Optimization of sorbitol, fructooligosaccharides and sugar levels in the syrup based on physicochemical properties and sensory acceptance of healthy, sweet egg yolk drop (a traditional egg-based dessert) using response surface methodology

ABSTRACT

Sweet egg yolk drops are one of the traditional Thai desserts made by dropping egg yolk batter into a boiling syrup with high sugar concentration to form teardrop shapes. The objectives of this study were to optimize the ratio of sugar substitutes (sorbitol and fructooligosaccharides (FOS)) and reduce the amount of sugar in syrup for the development of sugar-reduced egg yolk desserts. Effects of three ratios of sorbitol to FOS (25:75, 50:50, and 75:25) and four percentages of sugar reduction (25, 50, 75, and 100%) on the physical, chemical and sensory properties of desserts were also investigated. Results indicated that increasing the ratio of sorbitol and decreasing the percentage of sugar in syrup led to desserts with decreased color a* and b* values, bulk density, hardness, cohesiveness, soluble fiber, and reducing-sugar content. The opposite effect was observed in desserts made with syrup, which had higher ratios of FOS and lower levels of sugar. The optimum formulation of desserts based on response surface methodology (RSM) contained a ratio of 25:75–30:70 of sorbitol to FOS and 60–65% reduction of sugar in syrup. The comparison of the predicted values from the model of the optimized dessert with experimental values showed that the model was well-fitted. Moreover, the optimized dessert had higher sensory acceptability when compared to the full-sugar sweet egg yolk drops. Therefore, the results indicate that sorbitol and FOS could be used as sugar substitutes in syrup to produce a healthier and lower calorie egg-based dessert.

KEYWORDS:

• Egg-based desserts
• fructooligosaccharides
• response surface methodology
• sorbitol
• sugar reduction
• sweet egg yolk drops

Introduction

Non-communicable diseases (NCDs) have become an important public health issue due to inappropriate food consumption behavior.^[1] One of the NCDs that is of a great concern is type 2 Diabetes Mellitus, which is primarily associated with obesity and overweight. Type 2 Diabetes Mellitus is generally caused by prolonged elevation in blood glucose levels (hyperglycemia) that lead to insulin resistance.^[2] During the past few decades, the prevalence rates of obesity and type 2 diabetes have steadily increased around the world. According to worldwide estimates, there are approximately 463 million people living with diabetes in 2019, and by 2045 the number is anticipated to increase to 700 million.^[3,4] The World Health Organization (WHO), reports that type 2 diabetes tends to occur among adults, but recent studies have indicated that the prevalence of this disease is rising among children.^[5] Consequently, consumers are becoming more aware of the relationship between diet and health, which leads to an increase in the demand for healthy foods.^[6]

Desserts are popular in several cultures and often consist of sugar and fat.^[7,8] Since the prevalence of diabetes rates are increasing globally, the consumer demand for low-sugar foods is therefore increasing.^[9] In the past several years, the development of healthier desserts has emphasized adapting traditional formulas with less sugar (or sugar free) or using sugar substitutes with different types of sweeteners in order to lessen consumer’s concern. However, a sugar reduction affects not only the sweetness of the product, but also its viscosity, texture formation, volume, crust color and moisture content.^[10] A major challenge in modifying sweet ingredients is to create sugar-reduced products that possess similarities in characteristics and taste to full-sugar products. Therefore, the selection of appropriate sweetener types and proportions has been essential for the development of successful low-sugar desserts. In general, there have been numerous studies examining sugar reduction and sweetener substitutions in dairy- and plant-based foods, such as ice cream, chocolate, cakes, and biscuits, but there are very few studies focusing on egg-based foods. One of the popular Thai egg-based, sugar-rich desserts, so called “Sweet Egg Yolk Drop” or “Thong Yod,” is among the products that needed further investigation. This dessert contains a large amount of sugar, because it is made from egg yolks mixed with rice flour and dropped into boiling highly concentrated sugar syrups. The National Food Institute of Thailand^[11] reports that 100 g of sweet egg yolk drops contain 53.98 to 54.22 g of sucrose, which are approximately twice the recommended daily intake of sucrose (25 g or 6 teaspoons) according to WHO.^[12]

A sugar substitute, such as sorbitol, has a low-calorie content (2.6–3.0 kcal/g) and a relative sweetness of 0.6 as compared to sucrose’s (0.1). It is highly soluble in water, improves the stability of solutions, and reduces the crystallization of syrups. In addition, because sorbitol neither raises blood glucose levels nor requires insulin for its metabolism, it is suitable for a variety of sugar-free products (e.g., ice-cream, jams, baked goods, confectionary).^[13–15] However, only one kind of sugar substitute is not able to replace all of the major functions of sucrose; therefore, adding some agents to enhance other functions, such as structure-building substances (inulin, polydextrose, and dietary fiber), is an effective technique for improving product qualities.^[10] Fructooligosaccharide (FOS) is a soluble dietary fiber classified as a prebiotic. Because this fiber is indigestible by human digestive enzymes, it stimulates the growth of beneficial intestinal bacteria, such as bifidobacteria.^[16] FOS is low in calories (1–1.5 kcal/g), moderately sweet (one-third of sucrose), and can reduce energy intake by 65–75% when compared to other digestible carbohydrates.^[17,18] As a functional ingredient, FOS is not only beneficial as a prebiotic but also capable of improving the physicochemical properties of food products as a replacement for fats and sugars.^[19,20]

Therefore, the objectives of this study were (1) to evaluate the effects of substituting sorbitol and FOS in sugar syrup on the physicochemical and sensory properties of sweet egg yolk drops and (2) to develop a healthy egg-based dessert by using response surface methodology (RSM) to optimize sugar reduction levels and ratios of sugar substitutes (sorbitol mixed with FOS). From this study, a healthier and lower calorie egg-based dessert developed in this work will be beneficial for the food industries and health-conscious consumers.

Materials and methods

Materials

In order to prepare the sweet egg yolk drops, fresh duck eggs (Kasemchai Food (KCF), Nakhon Pathom, Thailand), rice flour (New Grade, Bangkok, Thailand), refined sugar (Lin, Kanchanaburi, Thailand), water (Nestle, Ayutthaya, Thailand), and pandan jasmine flavor (Winner’s, Bangkok, Thailand) were purchased from local markets. The sugar substitutes used were sorbitol powder (Shandong Tianli Pharmaceutical Co., Ltd., China) and FOS powder (FOS-P Powder, Meiji, Korea).

Preparation of sweet egg yolk drops

According to traditional recipes, sweet egg yolk drops were prepared in two steps. The first step is to dissolve refined sugar in boiling water to make a cooking syrup (1:0.6 w/w; 70 ± 2°Brix) and a soaking syrup (1:1 w/w; 56 ± 2°Brix). Artificial flavors (pandan-jasmine flavor) were then dropped into both syrups in small amounts. For the second step, the duck egg yolks were whipped with a wire whip for 8 min at a speed of 10 in a mixer (Model 5K5SS, KitchenAid_®). In the following step, the sifted rice flour was mixed with the beaten egg yolk for 2 min and allowed to stand at room temperature for 5 min. The battered mixtures were poured into a hand-held batter dispenser with a bottom opening of 16.20 mm. During the cooking process, a drop batter was added to the cooking syrup at 90°C for 15 min (each set contained 120 pieces). Immediately following cooking, the dessert drops were placed in the soaking syrup for 10 min at room temperature. The sweet egg yolk drops were immediately removed from the syrup and placed in a sealed plastic container in the refrigerator. Throughout the study, dessert samples were measured and stored at 8°C.

Experimental design

Based on the formulation of the full-sugar dessert, the sugar-reduced sweet egg yolk drops were prepared using the above-mentioned sugar substitutes in syrup. A 3$\times$4 factorial arrangement in a completely randomized design (CRD) was used to investigate the effects of three levels of sorbitol and FOS ratios (25:75, 50:50, and 75:25) and four levels of sugar reduction percentage (25, 50, 75, and 100%) on the properties of desserts.

The response surface methodology (RSM) was used to optimize the syrup formulation to produce the healthy, sweet egg yolk drops. The correlation between independent and dependent variables was modeled using a second-order polynomial model to predict the experiments following equation (1)(1) $Y={\beta }_0+{\beta }_i{X}_i+{\beta }_j{X}_j+{\beta }_{ii}{X}_i^2+{\beta }_{jj}{X}_j^2+{\beta }_{ij}{X}_i{X}_j$(1) .

(1) $Y={\beta }_0+{\beta }_i{X}_i+{\beta }_j{X}_j+{\beta }_{ii}{X}_i^2+{\beta }_{jj}{X}_j^2+{\beta }_{ij}{X}_i{X}_j$(1)

where Y is defined as the dependent variable, ${\beta }_0$ is the constant, ${\beta }_i$ and ${\beta }_j$ are the linear coefficients, ${\beta }_{ii}$ and ${\beta }_{jj}$ are the quadratic coefficients, and ${\beta }_{ij}$ is the interaction coefficient while $X_i$ and $X_j$ are levels of the independent variables. The coefficient of determination (R²) and the lack of fit test were used to assess the adequacy of the model.

The fitted quadratic polynomial equation was used to develop two-dimensional contour plots of the interaction between different factors mentioned above. Finally, each contour plot within the same variable set was superimposed to identify the optimum region that provided the optimal syrup formulation for the development of healthy, sweet egg yolk drops.

Physical properties

The color of dessert samples was measured using a HunterLab colorimeter (ColorFlex 45/0, Hunter Associates Laboratory Inc., Reston, VA., U.S.A.) with a D65 illuminant at 10° observation. Color values were reported in the CIELAB system as L* (lightness/darkness), a* (redness/greenness), and b* (yellowness/blueness) values. The sample color was determined by placing five dessert pieces from each preparation in glass containers and rotating them 90 degrees during the measurement (two determinations). The mean values of the color parameters were determined by a triplicate measurement for each formulation.

The water activity (a_w) was determined with a water activity meter (Aqualab_®, Series 4TE, Decagon, USA). The expansion ratio was calculated by comparing the diameter of ten samples with the diameter of the bottom openings of hand-held batter dispensers (16.20 mm) using a digital vernier caliper (Model 500-150-30, Mitutoyo, Japan). The bulk density of the desserts was determined by measuring their volume using the seed displacement method and weighing them (5 pieces per set), following a slightly modified method of Zhuang et al.^[21] The bulk density (g/cm³) was calculated by dividing the weight of the dessert by its bulk volume. These measurements were carried out in triplicate.

The texture properties of dessert samples were analyzed using a TA-XT plus Texture Analyzer (Stable Micro System, Texture Technologies Corp., NY, USA). A compression probe (P/35: aluminum cylinder) was used to determine the hardness (N) and cohesiveness of the sample. The texture analysis was performed at a pretest speed of 5.0 mm/s, a test speed of 2.5 mm/s, a posttest speed of 5.0 mm/s, a distance of 8.0 mm, and a force of 5 g. Each treatment was measured with ten pieces of dessert samples to determine the average values for hardness and cohesiveness.

Chemical properties

The moisture content was determined by gravimetric method (no. 925.10) using a vacuum oven (BINDER VD 115, Tuttlingen, Germany) at 70°C under pressure ≤100 mmHg until a constant weight was reached. The soluble dietary fiber content and the reducing sugar content (invert sugar per 100 g) were determined as described in method no. 993.19 and 925.36, respectively. The sucrose content in samples was measured using an HPLC method (no. 977.20), according to all previous methods and procedures provided by AOAC.^[22] The energy value of the sample was estimated based on the conversion factors for protein and carbohydrates (4 kcal/g) and lipids (9 kcal/g). All analyses were performed in triplicate.

As part of the comparison of nutrition values between the optimized sugar-reduced dessert and the full-sugar dessert, the FOS content was analyzed using gas-liquid chromatography as modified from the in-house method based on AOAC method 997.08 and Joye and Hoebregs,^[22,23] while the in vitro glycemic index (GI) was determined according to the methodology described by Goñi et al.^[24] All analyses were performed in triplicate.

Sensory properties

An untrained panel of 60 panelists was recruited from Ramkhamhaeng University, Department of Science Service, and Sukhothai Thammathirat Open University. This study was approved for ethical clearance by the Ethics Committee for Human Research, Ramkhamhaeng University, Bangkok, Thailand (RU-HRE 63/0058). The participants were selected based on general dessert preferences, no egg allergies, and no diabetes risk factors. An evaluation of sensory attributes including appearances, color, flavor, taste, juiciness, firmness, and overall liking was conducted with a 9‐point hedonic scale (1 = dislikeextremely, 5 = neither like nor dislike, and 9 = like extremely). In each sample, three pieces were served and coded using three-digit numbers. The panelists were instructed to rinse their mouths with water between sampling sets and to take a 10-minute break between sets of samples.

Statistical analysis

The experimental data were analyzed using the analysis of variance (ANOVA) in SPSS^© statistical package, version 12.0 (SPSS (Thailand) Co., Ltd., Bangkok, Thailand). The Duncan’s Multiple Range Test was used to establish the statistically significant differences between the mean values at p ≤ .05. The RSM was performed using a trial version of the STATISTICA 10 software package (StatSoft, Inc., Tulsa, OK, USA). For model validation, the observed and predicted values under the selection of syrup formulations were evaluated by calculating the correlation coefficient (r) and root mean-square error (RMSE) values.

Results and discussion

Color and water activity content (a_w)

The color of all sweet egg yolk drops desserts containing sorbitol mixed with FOS and with sugar reduction was yellowish-orange (Table 1 and Figure 1). This color effect was due to the presence of a high amount of carotenoid pigment in the duck egg yolks and the Maillard reaction occurring during the sugar syrup cooking process. Table 1 indicates that sorbitol/FOS ratios, sugar reduction percentages, and their interactions affected the color and a_w properties of dessert samples significantly (p ≤ .05). A reduction in sugar content in syrups led to a decrease in all color parameters for lightness (L*), redness (a*) and yellowness (b*) values in comparison to the full-sugar desserts (p ≤ .05). The food processing at high temperatures and pH changes can convert sucrose to glucose and fructose, which are both monosaccharides that exhibit the characteristics of reducing sugars.^[25] Consequently, the different sugar levels affected the amount of reducing sugar in syrup, which led to the varying degrees of Maillard reaction when reacted with amino acids/proteins. As a result, the dessert appeared in a variety of brown shades. However, sorbitol is a sugar alcohol, which does not have the reducing sugar properties and is not involved in the Maillard reaction.^[26] In the presence of FOS in syrup, the sugar-reduced desserts appeared to be darker in color, as indicated by an increase in redness and yellowness values, as well as a decrease in lightness. It is possible that the partial hydrolysis of this soluble fiber to low molecular weight compounds, such as monosaccharides, facilitates the Maillard reaction.^[27]

Figure 1. A variation in the characteristics of sweet egg yolk drops.

Table 1. Color and water activity properties of sweet egg yolk drops as affected by sorbitol with FOS ratio and percentage sugar reduction.

Experiments		Color			aw
Sorbitol : FOS	Sugar reduction (%)	L*	a*	b*
Control		95.17^b ± 0.62	4.28^b ± 1.27	83.32^c ± 2.95	0.837^d ± 0.001
25 : 75	25	95.88^a ± 0.15	5.62^a ± 0.22	86.04^ab ± 0.62	0.807^f ± 0.000
	50	95.94^a ± 0.08	5.66^a ± 0.22	86.43^a ± 0.22	0.819^e ± 0.004
	75	95.77^a ± 0.18	5.33^ab ± 0.50	85.68^abc ± 0.98	0.876^c ± 0.004
	100	95.56^ab ± 0.26	4.93^ab ± 0.50	84.95^abc ± 0.98	0.907^a ± 0.000
50 : 50	25	95.86^a ± 0.11	5.56^a ± 0.30	86.02^ab ± 0.15	0.800^g ± 0.000
	50	95.70^ab ± 0.04	5.19^ab ± 0.04	85.42^abc ± 0.04	0.820^e ± 0.000
	75	95.52^ab ± 0.11	4.91^ab ± 0.30	84.59^abc ± 0.15	0.839^d ± 0.001
	100	95.41^ab ± 0.15	4.72^ab ± 0.30	84.04^abc ± 0.52	0.893^b ± 0.001
75 : 25	25	95.78^a ± 0.09	5.41^a ± 0.18	85.80^ab ± 0.34	0.785^h ± 0.000
	50	95.73^a ± 0.20	5.37^a ± 0.36	85.54^abc ± 0.84	0.803^fg ± 0.003
	75	95.46^ab ± 0.34	4.79^ab ± 0.65	83.83^bc ± 2.05	0.803^fg ± 0.002
	100	95.40^ab ± 0.40	4.75^ab ± 0.74	84.40^bc ± 1.48	0.874^c ± 0.001

^a–hMeans ± SD of triplicate measurements with different superscripted letters in each column are significantly different (p ≤ .05).

The a_w values of all sugar-reduced desserts were significantly influenced by sorbitol and FOS levels (p ≤ .05). The level of sorbitol increased and the a_w decreased in sugar-reduced desserts, whereas this value increased with FOS when compared to the full-sugar desserts (Table 1). Because sorbitol has a low molecular weight and more effective hydroxyl groups, it can bind with water more effectively than sucrose, which increases the osmotic pressure and decreases the water activity.^[28–30] In contrast, FOS is a high molecular-weight substance that helps retain moisture, resulting in an enhanced water activity value in the dessert.^[16]

Physical properties

In general, the shape of sweet egg yolk drop is like a teardrop, as shown in Figure 1. The characteristics of this dessert are achieved by dropping egg yolk batter into boiling sugar syrup and submerging it until it is cooked. When the egg yolk batter droplets are immersed in sugar-water solutions, the starch and protein in the droplets absorb water molecules and expand. While sucrose molecules were absorbed (dependent on the concentration), they could limit the swelling of starch granules and inhibit the premature unfolding of protein molecules.^[31,32] The effects of sucrose on starch and protein may be caused by the increase in starch gelatinization temperatures and protein denaturation temperatures.^[33,34] As a result, egg yolk batter is able to expand and stabilize before being cooked in hot syrup (~70 ± 2 °Brix, 95 ± 2°C). This allows for the final product to have a proper size and shape. As a result, this study observed that sorbitol-to-fructose ratios and sugar reduction percentages in syrups, as well as interactions between these factors, influenced the properties of dessert samples (size, expansion ratio, and bulk density) (p ≤ .05). Therefore, reducing the amount of sugar in the syrup led to a significant increase in the size and expansion ratio, as well as a decrease in the bulk density of the dessert (Table 2).

Table 2. Physical properties of sweet egg yolk drops as affected by sorbitol with FOS ratio and percentage sugar reduction.

Experiments		Expansion ratio	Bulk density (g/cm^3)	Textures
Sorbitol : FOS	Sugar reduction (%)			Hardness (N)	Cohesiveness
Control		1.37^a ± 0.06	1.03^b ± 0.06	8.95^d ± 1.67	0.34^bcd ± 0.01
25 : 75	25	0.98^h ± 0.07	1.26^a ± 0.16	15.85^a ± 1.95	0.39^a ± 0.04
	50	1.06^g ± 0.07	1.12^ab ± 0.08	12.36^bc ± 1.68	0.38^ab ± 0.03
	75	1.17^de ± 0.05	1.12^ab ± 0.05	9.78^cd ± 2.10	0.33^bcd ± 0.03
	100	1.21^cd ± 0.08	1.05^f ± 0.11	5.46^abc ± 1.53	0.32^cd ± 0.04
50 : 50	25	1.79^gh ± 0.10	1.11^ab ± 0.12	13.11^b ± 1.99	0.34^abc ± 0.03
	50	1.81^fg ± 0.10	1.06^bc ± 0.09	10.22^cd ± 0.90	0.33^bcd ± 0.04
	75	1.91^cde ± 0.12	1.02^ef ± 0.05	5.78^abc ± 1.36	0.30^cd ± 0.05
	100	1.99^c ± 0.17	0.95^bc ± 0.07	3.85^fg ± 2.08	0.28^de ± 0.01
75 : 25	25	1.14^ef ± 0.09	1.05^bc ± 0.12	11.07^bcd ± 1.45	0.31^cd ± 0.03
	50	1.16^def ± 0.06	0.97^bc ± 0.08	8.29^de ± 1.66	0.29^cde ± 0.01
	75	1.30^b ± 0.10	0.97^bc ± 0.05	4.33^fg ± 1.62	0.25^ef ± 0.02
	100	1.36^ab ± 0.06	0.89^c ± 0.07	2.46^g ± 0.28	0.21^f ± 0.01

^a–hMeans ± SD of triplicate measurements with different superscripted letters in each column are significantly different (p ≤ .05).

Meanwhile, the substitution of sugar with sorbitol and FOS in syrup resulted in the significant changes in the physical properties of sugar-reduced desserts, as shown in Table 2. It was described that both substances differ in their physicochemical properties, particularly their molecular weights (MW) and the number of effective hydroxyl groups (N_OH,s/ν_s). The molecular weight of FOS (MW: 605, N_OH,s/ν_s: 0.0119) is greater and the number of effective hydroxyl groups is lower than those of sorbitol (MW: 182, N_OH,s/ν_s: 0.0272).^[35,36] As a result, FOS has less diffusion into egg yolk batter, but it still has the ability to form intermolecular hydrogen bonds with water in the outer dessert droplets. Consequently, the restricted water accessibility by FOS to starch-protein mixtures led to a delay in starch gelatinization and protein denaturation temperature.^[16] These reasons contribute to the limited swelling of starch/unfolding of proteins, which inhibits batter droplets from expansion before the outer layers of desserts gradually form and set. Therefore, the increased level of FOS in syrup led to a reduction in expansion ratio and an increase in the bulk density of desserts, whereas the opposite results were observed in the case of sorbitol.

Texture properties

As shown in table 2, hardness and cohesiveness values were used to describe the texture of sweet egg yolk desserts. There was a significant effect of changing the sorbitol/FOS ratio and percentage of sugar reduction in syrup on the hardness values (2.46–15.85 N, p ≤ .05) and cohesiveness values (0.21–0.39, p ≤ .05) of the desserts. It was also found that there was a significant interaction between the two factors in this study. The high levels of sorbitol combined with low sugar levels in the syrup resulted in a softer texture than other samples, especially the control. The high levels of FOS combined with the high levels of sugar in the syrup increased the hardness of the dessert. It is possible that these results are due to the changes in dessert size and shape caused by variations in bulk density and expansion ratio. According to Gok, et al.^[37] both sucrose and soluble wheat fiber increased the hardness of low-calorie gummy candies. In addition, according to Renuka et al.,^[38] it was reported that gulab jamuns soaked in FOS syrup are harder than those made with sucrose, which have some production processes similar to those of sweet egg yolk drops.

Chemical properties

In Table 3, the percentage of sugar reduction and the ratio of sorbitol/FOS significantly affected the chemical properties of dessert samples (p ≤ .05) and both variables interacted. As a soluble fiber, FOS has been added to the syrup for technical reasons in order to create the characteristics of a low sugar sweet egg yolk drop, which are not generally present in this dessert. As shown in Table 3, increasing the FOS ratio in the syrup expectedly led to significant increases in the soluble fiber content of desserts (p ≤ .05).

Table 3. Chemical properties of sweet egg yolk drops as affected by sorbitol with FOS ratio and percentage sugar reduction.

Experiments
Sorbitol : FOS	Sugar reduction (%)	Moisture content (% wb)	Soluble Fiber (g/100 g)	Reducing sugar (as invert sugar g/100 g)	Sucrose (g/100 g)	Energy (kcal)
Control		30.28^d ± 0.03	ND	0.68^b ± 0.03	43.10^a ± 0.56	351.28^a ± 1.07
25 : 75	25	23.18^h ± 0.06	0.19^a ± 0.01	0.73^a ± 0.04	31.95^c ± 0.64	331.76^b ± 0.92
	50	28.82^f ± 0.08	0.15^b ± 0.01	0.59^b ± 0.01	25.10^e ± 0.57	322.31^c ± 1.89
	75	29.55^def ± 0.31	0.15^b ± 0.01	0.54^b ± 0.02	16.80^g ± 0.56	321.31^c ± 2.02
	100	37.25^b ± 0.21	0.13^c ± 0.01	0.53^c ± 0.01	1.17^h ± 0.07	288.86^f ± 1.52
50 : 50	25	25.72^g ± 0.19	0.16^b ± 0.01	0.68^b ± 0.01	33.30^b ± 0.28	329.35^b ± 2.37
	50	29.13^ef ± 0.00	0.15^b ± 0.01	0.52^a ± 0.01	24.65^e ± 0.21	322.62^c ± 0.75
	75	31.14^c ± 0.36	0.13^c ± 0.01	0.51^b ± 0.01	17.95^f ± 0.21	301.92^e ± 0.13
	100	40.57^a ± 0.63	0.09^d ± 0.00	0.45^b ± 0.01	0.88^h ± 0.22	270.77^g ± 2.72
75 : 25	25	25.42^g ± 0.42	0.12^c ± 0.00	0.53^c ± 0.04	32.40^c ± 0.14	328.82^b ± 1.14
	50	29.72^de ± 0.06	0.10^d ± 0.00	0.49^b ± 0.01	26.65^d ± 0.21	314.24^h ± 0.64
	75	29.09^ef ± 0.19	0.09^d ± 0.01	0.47^c ± 0.01	17.35^fg ± 0.22	315.75^d ± 0.93
	100	40.40^a ± 0.55	0.09^d ± 0.00	0.42^b ± 0.01	0.71^h ± 0.01	268.56^g ± 1.31

^a–hMeans ± SD of triplicate measurements with different superscripted letters in each column are significantly different (p ≤ .05).

For the reducing sugar value, when the percentage of sugar reduction was increased from 25% to 100% and the sorbitol proportion was increased from 25 to 75 in syrup, the reducing sugar content of dessert decreased (Table 3). According to its properties, sorbitol is a non-reducing sugar like sucrose, which does not have exposed carbonyl groups that can react with amino acids to promote browning reactions.^[25,39] However, the dessert produced in the above conditions may contain a small amount of reducing sugar, as sucrose can also be hydrolyzed under thermal processing conditions into monosaccharides (glucose and fructose) which are characteristic properties of reducing sugar.^[40] While the FOS proportion in syrup was increased and the sugar percentage decreased, resulting in an increase in reducing sugar content in desserts, especially for the 25:75 ratio of sorbitol to FOS and the 25% of sugar reduction (0.73 g/100 g). It is possible that a high temperature (90–100°C) and a low pH (2.7–3.3) of food processing conditions led to the breakdown of oligosaccharides like FOS into monomers (glucose and fructose) and dimers (sucrose).^[41,42]

The desserts with sorbitol, FOS, and sugar content in the syrup had a decreased energy value in comparison to the control sample (Table 3). The decrease in energy can be attributed to a reduction in sucrose, which provides approximately 4 kcal/g of general energy, while the addition of sorbitol and FOS as sweeteners provides a dietary energy of about 2.6 kcal/g and 0–3 kcal/g, respectively.^[43–45]

Sensory properties

In replacing sugar with sweeteners in desserts, consumers expect desserts with sugar reduction to have the same pleasing taste and appearance as desserts with full sugar, especially for sweet tastes. This study compared the sensory evaluations of sugar-reduced sweet egg yolk drops and those of desserts that contained full sugar content, as shown in Table 4. It was found that there was a significant decrease in all sensory attributes that were influenced by the reduction of sugar content, especially appearances, sweetness, juiciness, and firmness of the desserts. The addition of sorbitol and FOS in different ratios can improve the liking scores of desserts that have been reduced in sugar content. Based on these data, analysis showed that both main variables (sorbitol/FOS ratio and sugar reduction (%)) and interaction between them affected all sensory attributes significantly (p ≤ .05).

Table 4. Sensory properties of sweet egg yolk drops as affected by sorbitol with FOS ratio and percentage sugar reduction.

Experiments		Appearances	Yellowish-orange	Overall flavor	Sweetness	Juiciness	Firmness	Overall liking
Sorbitol : FOS	Sugar reduction (%)
Control		6.5^abc ± 1.0	6.8^abc ± 1.2	6.2^bc ± 1.6	6.5^abc ± 1.3	6.5^ab ± 1.7	6.1^abcd ± 1.8	6.6^ab ± 1.3
25 : 75	25	6.0^cde ± 1.7	6.2^bcd ± 1.6	6.4^abc ± 1.6	6.0^bcd ± 1.5	5.7^bc ± 1.7	6.1^abcd ± 1.8	6.0^bc ± 1.5
	50	7.3^a ± 1.0	7.3^a ± 1.1	7.2^a ± 1.1	7.0^a ± 1.1	6.6^ab ± 1.5	6.8^b ± 1.6	7.2^a ± 0.9
	75	7.1^a ± 1.0	7.1^a ± 1.0	6.9^ab ± 0.9	6.8^ab ± 1.7	6.7^ab ± 0.9	7.0^a ± 1.2	6.6^ab ± 1.0
	100	5.5^def ± 1.8	6.2^bcd ± 1.5	6.5^abc ± 1.8	5.6^cde ± 2.0	5.6^bc ± 2.3	5.2^de ± 1.8	5.3^cd ± 1.8
50 : 50	25	6.5^abc ± 1.4	7.4^a ± 1.2	6.8^ab ± 1.6	6.0^bcd ± 1.4	6.6^ab ± 1.7	6.4^b ± 1.8	6.0^b ± 1.2
	50	7.0^a ± 1.2	7.0^ab ± 1.3	6.2^bc ± 1.4	6.2^abcd ± 1.5	6.4 ^ab ±1.5	6.0^abcd ± 2.2	6.5^ab ± 1.3
	75	7.2^a ± 1.0	6.1^bcd ± 1.1	6.0^bc ± 1.3	6.0^bcd ± 1.1	6.3^ab ± 1.5	6.2^abcd ± 1.2	7.0^a ± 0.7
	100	5.2^ef ± 1.4	5.9^cd ± 2.0	6.0^bc ± 1.7	4.5^f ± 1.2	6.0^b ± 2.0	5.8^abc ± 2.0	5.1^de ± 1.5
75 : 25	25	6.1^bcd ± 0.7	7.2^a ± 1.4	6.8^ab ± 1.5	6.2^abcd ± 1.6	7.2^a ± 1.4	6.1^abcd ± 1.0	6.0^b ± 1.0
	50	6.7^abc ± 1.1	6.7^abc ± 1.5	6.2^bc ± 1.6	6.2^abcd ± 1.2	6.5^ab ± 1.2	6.4^abc ± 1.5	6.7^ab ± 0.9
	75	6.8^ab ± 0.9	6.7^abc ± 1.9	6.8^ab ± 1.2	6.0^bcd ± 1.7	6.6^ab ± 1.9	5.5^cd ± 2.0	6.2^d ± 0.7
	100	5.0^f ± 1.8	5.4^d ± 1.7	5.7^c ± 1.8	5.0^ef ± 2.1	4.9^c ± 2.2	4.2^e ± 2.0	4.5^e ± 1.7

^a–hMeans±SD of triplicate measurements with different superscripted letters in each column are significantly different (p≤ .05).

In terms of taste and texture, the addition of sorbitol significantly maintained the liking scores for the sweetness and juiciness attributes of desserts until reducing the percentage of sugar in the syrup to 75%. A high FOS ratio did improve the liking scores for the firmness of desserts at reduced sugar levels between 25% and 75%. However, the overall liking of sugar-reduced desserts (50% and 75%) with sorbitol to FOS ratios (25:75 and 50:50) resulted in panelists giving the average liking scores between 7.0 and 7.2 (like moderately) as opposed to control desserts (6.6: like slightly), as shown in Table 4.

Optimization and verification

In order to select the optimal formulation for the healthy, sweet egg yolk desserts, the RSM was used to determine the optimal ratios of sorbitol and FOS and the reduction in sugar content in syrup. A quadratic polynomial model was used to investigate the relationship between two independent variables: the ratios of sorbitol to FOS (X₁) and the reduction of sugar content (X₂) in percentage with six dependent variables: physical properties (expansion ratios (Y₁) and textures (Y₂)), chemical properties (energy values (Y₃)) and sensory properties (appearance (Y₄), sweetness (Y₅), and overall liking (Y₆), as shown in Table 5. In all predictive models (Table 5), the R² values and lack of fit tests were higher than 0.70 and not significant (p > .05), indicating a good fit of the model and the ability of the independent variables to adequately explain the variation of the dependent variables. Therefore, the obtained model equation can be used to predict the optimal formulation for the healthy, sweet egg yolk drops.

Table 5. Predictive regression models for response variables of sweet egg yolk drops.

Dependent variables (Y)		Predictive models	R^2	Lack of fit (p-value)
Physical properties:
	Expansion ratio (Y1)	−1.237 + 0.115X1 +0.002X2-0.001X1^2-2.984E-18X1X2 +0.000008X2^2	0.98	>0.05
	Hardness (Y2)	23.808–0.180X1-0.161X2 +0.001X1^2 +0.000X1X2 +0.000X2^2	0.97	>0.05
Chemical property:
	Energy value (Y3)	333.907–0.782X1 +0.858X2 +0.008X1^2-0.004X1X2-0.011X2^2	0.93	>0.05
Sensory properties:
	Appearances (Y4)	3.2417 + 0.027X1 +0.1245X2-0.0003X1^2-0.0001X1X2-0.001X2^2	0.86	>0.05
	Sweetness (Y5)	6.1583–0.088X1 +0.0875X2 +0.0009X1^2-0.0002X1X2-0.0007X2^2	0.86	>0.05
	Overall liking (Y6)	3.4333 + 0.017X1 +0.1177X2-0.0001X1^2-0.0002X1X2-0.001 X2^2	0.78	>0.05

Independent variable: X1 = sorbitol (%), X2 = percentage of sugar reduction in syrup (%).

R2 = correlation coefficient of determination.

Second-order polynomial regression model: $Y={\beta }_0+{\beta }_i{X}_i+{\beta }_j{X}_j+{\beta }_{ii}{X}_i^2+{\beta }_{jj}{X}_j^2+{\beta }_{ij}{X}_i{X}_j$

Sugar reduction in processed foods has a significant impact on food characteristics and sensory preferences.^[46] It has also been reported that taste and flavor perception are related to the structure and texture of food, including food breakdown during oral processing.^[47] Thus, a superimposed optimal area for the development of healthy desserts was determined from six regions of response contour plots (expansion ratio, texture, energy values, appearance, sweetness, and overall liking) based on the sorbitol to FOS ratios and sugar reduction percentages, as shown in Figure 2(a-e). The critical boundaries of the superimposed regions were determined by the three hedonic scores of appearances, sweetness, and overall liking, each above 7.0, 6.5, and 7.0 (like slightly to like moderately), respectively (Figure 2(a). Moreover, the sugar content in syrup should be reduced to at least 50% to produce the healthier desserts. The results of the superimposed regions indicated that the optimal formulations had the sorbitol/FOS ratios between 25:75 and 35:65, with the reduction in sugar content in syrup ranging from 52% to 65% (Figure 3). This optimal formulation provided the characteristics of desserts that had an expansion ratio of 1.1–1.7, a texture of 9.4–12.0 N, and an energy level of 316.5–330.0 kcal, including the appearance scores of 7.4–7.5, the sweetness scores of 6.5–7.0, and the overall liking scores of 6.9–7.0.

Figure 2. RSM contour plots of expansion ratio (a), hardness (b), energy values (c), and sensory liking score of appearances (d), sweetness, (e) and overall liking (f).

Figure 3. The optimal condition in the shaded area for development of sweet egg yolk drop.

For the verification of the results, experimental and predictive values were compared to verify that the formulation selected was within the optimal range (Table 6). The experimental values were obtained from three formulations using various ratios of sorbitol to FOS accompanied with the various percentages of sugar reduction as follows: 25:75 with 60%, 25:75 with 65%, and 30:70 with 65%, respectively. Additionally, these three formulations were used to create the predicted values by substituting them into the generated model equations. The three syrup formulations were selected for the production of sweet egg yolk desserts because these syrup formulations reduced the sugar content as much as possible, while still producing the desserts that maintained a high level of sensory acceptance. The suitability of the generated models was determined by root mean squared error (RMSE) values. According to Table 6, the RMSE values of all model equations were relatively low. The RMSE value close to 0 indicates that the model predictions are in substantial agreement with the experimental data, indicating a well-fitted model.^[48,49] Thus, the regression model could be used to accurately predict the expansion ratios, textures, energy values, and sensory attributes (appearance, sweetness, and overall liking) of the desserts.

Table 6. Predicted and experimental values of the sweet egg yolk drop under selected conditions for verification of optimized region.

Properties			Formulation of syrup		Predicted values^a	Experimental values	RMSE^b
			Sorbitol : FOS (%)	Sugar reduction (%)
Expansion ratio			25 : 75	60	1.10	1.55	0.41
			25 : 75	65	1.11	1.65
			30 : 70	65	1.41	1.32
Hardness (N)			25 : 75	60	10.27	9.65	0.86
			25 : 75	65	9.47	8.17
			30 : 70	65	8.84	9.21
Energy values (kcal)			25 : 75	60	325.24	328.04	1.78
			25 : 75	65	322.15	321.44
			30 : 70	65	319.14	320.23
Appearances			25 : 75	60	7.4	7.5	0.18
			25 : 75	65	7.4	7.6
			30 : 70	65	7.5	7.2
Sweetness			25 : 75	60	7.0	6.8	0.17
			25 : 75	65	6.9	7.0
			30 : 70	65	6.7	6.9
Overall liking			25 : 75	60	7.0	7.2	0.28
			25 : 75	65	6.9	7.3
			30 : 70	65	6.9	7.0

^aValues derived from the predictive regression model (Table 5).

^bRMSE = Root mean squared error between predicted and experimental values.

Conclusion

This study demonstrated that sorbitol and FOS could be used as sugar substitutes for the preparation of sweet egg yolk drops that are comparable in quality and acceptability to those prepared with regular sugar syrups. Increasing the proportion of sorbitol in syrup (25% to 75%) and increasing the percentage of sugar reduction (from 25% to 100%), the properties of desserts, including a* and b* values, bulk density, hardness, cohesiveness, soluble fiber, and reducing sugar decreased, but the a_w, water content, and expansion ratio increased. In the same experiment, desserts with a reduction in sugar content and an increased FOS content produced the opposite results. The RSM was successfully used for the development of an optimal syrup formulation for the production of the healthy, sweet egg yolk drops. It is also possible to predict the quality of the desserts based on the model equations generated. The syrup formulations could be optimized by mixing sorbitol and FOS in a ratio of 25:75 and 30:70, as well as reducing sugar from 60 to 65%. Due to this condition, the developed desserts have lower energy values than 321 kcal (a reduction of approximately 25%) as well as higher sensory liking scores than the full-sugar sweet egg yolk desserts (appearance, sweetness, and overall liking > 6.8). Moreover, a formula randomly selected from the optimized area (sorbitol/FOS: 25:75 and sugar reduction at 65%) resulted in a developed dessert containing FOS of 17.44 g/100 g sample and a glycemic index (GI) of 71.62, in comparison to the sweet egg yolk desserts containing sugar (not detected and 78.14, respectively). Lower GI values indicate a slower rate of digestion and absorption of glucose into the bloodstream. From this work, a healthy, sweet egg yolk drop with positive properties was developed, which will be important for the dessert industries.

Author contributions

Phanlert Promsakha na Sakon Nakhon: Conceived and designed the experiments; Funding acquisition; Performed the experiments; Analyzed and interpreted the data; Wrote the paper. Montakan Aimkaew: Performed the experiments; Analyzed and interpreted the data.

Wannarat Leesuksawat: Conceived and designed the experiments; Performed the experiments.

Saynamphung Tongsai: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Revising manuscript.

Acknowledgments

The authors would like to thank the Department of Food Technology, Faculty of Science, Ramkhamhaeng University, for the support. We greatly appreciate the assistance with our experiments from Mr. Chanakun Srikongphan, Ms. Natthida Mahachairachan, Mr. Pongprasit Suwannasaksin, and Ms. Ploypailin Tulaphol.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data included in article/supplementary material/referenced in article..

Additional information

Funding

This research was financially supported by the Research and Development Institute, Ramkhamhaeng University.