietary quality with the aid of a low
glycemic index diet in Asian patients with Type 2 diabetes
mellitus. J Am Coll Nutr, 29(3): 161–170.
Prosky L, Asp N G, Furda I, de Vries J W, Schweizer T F, Harland
B F. 1985. Determination of total dietary fiber in foods and food
products: Collaborative study. J Assoc Anal Chem, 68(4):
677–679.
Robert S D, Ismail A A S. 2012. Glycemic responses of patients
with Type 2 diabetes to individual carbohydrate-rich foods and
mixed meals. Ann Nutr Metab, 60(1): 27–32.
Ryan A T, Luscombe-Marsh N D, Saies A A, Little T J, Standfield
S, Horowitz M, Feinle-Bisset C. 2013. Effects of intraduodenal
lipid and protein on gut motility and hormone release, glycemia,
appetite, and energy intake in lean men. Am J Clin Nutr, 98(2):
300–311.
Shyam S, Arshad F, Abdul Ghani R, Wahab N A, Safii N S,
Barakatun Nisak M Y, Chinna K, Kamaruddin N A. 2013. Low
glycaemic index diets improve glucose tolerance and body
weight in women with previous history of gestational diabetes:
A six months randomized trial. Nutr J, 12(1): 68.
Singh J, Dartois A, Kaur L. 2010. Starch digestibility in food
matrix: A review. Trends Food Sci Technol, 21(4): 168–180.
Sun L J, Ranawana D V, Leow M K S, Henry C J. 2014. Effect of
282
chicken, fat, and vegetable on glycaemia and insulinaemia to a
white rice-based meal in healthy adults. Eur J Nutr, 53(8):
1719–1726.
Sun Q, Spiegelman D, van Dam R M, Holmes M D, Malik V S,
Willett W C, Hu F B. 2010. White rice, brown rice, and risk of
Type 2 diabetes in US men and women. Arch Int Med, 170(11):
961–969.
Syahariza Z A, Sar S, Hasjim J, Tizzotti M J, Gilbert R G. 2013.
The importance of amylose and amylopectin fine structures for
starch digestibility in cooked rice grains. Food Chem, 136(2):
742–749.
Sydney University Glycemic Index Research Service. 2014.
Rice Science, Vol. 24, No. 5, 2017
Datafiles of Sydney University Glycemic Index Database.
Australia: The University of Sydney.
Wolever T M, Yang M, Zeng X Y, Atkinson F, Brand-Miller J C.
2006. Food glycemic index, as given in glycemic index tables, is
a significant determinant of glycemic responses elicited by
composite breakfast meals. Am J Clin Nutr, 83(6): 1306–1312.
Wolever T M S. 2013. Is glycaemic index (GI) a valid measure of
carbohydrate quality? Eur J Clin Nutr, 67(5): 522–531.
Yusof B N M, Talib R A, Karim N A. 2005. Glycaemic index of
eight types of commercial rice in Malaysia. Malay J Nutr, 11(2):
151–163.
(Managing Editor: FANG Hongmin)