Professional Documents
Culture Documents
ANNOTATED 4-Optimization of Natural Sweeteners For Sugar Reduction in Chocolate Flavoured
ANNOTATED 4-Optimization of Natural Sweeteners For Sugar Reduction in Chocolate Flavoured
ANNOTATED 4-Optimization of Natural Sweeteners For Sugar Reduction in Chocolate Flavoured
Optimization of natural sweeteners for sugar reduction in chocolate flavoured milk and
their impact on sensory attributes
Dipendra Kumar Mahato, Russell Keast, Djin Gie Liem, Catherine Georgina Russell,
Sara Cicerale, Shirani Gamlath
PII: S0958-6946(20)30292-2
DOI: https://doi.org/10.1016/j.idairyj.2020.104922
Reference: INDA 104922
Please cite this article as: Mahato, D.K., Keast, R., Liem, D.G., Russell, C.G., Cicerale, S., Gamlath, S.,
Optimization of natural sweeteners for sugar reduction in chocolate flavoured milk and their impact on
sensory attributes, International Dairy Journal, https://doi.org/10.1016/j.idairyj.2020.104922.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
12 ABSTRACT
13 Public Health bodies like the World Health Organization and Public Health England have
of
14 identified addition of sugar to processed foods and beverages as a health concern. Owing to the
ro
15 increasing consumption of sugar sweetened chocolate flavoured milk in Australia among all age
16
-p
groups, this study was conducted to investigate the possibility of achieving a 50% of added sugar
re
17 reduction in chocolate milk using natural non-nutritive sweeteners, namely stevia and monk fruit
lP
19 Methodology. Variables for 50% reduction of added sugar in chocolate milk were chosen as stevia
ur
20 sweetener (5-100 ppm) and monk fruit extract (50-100 ppm) based on preliminary trials. Sensory
21 attributes (overall liking, appearance, aroma, sweetness, mouthfeel and aftertaste) were taken as
Jo
22 the responses. The process was optimized with a combination of 56.27 ppm of stevia sweetener
23 and 81.90 ppm of monk fruit extract with sensory attributes of 6.78, 6.47, 6.31, 6.47, 6.45 and
24 6.33, respectively for overall liking, appearance, aroma, sweetness, mouthfeel and aftertaste. Use
25 of the two selected sweeteners in combination resulted in sweetness synergy. In addition, the
26 bitter and metallic aftertastes of stevia were masked by the monk fruit extract and helped enhance
27 the overall sensory attributes of the product compared to the control. This study provides
28 insightful information for optimizing sweeteners for sugar reduction in milk-based product.
29
30
1
31 1. Introduction
32 Consumption of added sugars in the diet is associated with an increased risk of several
33 chronic diseases, including overweight, obesity and type 2 diabetes (Malik, Schulze, & Hu, 2006;
34 Te Morenga, Mallard, & Mann, 2013; Patel et al., 2018; Singh et al., 2020). Globally,
35 consumption of added sugar has tripled in the last 50 years (Lustig, Schmidt, & Brindis, 2012).
36 Major contributors to consumption of added sugar are sugar-sweetened beverages which include
37 soft drinks (e.g. soda or pop), fruit and sports drinks, tea, coffee, energy drinks, sweetened milk or
38 milk alternatives, as well as other beverages to which sugar (mainly sucrose, fructose or high
of
39 fructose corn syrup) has been added (Blake, Lancsar, Peeters, & Backholer, 2019; Cleghorn, et al.,
ro
40 2019; Falbe, Thompson, Patel, & Madsen, 2019).
41
-p
Milk is the principal dairy product consumed by children worldwide. Approximately, 60-
re
42 80% of American children’s dairy product consumption is comprised of fluid milk (FPED, 2014).
lP
43 The majority of this milk is flavoured milk (86% of all the milk) in the USA, and further, the
44 majority of flavoured milk available is chocolate (Nicklas, O'neil, & Fulgoni, 2017). In Australia,
na
45 per capita consumption of drinking milk decreased from 105.1 litres in 2014-15 to 98.6 litres in
ur
46 2018-19 due to growth of flavoured milk products and coffee (Dairy Australia, 2019), suggesting
Jo
47 that Australians are increasingly incorporating flavoured milk in their diets. Flavoured milk is
48 considered a core food in the Australian Dietary Guidelines (Brownie, Muggleston, & Oliver,
49 2015) and is used to promote milk intake to ensure that children meet the recommended dietary
50 allowances (RDA) for vitamin D and calcium (Patel et al., 2018). Children’s dairy consumption is
51 typically lower than recommended in the Australian Dietary Guidelines (ADGs) and flavoured
52 milk might help to address this because children tend to consume more flavoured milk than plain
54
2
55 However, regular consumption of sweetened flavoured milk has been reported to increase
56 energy intake more than 10% compared to non-consumers (Krachler et al., 2006; Ludwig,
57 Peterson, & Gortmaker, 2001; Striegel-Moore et al., 2006) with the increased energy intake linked
58 to overweight, obesity and type 2 diabetes (Malik et al., 2006; Te Morenga et al., 2013; Patel et
59 al., 2018; Singh et al., 2020). The World Health Organization (WHO) recommends less than 10%
60 of total energy intake should come from free sugar and is in line with Public Health England
61 (PHE) recommendation of least a 20% sugar reduction by 2020 in processed foods and beverages
62 (PHE, 2017). Therefore, sugar reduction in chocolate milk is desired to maximise the health
63 benefits of chocolate milk consumption, specifically calcium. The vectors rule of a well-tested
of
64 model of an epidemiological triad (hosts, vectors, and environments) suggests ‘small changes ×
ro
65 large volumes = significant population benefits’ (Egger, Swinburn, & Rossner, 2003; Keast,
66
-p
Sayompark, Sacks, Swinburn, & Riddell, 2011). Even a small reduction can significantly benefit a
re
67 larger population in long terms (Mahato et al., 2020).
lP
68
na
69 Sweet taste is a driver of liking for many foods such as chocolates (de Melo, Bolini, &
ur
70 Efraim, 2009), yogurt and yogurt-like products (Bayarri, Carbonell, Barrios, & Costell, 2011;
71 Thompson, Lopetcharat, & Drake, 2007) along with chocolate milk (Li, Lopetcharat, & Drake,
Jo
72 2015). Sugar reduction greater than 30% affects the liking of dairy products including flavoured
73 yogurt (Chollet, Gille, Schmid, Walther, & Piccinali, 2013) and chocolate milk (Li et al., 2015).
74 Thus, finding a suitable sugar substitute to increase sugar reduction while maintaining sweet taste
75 and liking is important. Reduction of sugar without consumers’ noticing could be the most
76 effective strategy to maintain acceptance, however, this is difficult to achieve as sugar has
77 multiple sensory influences on a product, not just sweetness. This is the challenge for the food
78 industry.
79
80
3
81 Reduction of sugar and its impact on consumer acceptance can be overcome by the use of
82 high intensity sweeteners. Most previous studies have used artificial non-nutritive sweeteners
83 (NNS) like advantame, neotame, sucralose, or aspartame in dairy-based foods and beverages to
84 replace sucrose (Morais, Morais, Cruz, & Bolini, 2014). However, consumer trends suggest a
85 desire for healthier products with “all-natural” ingredients and therefore, products with natural
86 sweeteners are preferred over artificial sweeteners (Christoph et al., 2011; Li, Lopetcharat, &
87 Drake, 2014). For example, Li et al. (2014) found parents preferred to buy chocolate flavoured
88 milk with added natural NNSs or sucrose over the artificial NNSs for their children. Further, Li et
89 al. (2015) reported parents’ and children's acceptance of naturally sweetened chocolate milk.
of
90 Similarly, Oltman, Lopetcharat, Bastian, & Drake (2015) found general consumers’ (18-70 years)
ro
91 preference of “naturally sweetened” labels for protein beverages. Natural NNSs are derived from
92
-p
plants and comprise of natural compounds (Kim & Kinghorn, 2002), therefore, use of natural
re
93 NNSs like stevia and monk fruit could have better consumer acceptance. Also, an enzymatic
lP
94 treatment (lactase) can increase the sweetness and reduce sugar (up to 50%) which could be an
na
95 alternative to reduce sugar in a natural way in flavoured milk apart from the use of natural
ur
96 sweeteners (arlafoods.co.uk).
97 Stevia (Stevia rebaudiana) belongs to the Asteraceae family. The sweetness of stevia is due
Jo
98 to sweet glycosides, stevioside and different types of rebaudiosides (e.g. Reb A, Reb B, Reb D,
99 Reb E, Reb M, etc.) present in stevia leaves (Parker, Lopetcharat, & Drake, 2018). Stevia is
100 commonly used in dairy products such as yogurt and ice cream, baked goods and soft drinks
101 (Guggisberg, Piccinali, & Schreier, 2011; Ozdemir, Arslaner, Ozdemir, & Allahyari, 2015; Pon,
102 Lee, & Chong, 2015). Steviol glycosides are noncarcinogenic but consumption more than 4 mg
103 kg−1 body weight day−1 is unsafe and may change the composition of the gut microbiota (EU
104 regulation 1129/2011) (EFSA, 2010, 2015; Ruiz-Ojeda, Plaza-Díaz, Sáez-Lara, & Gil, 2019).
105 Further, rebaudiosides like Reb A and Reb M can be used in food and beverages for human
4
106 consumption in Australia and New Zealand at the same permitted levels as steviol glycosides
108 Monk fruit (Siraitia grosvenorii) also known as Luo Han Guo (DuBois & Prakash, 2012;
109 Pawar, Krynitsky, & Rader, 2013), is a fruit native to southern China, that contains sweet
110 glycosides mogroside IV, mogroside V, and mogroside VI (Pawar et al., 2013). Monk fruit has
111 several health beneficial effects such as anti-cancer (Liu et al., 2016), antioxidant (Chen, Wang,
112 Qi, & Xie, 2007), anti-inflammatory (Chun et al., 2014), anti-obesity (Sun et al., 2012) and anti-
113 diabetic properties (Lee, Jeong, Kim, & Nam, 2016; Qi, Chen, Zhang, & Xie, 2008). In a recent
114 study by Ban et al. (2020), rats fed symbiotic yoghurt sweetened with monk fruit extract showed
of
115 greater blood regulation and a significant reduction in insulin resistance and glycosylated
ro
116 haemoglobin compared to rats fed yoghurt sweetened with sucrose. Further, monk fruit is a low
117
-p
cost and high-intensity natural sweetener that can be utilized in food and beverage industries for
re
118 the development of low calories products for diabetics and health-conscious consumers (Pandey
lP
119 & Chauhan, 2019). The ADI has not been established for monk fruit extract because no adverse
na
120 effects have been reported, however, the ADI of monk fruit juice concentrate is 25 mg kg−1 body
121
122 To achieve sugar reduction in chocolate milk, some studies have tried reformulation. For
Jo
123 example, Rad, Delshadian, Arefhosseini, Alipour, & Jafarabadi (2012) used stevia as a sugar
124 replacer along with inulin as a thickening agent to evaluate the physical properties (sedimentation
125 and viscosity) of chocolate milk. Li et al., (2015) used monk fruit extract and stevia leaf extract
126 separately to reduce sugar in skim chocolate milk and studied the impact on consumer acceptance.
127 Similarly, Bordi Jr, Palchak, Verruma-Bernardi, & Cho (2016) used Reb A stevia to partially
128 substitute for sugar in chocolate milk. Further, Azami, Niakousari, Hashemi, & Torri (2018) used
129 liquorice extract as a sugar substitute in chocolate milk and studied consumer acceptance and
130 physicochemical properties. All these studies used a single sweetener for sugar reduction.
131 However, as no single sweetener has similar functionality to sucrose, the use of two or more
5
132 sweeteners as a blend may provide flavour and taste profiles similar to sucrose. For example, the
133 combined effect of stevia and sucralose improved sensory and physical properties of a sugar-free
134 dairy dessert (Furlán & Campderrós, 2017). Similarly, blending Reb M with Reb B/ Reb D
135 resulted in sweetness synergy (Prakash, Markosyan, & Bunders, 2014). No such experiment
136 involving two sweeteners and the effect of their interactions on the sensory attributes has been
137 reported in chocolate milk for sugar reduction. Therefore, this work aims to optimize two natural
138 NNSs (stevia sweetener containing steviol glycosides, Reb A and stevioside, and monk fruit
139 extract containing mogroside V) for sugar reduction in chocolate flavoured milk for optimum
of
ro
141 The novelty of this paper is its focuses on two sweeteners (stevia and monk fruit extract) of
142 -p
plant origin and introduces a Response Surface Methodology (RSM)-based approach to optimize
re
143 these sweeteners for sugar reduction in a milk-based product with optimum sensory attributes and
lP
144 consumer liking. The article would be a milestone for the dairy industry to reduce the sugar
145 content of their products in order to meet the demands and pressure from Public Health bodies and
na
146 consumers. Furthermore, as this article provides an insightful approach for optimizing sugar
ur
147 reduction in milk-based drinks, it would also be considered for its optimization methodology.
Jo
150 Coles cocoa powder (natural, 100% pure cocoa; 23.3% w/v, protein; 11.0% w/v, fat; and
151 18.7% w/v, carbohydrate), CSR white sugar (refined) and Australian homogenised and
152 pasteurised milk (3.5% w/v, protein; 3.3% w/v, fat; and 4.6% w/v, carbohydrate) were purchased
153 from a supermarket near Deakin University, Burwood, Australia. Starch (NOVATIONTM 3300)
154 was supplied by Ingredion, Westchester, city, IL, USA while the stabilizer (Ticaloid® 750
155 containing carrageenan standardized with maltodextrin) was provided by TIC GUMS,
156 Philadelphia, Ph, USA. Natural non-nutritive sweeteners namely Stevia (TASTEVA® containing
6
157 steviol glycosides, Reb A and stevioside) and Monk fruit extract (PUREFRUITTM containing
158 mogroside V) were made available by TATE & LYLE, London, UK.
160 The chocolate flavoured milk was prepared as per the process described by Ingredion
161 Singapore Pte Ltd, depicted in Fig. 1. The concentrations of the stabilizer (Ticaloid®750) and the
162 starch (NOVATIONTM 3300) were fixed to 0.02% and 0.5%, respectively based on pre-trials
163 performed to match a commercial reference sample. The stabilizer provides thermostability while
164 the starch improves the rheology or mouthfeel properties of the product (Agoda-Tandjawa, Le
of
165 Garnec, Boulenguer, Gilles, & Langendorff, 2017). The total sugar concentration of commercial
ro
166 chocolate milk in the Australian supermarket varies from 4-12% (including both reduced and full-
167
-p
sugar versions). Our previous consumer study on preference mapping (unpublished data) of ten
re
168 commercial chocolate milk samples with this sugar concentration range using 235 participants
lP
169 (19-63 years, 82.98% female, 65.53% Australian, 24.26% Asian and remaining from other
170 nations), identified OAK CHOCOLATE milk and OAK THICK DEATH BY CHOCOLATE milk
na
171 (Parmalat Food Products PTY LTD, Australia) with sugar concentrations 10.6% and 10.9%,
ur
172 respectively, to have the highest appeal satisfying between 70-80% of consumers assessed.
Jo
173 Therefore, the lower limit (round figure, 10%) was considered as the control point for sugar
174 reduction with OAK chocolate milk as a commercial reference sample in this study.
175 To formulate the control sample, 5% sucrose (white sugar) was added to obtain the final
176 10% total sugar. The remaining 5% sugar was contributed by other ingredients (milk and cocoa
177 powder) in the chocolate milk as calculated by FoodWorks (Xyris Software, Australia Pty Ltd).
178 The software calculates the final nutrient composition of the product based on the yield, nutrient
179 composition and proportion of individual ingredients added to the product. The control sample
180 was prepared to match the commercial reference (OAK CHOCOLATE milk as determined from
181 our previous study) in terms of the sensory perception (overall liking, 6.50; appearance liking,
7
182 6.37; aroma liking 6.34; sweetness liking, 6.17; mouthfeel liking, 6.30; and aftertaste liking, 3.11)
183 and other physicochemical properties (pH, 6.73; total soluble solids, 20.5 °Brix; viscosity, 7.50
184 mPa.s; and water activity, 0.94) as benchmarking trials by 6-10 participants in CASS sensory
185 laboratory. To reduce sugar in chocolate flavoured milk, Stevia sweetener (TASTEVA®) and
188 In the initial trials to decide the concentrations of sweeteners, a single sweetener (stevia
189 sweetener or monk fruit extract) was compared to formulations with combinations of these
of
190 sweeteners. Different concentrations (low, middle and high) of stevia sweetener and monk fruit
ro
191 extract between 5-200 ppm (within the ADI limits for safe consumption) were used for
192
-p
formulations with 20, 50, 80 and 100% reduction of added sugar as preliminary trials, taking
re
193 chocolate milk with 10% total sugar as the control sample. It was found that an individual
lP
194 sweetener needed a higher concentration to achieve a sweetness level compared to the two in
195 combination. Also, other properties like mouthfeel and aftertaste were better when the two
na
196 sweeteners were used together which would account for their synergistic effects. Monk fruit
ur
197 extract showed a better sensory profile than stevia. Therefore, for further refining, the
Jo
198 concentration of stevia was reduced (5-100 ppm) while monk fruit extract (50-100 ppm) was
199 increased. The preliminary trials were repeated using the low, middle and high values of these
200 sweeteners in combination for 20, 50, 80 and 100% reduction of added sugar.
201 To determine maximum added sugar reduction that could be achieved without much
202 impact on sensory perception compared to the control chocolate milk (10% total sugar), 6-10
203 participants (25-60 years, 50% female, experts in sensory science and product development) were
204 used to evaluate the reformulations. One participant at a time was provided with samples from
205 low to high concentration and feedback was noted for sensory perception (sweetness, appearance,
8
206 colour, flavour, mouthfeel and aftertaste). Participants were allowed to rinse their mouth between
208
209 Based on feedback, 50% of added sugar reduction was deemed possibly with lower and
210 upper limits of Stevia sweetener (TASTEVA®, 5 and 100 ppm, respectively) and Monk fruit
211 extract (PUREFRUITTM, 50 and 100 ppm, respectively) and considered for the optimization by
212 RSM experimental design software. Other ingredients like cocoa powder (approx. 20% solution),
213 white sugar (sucrose, 5%), starch (0.5%) and stabilizer (0.02%) were kept fixed for all the 13
of
214 reformulations obtained by RSM design as described in section 2.5. (Experimental design and
ro
215 data analysis). These reformulated samples were taken for sensory evaluation.
216
-p
re
217 2.4. Sensory evaluation
lP
219 Typically, 100 consumers are required for sensory evaluation using a hedonic scale, which
ur
220 is generally representative of the target market or current users (Kemp, Hollowood, & Hort,
Jo
221 2009). Therefore, 120 participants were recruited allowing a 20% dropout rate. To be included,
222 participants were required to be above 18 years, consume milk and flavoured milk at least once
223 per week, be without diabetes and/or allergies or sensitivities. All the interested candidates
224 provided informed consent before participation. The consumer study was performed as per the
226
228 Participants were recruited using two strategies. First, members of the CASS consumer
229 database (n=2000+) were contacted via email through Compusense Cloud software (Compusense
9
230 Inc., Guelph, Ontario, Canada). Members of the CASS consumer database are members of the
231 general community who were recruited via social media, flyers, and posters in the greater
232 Melbourne area. Second, recruitment was conducted in areas surrounding and including Deakin
233 University, Burwood campus utilising social media, flyers, posters and faculty emails. Interested
234 candidates were screened by asking them to complete a short online questionnaire through
235 Compusense Cloud software (Compusense Inc., Guelph, Ontario, Canada) which included
237
of
238 2.4.3. Consumer study
ro
239 Among the 120 participants recruited considering a 20% dropout rate, 107 participants (43
240
-p
male and 64 female) aged between 20 and 65 years old evaluated the chocolate milk samples. The
re
241 reformulated chocolate flavoured milk with varying concentrations of sweeteners (Table 1) were
lP
242 evaluated for sensory acceptance. All the reformulations were prepared fresh (as depicted in Fig.
na
243 1) on the day of evaluation and stored at 4°C to ensure the safety of the product.
244 Panelist first evaluated samples for overall liking using 9-Point Hedonic Scale ranging from
ur
245 “1-Dislike Extremely” to “9-Like Extremely”. Second, evaluation focused on the attributes
Jo
246 (namely appearance, aroma, sweetness, mouthfeel and aftertaste) using the same 9-Point Hedonic
247 Scale. Samples (15-20 ml) were presented in cups with three-digit codes in randomised order
248 (Williams' Latin Square design) to avoid carryover effects and bias in the data collection.
249 Evaluation was performed under white light, at room temperature and in sound controlled
250 partitioned sensory booths in the CASS sensory laboratory of Deakin University. All the data
251 collection was done using Compusense software (Compusense Inc., Guelph, Ontario, Canada).
252 The mean values of the liking scores were used for the optimization of sweeteners using RSM
254
10
255 2.5. Experimental design and data analysis
256 Experiments were designed using RSM of Design-Expert 12 statistical software (State Ease,
257 Inc. Minneapolis, Mn, USA). The central composite design (CCD) of RSM was used to optimize
258 concentrations of natural non-nutritive sweeteners for 50% reduction of added sugar. It is based
259 on 2-level factorial designs, having centre and axial points to fit quadratic models. In total, it has 5
260 levels which include +/- axial points, +/- factorial points and the centre point. The number of
261 experiments is calculated based on the formula in the equation 1 (Aslan, 2007):
262
of
263 No. of Experiments = 2k + 2k + centre points………….……(1)
ro
264 Where, k is the number of variables. In the study design with 2 variables (Table 1), the number of
266 Therefore, 13 different formulations (5 centre points and 8 non-centre points) of chocolate
267 milk comprising of two independent variables: Stevia sweetener (5-100 ppm) and Monk fruit
na
268 extract (50-100 ppm) were used to fit a quadratic response surface and approximately locate the
ur
269 optimum combination. The centre points help to anchor the design space i.e. to control the design
Jo
270 from external environmental influence, thereby, ensuring that the design space is stable. The
271 effect of these independent variables on the sensory attributes (overall liking, appearance liking,
272 aroma liking, sweetness liking, mouthfeel liking and aftertaste liking) was evaluated by a second-
274 = +∑ +∑ + …………………………….(2)
275 Where, Y is the response variable; β0 is the intercept; Xi and Xj are independent variables; βi, βij
276 and βii are regression coefficients for linear, interaction and quadratic terms; and k is number of
277 variables. 3D graphs were generated using regression coefficient and analysis of variance
278 (ANOVA) was conducted to find the significant (p < 0.05) effect of the model terms. For model
11
279 fitness, model F-value, lack of fitness and coefficient of determination (R2) were estimated for
281
284 The CCD of the RSM software was used to generate levels of variables namely stevia
285 sweetener (X1) and monk fruit extract (X2) with 13 sets of experiments as shown in Table 1.
286 Based on the suggested model, the significance of the model and lack-of-fit (non-significant) were
of
287 individually checked for all responses. The statistics and adequacy of the models for all responses
ro
288 are shown in Table 2. For the model adequacy, it was found that the values of adjusted R2 were
290 After the model fit and adequacy, the numerical and graphical optimizations were carried
291 out for the sugar reduction in chocolate flavoured milk using stevia and monk fruit sweeteners. As
na
292 a part of numerical optimization, the criteria for desired goals, weightage and importance of each
ur
293 variable and response were set as summarized in Table 3. The goal for each sweetener
Jo
294 concentration was kept in range while the output for response variables (i.e. sensory attributes)
295 was desired for maximum. Using the desirability function method, the software predicted the most
296 desirable solution with the desirability value of 0.952 (Table 4). Based on the same set of criteria
297 and outputs for numerical optimization, relevant and statistically significant responses were
298 depicted through the 3D graphs for each response (Fig. 2 a-f). The shaded area represents the
299 independent variable (stevia sweetener and monk fruit extract) conditions and their effects on the
300 response variables (shown by 3D graphs) for optimizing sugar reduction in chocolate flavoured
301 milk.
12
303 3.2.1. Effect on overall liking
304 Consumers are usually asked to provide the hedonic ratings for the overall liking first,
305 before a specific attribute (e.g., sweetness) because they evaluate the product using an integrative
306 frame of mind where all attributes are considered collectively. However, when they evaluate a
307 specific attribute, they switch into an analytical frame of mind and focus only on that specific
308 attribute (Lawless & Heymann, 2010). This prevents the influence of the memory of any
310 The effect of non-nutritive sweeteners (stevia sweetener and monk fruit extract) on the
of
311 overall liking for 50% reduction of added sugar is presented in Table 1. Both stevia sweetener and
ro
312 monk fruit extract positively affected the overall liking score, however, the effect of stevia
313
-p
sweetener was more prominent than monk fruit extract. This effect is illustrated by the quadratic
re
314 equation obtained by the response surface analysis (RSA) of the data as expressed in eq. (3).
lP
315 Overall liking = + 6.76 + 0.13 × X1 + 0.074 × X2 - 0.21 × X1X2 - 0.17 × X12 - 0.08 × X22 …..(3)
na
316 Where,
ur
318 The range of overall liking scores varied between 6.06 to 6.80 (Table 1 and Table 3) on a
319 9-point hedonic scale for 50% reduction of added sugar. Similar findings were reported by earlier
320 researchers, Li et al. (2015) reported overall liking scores of 6.5 for both monk fruit and stevia
321 leaf extract for 50% reduction as compared to control skim chocolate milk (5.14% sucrose with
323 The overall liking scores followed a quadratic model as depicted in Fig. 2a and Fig. 3.
324 Initially, with an increase of the concentration of stevia sweetener, the overall liking increases up
325 to 6.76 and then tails off while for the monk fruit extract, the tailing off is less prominent (Fig. 3).
326 This can also be seen from the eq. (3) where the quadratic coefficient for stevia sweetener (X12) is
13
327 greater than that of monk fruit extract (X22). Further looking at the interaction between these two
328 sweeteners (X1X2 term in eq. 3), there is a slightly negative impact on the overall liking score.
329 Overall, both sweeteners influenced positively and synergistically increased the overall liking
331
333 The appearance liking is based on the visual appearance of the chocolate milk. The scores
334 for appearance liking ranged in between 6.07 and 6.50 (Table 1 and Table 3). The effect of stevia
of
335 sweetener and monk fruit extract on the appearance is describes as in quadratic eq. (4).
ro
336 Appearance = + 6.46 + 0.06 × X1 + 0.03 × X2 - 0.09 × X1X2 - 0.13 × X12 - 0.09 × X22 …..(4)
337 Where,
-p
re
338 X1 = Stevia sweetener (ppm); X2 = Monk fruit extract (ppm)
lP
339 Both the sweeteners (stevia sweetener and monk fruit extract) had a positive effect on
na
340 appearance liking of the chocolate milk (Fig. 2b and Fig. 3). Stevia sweetener showed a positive
ur
341 effect double compared to monk fruit extract as evident from their respective coefficient X1 and
Jo
342 X2 (eq. 4). Stevia sweetener and monk fruit extract showed a similar pattern of quadratic
343 behaviour as observed for overall liking (Fig. 3). Similarly, the quadratic coefficient for stevia
344 sweetener (X12) is greater than that of monk fruit extract (X22) as shown in eq. (4) and both the
345 sweeteners contributed positively and synergistically to appearance liking with a prediction of
346 6.47 (Fig. 4 and Table 4). The appearance liking of the current study was similar to Li et al.
347 (2015), where the appearance liking were 6.6 and 6.7, respectively for 50% reduction with monk
348 fruit extract and stevia leaf extract as compared to control skim chocolate milk (5.14% sucrose
350
14
351 3.2.3. Effect on aroma liking
352 Aroma is the odour perceived through the nasal cavity while flavour combines both aroma
353 and taste. The quadratic equation describing the relation between stevia sweetener and monk fruit
354 extract and their effect on the aroma is represented in eq. (5).
355
356 Aroma = +6.30 + 0.06 × X1 + 0.05 × X2 - 0.15 × X1X2 - 0.16 × X12 - 0.09 × X22 ……..(5)
357 Where,
of
358 X1 = Stevia sweetener (ppm); X2 = Monk fruit extract (ppm)
ro
359 The range for aroma liking scores varied between 5.80 to 6.34 (Table 1 and Table 3). Both
360
-p
sweeteners had a positive effect on the aroma liking of chocolate milk (Fig. 2b and Fig. 3). Their
re
361 effects were almost equivalent as evident from the coefficients of X1 and X2 (eq. 5) while monk
lP
362 fruit extract tails off slightly slower than stevia sweetener as observed from their quadratic
na
363 behaviour graph for aroma liking (Fig. 3). Their overall effects on aroma liking were positive and
364 synergistic with a prediction of 6.31 (Fig. 4 and Table 4) which is supported by previous findings
ur
365 where aroma profile was improved in sugar-free dairy dessert with the combined effect of stevia
Jo
366 and sucralose (Furlán & Campderrós, 2017). The aroma liking for chocolate milk in this study
367 was similar to the flavour liking of skim chocolate milk by Li et al. (2015).
368 The aroma of chocolate milk in the current study is mainly due to cocoa powder, sugar and
369 milk. No additional aroma was added during the preparation of the chocolate milk. The addition
370 of aroma could be employed to enhance the sensory perception while reducing sugar. For
371 example, Alcaire, Antunez, Vidal, Gimenez, & Ares (2017) used vanilla flavour to mitigate the
372 effect of a 20% reduction of added sugar in milk desserts. Further, aroma is also associated with
373 perception of sweetness in a particular product (Prescott, Johnstone, & Francis, 2004) and the
374 taste-smell integration in the brain is related to existing experiences with taste-smell
15
375 combinations. Both tastants and odourants are responsible for overlapping activations in a specific
376 part of the brain that leads to enhanced sweetness perception (Small et al., 2004).
377
378
379
381 Sweetness is a key driver of liking for food products and a heightened liking for sweet tastes
of
382 has been associated with increased intakes of foods with added sucrose (Drewnowski, Mennella,
ro
383 Johnson, & Bellisle, 2012). Therefore, sweetness is given higher importance compared to the
384 -p
other sensory attributes as presented in Table 3. The quadratic equation obtained from RSA data
re
385 shows the effect of stevia sweetener and monk fruit extract on sweetness liking is represented in
lP
Sweetness = + 3.45 + 0.04 × X1 + 0.05 × X2 - 0.0002 × X1X2 - 0.0002 × X12 - 0.0002 × X22….(6)
na
387
Where,
ur
388
Jo
390 Sweetness liking scores varied from 5.50 to 6.50 (Table 1 and Table 3) for a 50% reduction
391 of added sugar. Similar findings were reported by Li et al. (2015) where the sweetness liking
392 scores were 6.3 and 6.2, respectively for monk fruit and stevia leaf extract for 50% reduction as
393 compared to control skim chocolate milk (5.14% sucrose with sweetness liking 6.9).
394 Sweetness liking scores followed a quadratic model as shown in Fig. 2d and Fig. 3.
395 Psychophysical curve provides practical formulation information between the perceived sweetness
396 intensity (on Y-axis) and sweetener concentration (on X-axis) (Cardello, Da Silva, & Damasio,
397 1999). Psychophysical behaviour has been studied for dose-response behaviour across 16
16
398 sweeteners (Wee, Tan, & Forde, 2018); for sweet taste perception associated with sweet food
399 liking and intake (Jayasinghe et al., 2017) as well as oral sensitivity to complex carbohydrates and
400 sweet taste in humans (Low, Lacy, McBride, & Keast, 2017). Unlike the psychophysical curve,
401 where sweetness intensity is on the Y-axis, quadratic modelling graph (Fig. 3) uses sweetness
402 liking on Y-axis. Usually, as sweetness levels increase, a pattern of hedonic responses follows an
403 inverted-U as reported across different cultures (Pfaffmann, 1980; Prescott et al., 1992; Prescott et
404 al., 1997; Prescott, 1998). A similar pattern was observed (Fig. 3) where sweetness liking was
405 increased to a point and then gradually tails off. The tailing off is quicker for stevia sweetener
406 compared to monk fruit extract. This is because both stevia (Reb A) and monk fruit have varied
of
407 temporal sweetness and side taste profiles compared to sucrose with monk fruit having prolonged
ro
408 sweet and honey taste in comparison to stevia (Reb A) (Tan, Wee, Tomic, & Forde, 2019). A
409
-p
similar curve referring to “sensory liking relation” was used for the selection of tomato sauce
re
410 (Moskowitz & Marketo, 2001).
lP
411 Further looking at the interaction coefficient (X1X2) and quadratic coefficients for stevia
na
412 sweetener (X12) and monk fruit extract (X22) in eq. 6, the value (0.0002) is almost negligible
ur
413 suggesting a strong positive and synergistic effect of stevia sweetener and monk fruit extract on
Jo
414 sweetness liking with a prediction of 6.47 (Fig. 4 and Table 4) when used in combination. This
415 finding is supported by earlier research where the use of two sweeteners as a blend
416 (Cyclamate/Saccharin blend, 2:1) enhanced sweetness to meet sweetness equivalent to sucrose
417 probably by minimising off-flavour or bitter after-taste in peach nectar (Cardoso & Bolini, 2007).
418 Similarly, blending aspartame (APM) and acesulfame-K (ACE-K) resulted in a sweetness synergy
419 of approximately 30% (DuBois & Prakash, 2012) as well as blending Reb M with Reb B/ Reb D
420 resulted in sweetness synergy with improvement in sweetness intensity (Prakash et al., 2014).
421
17
423 Consumer liking of chocolate milk is largely dependent on the mouthfeel profile next to
424 sweetness (Paixão, Rodrigues, Esmerino, Cruz, & Bolini, 2014). Mouthfeel depends on various
425 ingredients such as sugar, fat, protein, stabilisers like carrageenan and other thickening agents in
426 the chocolate milk (Li et al., 2015). For this study, we fixed the other ingredients (Fig. 1) for all
427 reformulations except for the variables (stevia sweetener and monk fruit extract) to evaluate their
428 effect on sensory attributes. The quadratic equation showing the effect of stevia sweetener and
430
of
431 Mouthfeel = + 6.43 + 0.08 × X1 + 0.09 × X2 - 0.19 × X1X2 - 0.21 × X12 - 0.16 × X22……(7)
ro
432 Where,
-p
re
433 X1 = Stevia sweetener (ppm); X2 = Monk fruit extract (ppm)
lP
434 Values of mouthfeel liking vary from 5.64 to 6.48 (Table 1 and Table 3) for a 50%
435 reduction of sucrose. Li et al. (2015) reported thickness liking scores similar to the mouthfeel
na
436 scores of the present findings i.e. 6.3 and 6.4, respectively for monk fruit and stevia leaf extract
ur
437 for 50% reduction as compared to control skim chocolate milk (5.14% sucrose with thickness
Jo
438 liking 6.5). The mouthfeel liking revealed a quadratic model similar to other sensory attributes as
439 depicted in Fig. 2e and Fig. 3. Mouthfeel liking increased with an increase in concentration up to
440 a certain point and then gradually tails off. Tailing off is slightly quicker for stevia sweetener
441 compared to monk fruit extract (Fig. 3). This could be due to their varying mouth-drying
442 properties (i.e. the sensation of the lack of lubrication or moistness) as studied by Tan et al. (2019)
443 where mouth-drying was higher for stevia (Reb A) compared to monk fruit. Further, the
444 sweeteners (stevia sweetener and monk fruit extract) showed an almost similar positive effect on
445 mouthfeel liking as evident from the coefficients X1 and X2 (eq. 7). The monk fruit extract tails off
446 slightly slower than stevia sweetener for mouthfeel liking, similar to that observed for aroma
18
447 liking (Fig. 3). Overall, the combined effect of both the sweeteners on the mouthfeel liking was
448 positive and synergistic with a prediction of 6.45 (Fig. 4 and Table 4).
449
451 Aftertaste is another important sensory attribute for consumer liking and acceptance or
452 rejection of any product. Presence of aftertastes like bitterness and metallic can limit the
453 applications of sweeteners for sugar reduction in foods and beverages (Belščak-Cvitanović et al.,
454 2015). The quadratic equation for the effect of stevia sweetener and monk fruit extract on
of
455 aftertaste liking is represented in eq. (8).
ro
456 Aftertaste = + 6.30 + 0.11× X1 + 0.09 × X2 - 0.22 × X1X2 - 0.28 × X12 - 0.03 × X22…..(8)
-p
re
457 Where,
lP
459 Aftertaste liking scores range from 5.52 to 6.41 (Table 1 and Table 3). Li et al. (2015)
ur
460 reported similar values for aftertaste liking i.e. 5.8 for both monk fruit and stevia leaf extract for
Jo
461 50% reduction as compared to control skim chocolate milk (5.14% sucrose with aftertaste liking
462 6.3). Aftertaste liking scores followed a quadratic model as depicted in Fig. 2f and Fig. 3. Stevia
463 sweetener showed an inverted-U shape while the graph for monk fruit extract is almost linear.
464 This is because stevia sweetener contains Reb A which has metallic and bitter aftertastes (DuBois
465 & Prakash, 2012) that causes the graph to tail off after reaching the optimum point. Tan et al.
466 (2019) further reported bitterness, metallic and chemical taste to be more prominent in the
467 temporal profile of stevia (Reb A) compared to the monk fruit which had a prolonged sweet and
468 honey taste. This accounts for the lower coefficient values of (X2) and (X22) in eq. (8).
469 Therefore, when both the sweeteners were used in combination, the bitter and metallic
470 aftertastes of stevia sweetener were masked by the monk fruit extract. This led to the overall
19
471 positive and synergistic effect of stevia sweetener and monk fruit extract on the aftertaste liking
472 with a prediction of 6.33 (Fig. 4 and Table 4). Similar findings were reported in earlier research,
473 where a blend of two sweeteners (Cyclamate/Saccharin blend, 2:1) minimised the off-flavour and
474 bitter after-taste in peach nectar (Cardoso & Bolini, 2007). Further, blending Reb M with Reb B/
475 Reb D improved the onset and bitterness perception (Prakash et al., 2014). This collectively leads
476 to improved sweetness perception and ultimately the overall liking and acceptance of the product
of
479 Stevia sweetener (5-100 ppm) and Monk fruit extract (50-100 ppm) were used for the
ro
480 optimization for 50% reduction of added sugar (i.e. sucrose, 5% to 2.5%) in chocolate flavoured
481
-p
milk using central composite design (CCD) of the RSM software. Following the numerical (Table
re
482 3) and graphical (Fig. 2) optimization processes, an optimum amount of these sweeteners within
lP
483 the selected range i.e. 56.27 ppm of stevia sweetener and 81.90 ppm of monk fruit extract were
484 predicted for achieving the optimum sensory attributes for 50% sugar reduction in chocolate
na
485 flavoured milk. These concentrations resulted in a positive and synergistic effect on the sensory
ur
486 attributes to achieve optimum liking of 6.78, 6.47, 6.31, 6.47, 6.45 and 6.33, respectively for
Jo
487 overall liking, appearance, aroma, sweetness, mouthfeel and aftertaste as the best outcome in
488 terms of desirability function (0.952) (Table 4). Also, the viscosity (8.38 mPa.s) of the chocolate
489 milk with optimum sweetener concentration was close to the control sample (8.34 mPa.s). Further,
490 the predicted sensory attributes’ liking of the 50% reduced added sugar chocolate milk with stevia
491 and monk fruit extract is comparable to the control with full sugar (Table 1).
492 4. Conclusion
493 Optimization of natural non-nutritive sweeteners namely stevia sweetener (5-100 ppm) and
494 monk fruit extract (50-100 ppm) for 50% sucrose reduction (5% to 2.5%) in chocolate flavoured
495 milk could be predicted based on sensory attributes using a central composite design (CCD) from
20
496 RSM. Multiple response optimizations predicted the concentrations of stevia sweetener and monk
497 fruit extract to be 56.27 ppm and 81.90 ppm, respectively for the optimum sensory attributes for
498 sugar reduction in chocolate flavoured milk without much impact as compared to the control with
499 full sugar. Therefore, the combined use of natural non-nutritive sweeteners could be used to
500 partially reduce sugar in food products to meet WHO and PHE sugar guidelines but still maintain
501 consumer liking and acceptance of the product. From an industrial point of view, the results of
502 this study can assist food companies in overcoming the technical challenges underlying sugar
503 reduction in milk-based beverages and achieve at least a small reduction that could benefit the
504 health of the population significantly. The future recommendation is to consider children as
of
505 consumers as well as their demographic variations for the acceptance of the reduced sugar
ro
506 product.
-p
re
507
lP
508 Acknowledgments
na
510 Scholarship and the authors would like to thank TATE & LYLE, London, UK for providing the
ur
511 natural non-nutritive sweeteners as well as Simone Lewin, for her support.
Jo
513 The authors ensure that the consumer study involving humans was carried out in accordance with
514 The Code of Ethics of the World Medical Association (Declaration of Helsinki), as per the
515 approved ethics, Human Ethics Advisory Group, Faculty of Health (HEAG-H 155_2018),
519
21
520 References
521 Agoda-Tandjawa, G., Le Garnec, C., Boulenguer, P., Gilles, M., & Langendorff, V. (2017).
523 biopolymer type and chemical characteristics. Food Hydrocolloids, 73, 300-312.
524 Alcaire, F., Antunez, L., Vidal, L., Gimenez, A., & Ares, G. (2017). Aroma-related cross-modal
525 interactions for sugar reduction in milk desserts: Influence on consumer perception. Food
527 Aslan, N. (2007). Application of response surface methodology and central composite rotatable
528 design for modeling the influence of some operating variables of a multi-gravity separator
of
529 for coal cleaning. Fuel, 86, 769-776.
ro
530 Australia and New Zealand Food Standards Code. (2018). Application to Amend the
531
-p
Specifications for Rebaudioside M Under Australia and New Zealand Food Standards
re
532 Code – Standard.
lP
533 Azami, T., Niakousari, M., Hashemi, S. M. B., & Torri, L. (2018). A three-step sensory-based
na
535 flavored milk drink. LWT- Food Science and Technology, 91, 375-381.
536 Ban, Q., Cheng, J., Sun, X., Jiang, Y., Zhao, S., Song, X., & Guo, M. (2020). Effects of a
Jo
537 synbiotic yogurt using monk fruit extract as sweetener on glucose regulation and gut
538 microbiota in rats with type 2 diabetes mellitus. Journal of Dairy Science, 103, 2956-2968.
539 Bayarri, S., Carbonell, I., Barrios, E. X., & Costell, E. (2011). Impact of sensory differences on
540 consumer acceptability of yoghurt and yoghurt-like products. International Dairy Journal,
542 Belščak-Cvitanović, A., Komes, D., Dujmović, M., Karlović, S., Biškić, M., Brnčić, M., & Ježek,
543 D. (2015). Physical, bioactive and sensory quality parameters of reduced sugar chocolates
544 formulated with natural sweeteners as sucrose alternatives. Food Chemistry, 167, 61-70.
22
545 Blake, M. R., Lancsar, E., Peeters, A., & Backholer, K. (2019). Sugar-sweetened beverage price
546 elasticities in a hypothetical convenience store. Social Science & Medicine, 225, 98-107.
547 Bordi Jr, P. L., Palchak, T., Verruma-Bernardi, M. R., & Cho, H. C. (2016). Adult Acceptance of
548 Chocolate Milk Sweetened with Stevia. Journal of Culinary Science & Technology, 14,
549 216-221.
550 Brownie, S., Muggleston, H., & Oliver, C. (2015). The 2013 Australian dietary guidelines and
551 recommendations for older Australians. Australian Family Physician, 44, 311.
552 Cardello, H. M. A. B., Da Silva, M. A. P. A., & Damasio, M. H. (1999). Measurement of the
553 relative sweetness of stevia extract, aspartame and cyclamate/saccharin blend as compared
of
554 to sucrose at different concentrations. Plant Foods for Human Nutrition, 54, 119-129.
ro
555 Cardoso, J. M. P., & Bolini, H. M. A. (2007). Different sweeteners in peach nectar: Ideal and
556
-p
equivalent sweetness. Food Research International, 40, 1249-1253.
re
557 Chen, W., Wang, J., Qi, X., & Xie, B. (2007). The antioxidant activities of natural sweeteners,
lP
558 mogrosides, from fruits of Siraitia grosvenori. International Journal of Food Sciences and
na
560 Chollet, M., Gille, D., Schmid, A., Walther, B., & Piccinali, P. (2013). Acceptance of sugar
562 Christoph, I. B., Peter, G., Rothe, A., Salamon, P., Weber, S. A., & Weible, D. (2011). School
563 milk consumption in Germany–what are important product attributes for children and
564 parents? In EAAE 2011 Congress Change and Uncertainty. ETH Zurich, Zurich,
565 Switzerland.
566 Chun, L., Li-Mei, L., Feng, S., Zhi-Min, W., Hai-Ru, H., Li, D., & Jiang, T.-L. (2014). Chemistry
23
569 Cleghorn, C., Blakely, T., Mhurchu, C. N., Wilson, N., Neal, B., & Eyles, H. (2019). Estimating
570 the health benefits and cost-savings of a cap on the size of single serve sugar-sweetened
573 https://www.dairyaustralia.com.au/industry/production-and-sales/consumption-summary
574 de Melo, L. L. M. M., Bolini, H. M. A., & Efraim, P. (2009). Sensory profile, acceptability, and
575 their relationship for diabetic/reduced calorie chocolates. Food Quality and Preference,
577 Drewnowski, A., Mennella, J. A., Johnson, S. L., & Bellisle, F. (2012). Sweetness and food
of
578 preference. The Journal of Nutrition, 142, 1142S-1148S.
ro
579 DuBois, G. E., & Prakash, I. (2012). Non-caloric sweeteners, sweetness modulators, and
580
-p
sweetener enhancers. Annual Review of Food Science and Technology, 3, 353-380.
re
581 EFSA. (2010). Scientific opinion on the safety of steviol glycosides for the proposed uses as a
lP
582 food additive. EFSA Panel of Food Additive and Nutrients Sources added to Food (ANS).
na
584 EFSA. (2015). Scientific opinion on the safety of the proposed amendment of the specifications
585 for steviol glycosides (E 960) as a food additive. EFSA Panel of Food Additive and
Jo
586 Nutrient Sources added to Food (ANS). EFSA Journal, 13, 4316.
587 Egger, G., Swinburn, B., & Rossner, S. (2003). Dusting off the epidemiological triad: could it
589 Falbe, J., Thompson, H. R., Patel, A., & Madsen, K. A. (2019). Potentially addictive properties of
591 FPED. (2014). What We Eat in America. NHANES 2011-2012, individuals 2 years and over
592 (excluding breast-fed children), day 1 dietary intake data, weighted. Food Patterns
24
594 Furlán, L. T. R., & Campderrós, M. E. (2017). The combined effects of Stevia and sucralose as
595 sugar substitute and inulin as fat mimetic on the physicochemical properties of sugar-free
596 reduced-fat dairy dessert. International Journal of Gastronomy and Food Science, 10, 16-
597 23.
598 Guggisberg, D., Piccinali, P., & Schreier, K. (2011). Effects of sugar substitution with Stevia,
600 characteristics of low-fat and whole milk set yogurt. International Dairy Journal, 21, 636-
601 644.
602 Hanks, A. S., Just, D. R., & Wansink, B. (2014). Chocolate milk consequences: a pilot study
of
603 evaluating the consequences of banning chocolate milk in school cafeterias. PLOS One, 9,
ro
604 e91022.
605
-p
Henry, C., Whiting, S. J., Phillips, T., Finch, S. L., Zello, G. A., & Vatanparast, H. (2015). Impact
re
606 of the removal of chocolate milk from school milk programs for children in Saskatoon,
lP
608 Jayasinghe, S. N., Kruger, R., Walsh, D. C., Cao, G., Rivers, S., Richter, M., & Breier, B. H.
ur
609 (2017). Is sweet taste perception associated with sweet food liking and intake? Nutrients,
610 9, 750.
Jo
611 Keast, R., Sayompark, D., Sacks, G., Swinburn, B., & Riddell, L. (2011). The influence of
614 Kemp, D. S. E., Hollowood, D. T., & Hort, D. J. (2009). Sensory test methods. Sensory
616 Kim, N.-C., & Kinghorn, A. D. (2002). Highly sweet compounds of plant origin. Archives of
25
618 Krachler, B., Eliasson, M., Stenlund, H., Johansson, I., Hallmans, G., & Lindahl, B. (2006).
619 Reported food intake and distribution of body fat: a repeated cross-sectional study.
621 Lawless, H. T., & Heymann, H. (2010). Sensory evaluation of food: principles and practices:
623 Lee, Y. J., Jeong, J., Kim, M. O., & Nam, J.-O. (2016). The positive effect of luohanguo as sugar
626 Li, X. E., Lopetcharat, K., & Drake, M. A. (2014). Extrinsic attributes that influence parents’
of
627 purchase of chocolate milk for their children. Journal of Food Science, 79, S1407-S1415.
ro
628 Li, X. E., Lopetcharat, K., & Drake, M. A. (2015). Parents’ and children's acceptance of skim
629
-p
chocolate milks sweetened by monk fruit and stevia leaf extracts. Journal of Food Science,
re
630 80, S1083-S1092.
lP
631 Liu, C., Dai, L., Liu, Y., Rong, L., Dou, D., Sun, Y., & Ma, L. (2016). Antiproliferative activity of
na
632 triterpene glycoside nutrient from monk fruit in colorectal cancer and throat cancer.
ur
634 Low, J. Y. Q., Lacy, K. E., McBride, R. L., & Keast, R. S. J. (2017). Evidence supporting oral
Jo
637 Ludwig, D. S., Peterson, K. E., & Gortmaker, S. L. (2001). Relation between consumption of
638 sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. The
640 Lustig, R. H., Schmidt, L. A., & Brindis, C. D. (2012). Public health: the toxic truth about sugar.
26
642 Mahato, D. K., Keast, R., Liem, D. G., Russell, C. G., Cicerale, S., & Gamlath, S. (2020). Sugar
643 Reduction in Dairy Food: An Overview with Flavoured Milk as an Example. Foods, 9(10),
644 1400.
645 Malik, V. S., Schulze, M. B., & Hu, F. B. (2006). Intake of sugar-sweetened beverages and weight
646 gain: a systematic review. The American Journal of Clinical Nutrition, 84, 274-288.
647 Morais, E., Morais, A., Cruz, A., & Bolini, H. (2014). Development of chocolate dairy dessert
648 with addition of prebiotics and replacement of sucrose with different high-intensity
650 Moskowitz, H. R., & Marketo, C. (2001). Selecting products for Category Appraisal Studies—
of
651 fewer products do almost as well as many products. Journal of Sensory Studies, 16, 537-
ro
652 549.
653
-p
Nicklas, T. A., O'neil, C., & Fulgoni, V. (2017). Flavored milk consumers drank more milk and
re
654 had a higher prevalence of meeting calcium recommendation than nonconsumers. Journal
lP
656 Oltman, A. E., Lopetcharat, K., Bastian, E., & Drake, M. A. (2015). Identifying key attributes for
ur
658 Ozdemir, C., Arslaner, A., Ozdemir, S., & Allahyari, M. (2015). The production of ice cream
Jo
659 using stevia as a sweetener. Journal of Food Science and Technology, 52, 7545-7548.
660 Paixão, J., Rodrigues, J., Esmerino, E., Cruz, A., & Bolini, H. (2014). Influence of temperature
661 and fat content on ideal sucrose concentration, sweetening power, and sweetness
662 equivalence of different sweeteners in chocolate milk beverage. Journal of Dairy Science,
664 Pandey, A. K., & Chauhan, O. P. (2019). Monk fruit (Siraitia grosvenorii)-health aspects and
666 Parker, M. N., Lopetcharat, K., & Drake, M. A. (2018). Consumer acceptance of natural
27
668 Patel, A. I., Moghadam, S. D., Freedman, M., Hazari, A., Fang, M. L., & Allen, I. E. (2018). The
669 association of flavored milk consumption with milk and energy intake, and obesity: A
671 Pawar, R. S., Krynitsky, A. J., & Rader, J. I. (2013). Sweeteners from plants—with emphasis on
672 Stevia rebaudiana (Bertoni) and Siraitia grosvenorii (Swingle). Analytical and
674 Pfaffmann, C. (1980). Wundt's schema of sensory affect in the light of research on gustatory
676 PHE. (2017). Sugar reduction: achieving the 20%. A technical report outlining progress to date,
of
677 guidelines for industry, 2015 baseline levels in key foods and next steps. London, UK:
ro
678 Public Health England. In.
679
-p
Pon, S., Lee, W., & Chong, G. (2015). Textural and rheological properties of stevia ice cream.
re
680 International Food Research Journal, 22, 1544.
lP
681 Prakash, I., Markosyan, A., & Bunders, C. (2014). Development of next generation stevia
na
683 Prescott, J., Laing, D., Bell, G., Yoshida, M., Gillmore, R., Allen, S., Yamazaki, K., & Ishii, R.
684 (1992). Hedonic responses to taste solutions: a cross-cultural study of Japanese and
Jo
686 Prescott, J., Bell, G. A., Gillmore, R., Yoshida, M., O'sullivan, M., Korac, S., Allen, S., &
689 Prescott, J. (1998). Comparisons of taste perceptions and preferences of Japanese and Australian
690 consumers: Overview and implications for cross-cultural sensory research. Food Quality
692 Prescott, J., Johnstone, V., & Francis, J. (2004). Odor–taste interactions: effects of attentional
28
694 Qi, X.-Y., Chen, W.-J., Zhang, L.-Q., & Xie, B.-J. (2008). Mogrosides extract from Siraitia
695 grosvenori scavenges free radicals in vitro and lowers oxidative stress, serum glucose, and
696 lipid levels in alloxan-induced diabetic mice. Nutrition Research, 28, 278-284.
697 Rad, A. H., Delshadian, Z., Arefhosseini, S. R., Alipour, B., & Jafarabadi, M. A. (2012). Effect of
698 inulin and stevia on some physical properties of chocolate milk. Health Promotion
700 Ruiz-Ojeda, F. J., Plaza-Díaz, J., Sáez-Lara, M. J., & Gil, A. (2019). Effects of sweeteners on the
701 gut microbiota: a review of experimental studies and clinical trials. Advances in Nutrition,
of
703 Singh, J., Rasane, P., Kaur, S., Kumar, V., Dhawan, K., Mahato, D. K., Malhotra, S., Sarma, C.,
ro
704 Kaur, D. & Bhattacharya, J. (2020). Nutritional interventions and considerations for the
705
-p
development of low calorie or sugar free foods. Current Diabetes Reviews, 16, 301-312.
re
706 Small, D. M., Voss, J., Mak, Y. E., Simmons, K. B., Parrish, T., & Gitelman, D. (2004).
lP
707 Experience-dependent neural integration of taste and smell in the human brain. Journal of
na
709 Striegel-Moore, R. H., Thompson, D., Affenito, S. G., Franko, D. L., Obarzanek, E., Barton, B.
710 A., Schreiber, G. B., Daniels, S. R., Schmidt, M., & Crawford, P. B. (2006). Correlates of
Jo
711 beverage intake in adolescent girls: the National Heart, Lung, and Blood Institute Growth
713 Sun, B.-S., Chen, Y.-P., Wang, Y.-B., Tang, S.-W., Pan, F.-Y., Li, Z., & Sung, C.-K. (2012).
714 Anti-obesity effects of mogrosides extracted from the fruits of Siraitia grosvenorii
716 Tan, V. W. K., Wee, M. S. M., Tomic, O., & Forde, C. G. (2019). Temporal sweetness and side
29
719 Te Morenga, L., Mallard, S., & Mann, J. (2013). Dietary sugars and body weight: systematic
720 review and meta-analyses of randomised controlled trials and cohort studies. British
722 Thompson, J. L., Lopetcharat, K., & Drake, M. A. (2007). Preferences for commercial strawberry
723 drinkable yogurts among African American, Caucasian, and Hispanic consumers in the
725 Wee, M., Tan, V., & Forde, C. (2018). A comparison of psychophysical dose-response behaviour
of
727
ro
728
-p
re
lP
na
ur
Jo
30
Fig. 1. Flow diagram for the preparation of chocolate flavoured milk.
Fig. 2. Effect of stevia sweetener and monk fruit extract concentration on different sensory
attributes: (a) Overall liking; (b) Appearance liking; (c) Aroma liking; (d) Sweetness liking;
(e) Mouthfeel liking and (f) Aftertaste liking, for 50% reduction of added sugar in chocolate
flavoured milk.
Fig. 3. Quadratic modelling graph for concentration of natural sweeteners (stevia sweetener
and monk fruit extract) and the prediction for individual sensory attribute: overall liking;
appearance; aroma; sweetness; mouthfeel and aftertaste. The cross point on the middle
prediction line denotes the optimum concentration of each sweetener for individual sensory
of
attribute for maximum desirability function.
ro
Fig. 4. Interaction effects of stevia sweetener (denoted by black prediction lines) and monk
fruit extract (denoted by red prediction lines) for the sugar reduction in chocolate flavoured
-p
milk. The dotted lines denote 95% confidence interval (CI) while the cross point denotes the
re
optimum prediction for individual sensory attribute after the synergistic effect of the two
sweeteners.
lP
Table 1 The experimental design and sensory scores* for optimization with different
na
concentrations of natural sweeteners (Stevia sweetener and Monk fruit extract) for 50%
reduction of added sugar in chocolate flavoured milk by Response Surface Methodology
ur
(RSM).
Jo
Table 2 Model statistics (Analysis of Variance, ANOVA) and adequacy of the models for
sensory attributes of chocolate milk.
Table 3 Criteria and outputs of the numerical optimization for added sugar reduction in
chocolate flavoured milk using stevia sweetener and monk fruit extract.
X1: Stevia X2: Monk fruit Overall Appearance Aroma Sweetness Mouthfeel Aftertaste
sweetener (ppm) extract (ppm) liking
1 5.00 50.00 6.06 6.07 5.80 5.50 5.64 5.52
f
oo
2 100.00 50.00 6.70 6.35 6.20 6.17 6.25 6.15
3 52.50 75.00 6.76 6.42 6.26 6.48 6.40 6.29
pr
4 100.00 100.00 6.46 6.20 5.95 5.75 6.00 5.92
e-
5 52.50 75.00 6.72 6.45 6.30 6.42 6.38 6.35
Pr
6 119.68 75.00 6.67 6.32 6.12 5.89 6.15 5.95
al
7 52.50 75.00 6.74 6.50 6.28 6.45 6.44 6.33
rn
8 52.50 39.64 6.55 6.25 6.03 6.08 6.02 6.18
9 52.50 75.00 u 6.80 6.46 6.32 6.46 6.46 6.25
Jo
10 0.00 75.00 6.24 6.13 5.90 5.63 5.98 5.62
11 5.00 100.00 6.65 6.29 6.15 6.21 6.18 6.18
12 52.50 110.36 6.72 6.37 6.24 6.35 6.30 6.41
13 52.50 75.00 6.78 6.48 6.34 6.50 6.48 6.30
Control - - 6.88 6.64 6.39 6.28 6.55 6.50
*Scored on a 9-point hedonic scale; ppm- parts per million; Control- Full sugar version (10%)
Table 2
Response Model Lack-of-fit R2 Adjusted R2 Std. dev. C.V.% F value
(P value) (P value)
(< 0.0001)
f
oo
*Appearance Quadratic 0.27 0.96 0.93 0.04 0.56 34.07
pr
(< 0.0001)
e-
*Aroma Quadratic 0.12 0.96 0.93 0.05 0.75 33.08
Pr
(< 0.0001)
al
rn
*Sweetness Quadratic 0.06 0.99 0.98 0.05 0.88 97.59
(< 0.0001) u
Jo
*Mouthfeel Quadratic 0.09 0.96 0.93 0.07 1.07 31.54
(0.0001)
(< 0.0001)
f
oo
Appearance maximize 6.07 6.50 1 1
pr
Aroma maximize 5.80 6.34 1 1 3
e-
Sweetness maximize 5.50 6.50 1 1 5
Pr
Mouthfeel maximize 5.64 6.48 1 1 3
Aftertaste maximize 5.52 6.41 1 1 3
al
rn
*Weight range (0.10 to 10)-fine tunes the optimization process for the best solution. A low weight (near 0.10) allows more solutions that don’t quite meet the optimal goal while a
u
high weight (close to 10) causes the optimization to seek a solution close to or beyond the stated goal. So, from a practical standpoint, the weight is kept at 1. **Importance: 1 is
Jo
assigned for the response with the lowest importance (+); 3 for medium importance (+++) and 5 for critical importance (+++++).
Table 4
Solutions Stevia Monk Fruit Overall Appearance Aroma Sweetness Mouthfeel Aftertaste Desirability
(TASTEVA®) (PUREFRUITTM)
(ppm) (ppm)
1 56.268 81.903 6.779 6.465 6.307 6.471 6.445 6.330 0.952 *Selected
f
oo
2 56.271 81.619 6.779 6.466 6.307 6.472 6.455 6.329 0.952
pr
* Response Surface Methodology (RSM) selected solution based on the criteria set in Table 3
e-
Pr
al
u rn
Jo
Cocoa powder
Premix stabilizer (Ticaloid®750, 0.02%), sucrose (5%), and NOVATIONTM 3300 starch
(0.5%)
Add Stevia sweetener (TASTEVA®, 5-100 ppm) and Monk fruit extract (PUREFRUITTM,
50-100 ppm)
of
ro
Add cocoa solution to the premix
-p
Cool the solution to 20°C
re
Add pasteurized milk (3.5% w/v, protein; 3.3% w/v, fat; and 4.6% w/v, carbohydrate)
lP
Packaging
Storage at 4°C
Fig. 1.
(a) (b) (c)
f
oo
pr
e-
Pr
(d) (e)
(f)
al
rn
u
Jo
Fig. 2.
Jo
ur
na
lP
re
-p
ro
Fig. 3.
of
Fig. 4.
Jo
urn
al
Pr
e-
pr
oo
f
Highlights
1. It is possible to achieve a 50% of added sugar reduction in chocolate milk using stevia
and monk fruit compounds
2. The use of two sweeteners provide synergistic effects on sensory attributes
3. 56.27 ppm of stevia sweetener and 81.90 ppm of monk fruit extract showed optimum
sensory response with overall liking score of 6.78
4. The optimum sensory response is comparable to the control sample with full sugar
having overall liking score of 6.88
of
ro
-p
re
lP
na
ur
Jo
Credit Author Statement
Kumar Mahato: Data collection, Writing- Original draft preparation. Djin Gie Liem:
of
ro
-p
re
lP
na
ur
Jo
Declaration of interests
☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.
☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:
Signature:
of
ro
-p
re
lP
na
ur
Jo