Production trials to improve the nutritional quality of biscuits and to enrich them with natural anthocyanins

Abstract

The aim of the work was to improve the nutritional quality of biscuits and to enrich them with anthocyanins. The steps were (i) evaluation of various raw materials as sources of anthocyanins and optimization of anthocyanin extraction procedure; (ii) conduction of six biscuit-production trials by using anthocyanins extracted from the selected source, at different enrichment levels and with progressively less fat and sugar; and (iii) conduction of a paired-preference test to compare the best anthocyanin-enriched biscuits to control plain biscuits. Grape marc extract was selected for biscuit making. Also, the use of durum wheat semolina and extra virgin olive oil was proposed. A fat level as low as 100 g/kg was achieved, with a notable decrease of saturated fatty acids. Total anthocyanins accounted for 103 × 10⁻³ g/kg. The color of biscuits was characterized by more intense redness and minor lightness compared to the control, and acceptability was adequate, as assessed by 103 consumers.

El objetivo de este trabajo fue mejorar la calidad nutricional de las galletas y enriquecerlas con antocianinas. Para ello se llevaron a cabo los siguientes pasos: (i) evaluación de varias materias primas como posibles fuentes de antocianinas para seleccionar una de ellas y optimización del procedimiento de extracción de antocianinas; (ii) realización de ensayos para la producción de seis galletas usando antocianinas extraídas de la fuente seleccionada, modificando sus niveles de enriquecimiento con cada vez menos grasa y azúcar; (iii) ejecución de una prueba de preferencia entre dos muestras de galletas, a fin de comparar las mejores galletas enriquecidas a base de antocianinas con las galletas sin enriquecer utilizadas como control. Para hacer las galletas, entre las materias primas evaluadas como fuentes de antocianinas se seleccionó el extracto de orujos de uva, empleándose, además, sémola de trigo duro y aceite de oliva extra virgen. Las galletas obtenidas presentaron un nivel de grasas mínimo, de 100 g/kg, constatándose una reducción notable de los ácidos grasos saturados. El total de antocianinas fue de 103 × 10⁻³ g/kg. Asimismo, mostraron un color rojizo de carácter más intenso y menor ligereza que las utilizadas como control, logrando una aceptabilidad adecuada, según la opinión de 103 consumidores.

Keywords:

• anthocyanins
• biscuits
• nutritional quality
• durum wheat semolina
• extra virgin olive oil

Palabras clave:

• antocianinas
• galletas
• calidad nutricional
• sémola de trigo duro
• aceite de oliva extra virgen

Introduction

Many foods contain non-nutritive components that may provide protection against chronic diseases through multiple effects. Anthocyanins are flavonoid compounds responsible for the coloration of flowers, fruits, and vegetables. They exhibit many health-related properties such as anti-inflammatory activity, antioxidant activity on human low-density lipoproteins (Satué-Gracia, Heinonen, & Frankel, 1997), and have positive effects on micro-circulation diseases and eye function (Ghosh & Konishi, 2007).

In recent years, the interest of consumers toward the health implications of food has progressively increased, and many functional foods have been released. The most consumed foodstuffs, such as dairy products, as well as pasta and bakery products, have been functionalized by enriching them with a wide array of antioxidants and vitamins.

Moreover, various ingredients have been evaluated as sources of active healthy components to prepare functional biscuits: defatted mustard meal, flaxseed meal, flaxseed oil (Hassan, Rasmy, Foda, & Bahgaat, 2012), and black cumin fixed oil (Sultan et al., 2011). Other studies proposed the use of plants with antioxidant properties to replace synthetic antioxidants (Bajaj, Urooj, & Prabhasankar, 2006; Bassiouny, Hassanien, Ali, Sohair, & El-Kayati, 1990; Nanditha, Jena, & Prabhasankar, 2009; Reddy, Urooj, & Kumar, 2005). Pure antioxidant compounds (ascorbyl-2-phosphate, α-tocopherol, and their combinations) were tested as replacements for butylated hydroxyanisole (Hix, Klopfenstein, & Walker, 1997). Other researchers, mainly to increase fiber content, proposed the enrichment of soft dough biscuits with several by-products, such as mango peel (Ajila, Leelavathi, & Prasada Rao, 2008), potato peel (Abdel-Magied, 1991), finger millet seed coat (Krishnan, Dharmaraj, Sai Manohar, & Malleshi, 2011), or cereal bran (Sudha, Vetrimani, & Leelavathi, 2007). To our knowledge, up to now, no studies have been aimed at enriching biscuits with anthocyanins.

Biscuits are daily present in the diet, and their formulation could positively contribute to the overall intake of nutrients. Although not perceived as fatty foods by consumers, they contain a consistent quantity of lipids. Amounts of fat ranging from 75 to 230 g/kg, mainly constituted by refined vegetable oils and margarine, have been determined in Italian biscuits (Caponio, Summo, Delcuratolo, & Pasqualone, 2006). Lipids have a significant effect on the rheological characteristics of dough and induce easier breaking and friability of biscuits (Chevallier, Colonna, Della Valle, & Lourdin, 2000). The choice of lipid to be used in biscuit production is mainly done on the basis of technological and economic parameters, without considering the nutritional implications. In fact, in some cases, high amounts of trans isomers of unsaturated fatty acids have been reported (Caponio et al., 2006). Similar problems affect gluten-free biscuits, with a fatty fraction present in mean amounts of 150 g/kg, where high contents of compounds derived from lipid oxidation and trans isomers of fatty acids have been detected (Caponio, Summo, Clodoveo, & Pasqualone, 2008a).

Attempts to improve nutritional characteristics of biscuits were directed at reducing fats, sugars, and energy level (Sultan et al., 2011). Taylor, Fasina, and Bell (2008) proposed the replacement of sucrose with tagatose, a minimally absorbed prebiotic monosaccharide. However, the substitution of fats with extra virgin olive oil, known for its positive healthy features, has not been considered in any report. Moreover, the use of durum wheat semolina, richer in carotenoid compounds than common wheat flour (Abdel-Aal, Young, Rabalski, Hucl, & Fregeau-Reid, 2007), could further improve the nutritional profile of the final product. Additionally, this has not been considered by previous research works.

The aim of the present work was to improve the nutritional quality of biscuits by using extra virgin olive oil and durum wheat semolina and to enrich them with anthocyanins. In particular, the research was structured in three subsequent steps: (i) evaluation of different raw materials as sources of anthocyanins, and optimization of anthocyanin extraction procedure; (ii) conduction of different biscuit-production trials by using anthocyanins extracted from the selected source to investigate how the proposed changes of a plain biscuit recipe affect the nutritional quality and to point out the best final product; and (iii) conduction of a paired-preference test to compare the anthocyanin-enriched biscuits with best nutritional features to control plain biscuits.

Materials and methods

Anthocyanin sources

Commercial samples of red and black berries mix (red and black currants, blackberries, raspberries, and bilberries), sour cherries, grape (table grape cv. Black magic, whole berry or skin, the latter obtained by manual peeling), and pressed grape marc (from wine grape cv. Sangiovese) obtained from a local wine-making factory were collected for anthocyanin extraction.

Extraction and determination of total anthocyanins

Samples were homogenized (Omni Mixer 17108; Sorvall Omni International, Kennesaw, GA, USA), at a solid:liquid ratio 1:10, with a solution of absolute ethanol acidified by citric acid (85:15 absolute ethanol:1 mol L⁻¹ citric acid, pH 4). Grape marc and grape skins were submitted to extraction both fresh and after lyophilization (Heto Power Dry 9000 lyophilizer; Heto-Holten, Allerød, Denmark). Homogenates were submitted to orbital shaking at 250 rpm, at room temperature (20°C), in the dark, for different times (30 min, 1 h, and 2 h). After centrifugation at 2000 × g for 10 min, the supernatant was stored in amber flasks until analysis. Each extraction was carried out in triplicate. The total monomeric anthocyanin content was spectrophotometrically quantified at 535 nm after setting up a calibration curve with cyanidin chloride (CyCl) (PhytoLab, Vestenbergsgreuth, Germany) as standard, according to Lohachoompol, Mulholland, Srzednicki, and Craske (2008).

Sample cooking and preparation

Commercial samples of white soft wheat flour (from Triticum aestivum L.), durum wheat semolina (from Triticum turgidum L., var. durum), sucrose, butter, and extra-virgin olive oil were purchased at local retailers for the biscuit-making trials. Soft wheat flour had the following composition: proteins 85.0 g/kg, carbohydrates 780.3 g/kg, and moisture 125.0 g/kg, while the composition of durum wheat semolina was: proteins 110.0 g/kg, carbohydrates 755.0 g/kg, and moisture 125.0 g/kg. Flour and semolina compositional data were obtained from the producer, based on the production lot. Six trials of production of biscuits were carried out in laboratory plus a control trial, according to the formulations reported in Table 1, and each production trial was repeated in triplicate. Twenty biscuits were obtained from each trial, sufficient for subsequent compositional analyses. Trial 6 and control-making trial were scaled up and repeated to obtain 150 biscuits each, in order to perform also a paired-preference test. A total amount of about 750 mL of anthocyanin extract, having the concentration of 353.70 mg/L CyCl, was needed for all the enrichment trials. The extract was obtained starting from 300 g lyophilized grape marc; the extraction duration was 1 h. The other conditions were as described above, apart a final vacuum evaporation to a quarter of the initial volume. Control biscuits were produced with the same composition of enriched biscuits but without anthocyanin extract. The biscuit-making process, according to the technology of shortbreads, consisted of mixing sugar, fat, and flour by a spiral kneader for 5 min (“creaming” step, to reduce subsequent gluten formation), adding water and/or anthocyanin extract, mixing for 10 min, then rolling dough, shaping (round biscuits, 4.5 cm diameter, 4 mm thickness), and baking in an electric oven (Smeg SI 850 RA-5 oven; Smeg S.p.A., Guastalla, Italy) for 15 min at 150 °C.

Table 1. Formulation of biscuits derived from different production trials.
Tabla 1. Formulación de galletas derivada de diversas pruebas de producción.

Ingredient	Trial
	1	2	3	4	5	6	Control
Soft wheat flour (g)	100	–	–	–	–	–	–
Durum wheat semolina (g)	–	100	100	100	100	100	100
Sucrose (g)	40	40	40	35	35	35	35
Butter (g)	28	28	28	16	–	–	–
Extra virgin olive oil (mL)	–	–	–	–	16	16	16
Anthocyanin extract (mL)	20	20	28	28	28	45	–
Water (mL)	8	8	–	–	–	10	25

Analyses of biscuits

Compositional analyses were performed for all experimental samples (biscuits derived from trials 1–6). A paired-preference test was performed only for samples from trial 6 versus control biscuits. Moisture content was assessed at 105°C by automatic moisture analyzer (Mod. MAC 110/NP; Radwag Wagi Elektroniczne, Radom, Poland). Protein content (N × 5.7) was determined according to the AACC approved method 46-11A (AACC, 2000). Fat content was assessed by acid hydrolysis and subsequent extraction of hydrolyzed lipid materials according to AACC method 30-10 (AACC, 2000). Fatty acid composition was determined by gas chromatographic analysis of fatty acid methyl esters according to AOCS method Ch 1–91 (AOCS, 1993), using C_15:0 as internal standard. The conditions of the analysis were as described in a previous work (Caponio, Summo, Pasqualone, & Bilancia, 2008b). The determination of carotenoid pigment concentration was made according to AACC approved method 14–50.01 (AACC, 2000) with slight modifications: 1 g of each sample, ground in a mortar, was extracted with 5 mL of water-saturated n-butyl alcohol on an orbital shaker for 3 h at 260 rpm. Samples were centrifuged for 7 min at 2400 × g, and spectrophotometrical absorbance of water-saturated n-butyl alcohol extracts was measured at 435.8 nm. Total carotenoid content was expressed as β-carotene, and calculations were made based on the extinction coefficient of 1.6632 for a solution of 1 mg β-carotene in 100 mL water-saturated n-butyl alcohol. Each analysis was performed in triplicate. Color analyses were carried out by using a spectro-colorimeter CM-700d (Konica Minolta Sensing, Osaka, Japan) equipped with a pulsed xenon lamp. The following coordinates in the L*a*b* color space, derived from the Commission Internationale de l'Eclairage (CIE, 1986) XYZ tristimulus values, were determined under D65 illuminant: b* (yellow/blue balance), a* (red/green balance), and L* (lightness), the latter used to calculate brown index (BI) as 100 complement of the L* index (BI = 100 − L*) (Feillet, Autran, & Icard-Vernière, 2000). Besides, color saturation or chroma (C*) and hue (h°) were calculated as follows: C* = [(a*)² + (b*)²]^1/2; h° = arctan (b*/a*). Color measurements were performed considering three different points of the central surface of three different biscuits. A paired-preference test (trial 6 versus control) was carried out to determine the acceptability of biscuits and the willingness of consumers to purchase. The test involved 103 voluntary, untrained, biscuit consumers aged 14–70 years (44 male and 59 females). Each consumer received the samples (one control and one anthocyanin-enriched biscuit from trial 6) coded A and B, as well as the evaluation sheet, reported in Figure 1, and a pencil.

Figure 1. Evaluation sheet for the assessment of acceptability and willingness to buy.

Figura 1. Hoja de evaluación para valorar la aceptabilidad y la disposición de compra.

Statistical analyses

Data were submitted to statistical analysis by XLStat software (Addinsoft SARL, New York, NY, USA). One-way analysis variance (ANOVA) was performed, followed by the Tukey HSD test for multiple comparisons. Friedman test was performed to assess significant differences between the number of responses given by consumers involved in paired-preference test. Pearson's simple correlation coefficients were calculated between total anthocyanin content, total carotenoid content, and colorimetric indices.

Results and discussion

Evaluation of different raw materials as sources of anthocyanins and optimization of extraction procedure

In order to obtain an edible, non-toxic, anthocyanin extract suitable to be used for food enrichment, the combination of ethyl alcohol and citric acid was proposed in our trials as extracting medium. The anthocyanins are soluble in polar solvents, and they are commonly extracted in acidic conditions to prevent the degradation of the non-acylated anthocyanin pigments. Established anthocyanin extraction methods involve maceration or soaking the plant material in a low boiling point polar solvent containing small amounts of mineral acid, to keep the anthocyanins in their stable red flavylium cation form. However, these methods are not intended to an edible use of the extract due to the solvent used (Bechtold, Mahmud-Ali, & Mussak, 2007; Burin, Rossa, Ferreira-Lima, Hillmann, & Boirdignon-Luiz, 2011; Garcia-Viguera, Zafrilla, & Tomàs-Barberàn, 1998; Gòmez-Plaza, et al., 2008; Liazid, Guerrero, Cantos, Palma, & Barroso, 2011; Manfra et al., 2011; Manhita, Teixeira, & da Costa, 2006; Núñez, Monagas, Gomez-Cordovés, & Bartolomé, 2004).

Various raw materials were considered to set up and optimize the extraction procedure, i.e. red and black fruits, grape and grape marc, known to be potentially suitable for anthocyanin extraction (Bechtold et al., 2007; Burin et al., 2011; Kähkönen, Heinämäki, Ollilainen, & Heinonen, 2003; Liazid et al., 2011; Viskelis, Bobinaite, Rubinskiene, Sasnauskas, & Lanauskas, 2012). Assuming a certain variability due to both cultivar and extraction conditions, the obtained results (Table 2) were in the range of literature data (Bechtold et al., 2007; Burin et al., 2011; Kähkönen et al., 2003; Liazid et al., 2011; Manfra et al., 2011; Núñez et al., 2004; Viskelis et al., 2012), with the highest levels, as expected, in red and black berries. Major interest was put toward grape and grape marc because of their large availability in the area where the research was carried out. In particular, the use of grape marc, an abundant and low-cost by-product of wine making, has been recently proposed for the extraction of antioxidant phenolic compounds (Fiori, De Faveri, Casazza, & Perego, 2009; Louli, Ragoussis, & Magoulas, 2004; Pinelo et al., 2007). Higher levels of anthocyanins were observed in grape skin than in whole berries (table grape, cv. Black magic), due to peripheral localization of anthocyanins. Lower amounts were extracted from marc of wine grape, cv. Sangiovese, mainly due to anthocyanin transfer from berries to future wine during the maceration phase. In order to increase the anthocyanin concentration of final extract, both grape skin and grape marc were subjected to lyophilization. In both the cases, this treatment markedly increased the level of recovered anthocyanins respect to the fresh source. With the subsequent aim of achieving significant enrichments of biscuits, at reasonably low volumes of anthocyanin extracts, lyophilization was considered essential. Besides, different extracting times were tested. The obtained data indicate that the majority of anthocyanins were extracted within 30 min, in agreement with other authors (Bechtold et al., 2007). Prolonging the extraction time to 1 h led to a significant increase of anthocyanin content (P < 0.001), while at 2 h were observed contrasting results. Hence, to prepare the anthocyanin extract for the subsequent biscuit-production trials, lyophilized grape marc was selected among the different sources tested, and the extraction was prolonged for 1 h.

Table 2. Total extractable anthocyanins (expressed as g CyCl × 10⁻³/kg sample, mean ± SD) of different vegetable sources at different extraction times.
Tabla 2. Total de antocianinas extraídas (expresado como CyCl × 10^{− 3}/kg muestra, media ± DS) de diferentes fuentes vegetales usando diferentes tiempos de extracción.

Sample	Extraction time
	30 min	1 h	2 h
Red and black berries commercial mix (red and black currants, blackberries, raspberries, and bilberries)	700.77^a ± 27.87	1000.46^b ± 20.33	984.24^b ± 12.23
Dried sour cherries	42.31^a ± 1.43	109.61^b ± 2.81	108.63^b ± 4.23
Grape berry^1 (fresh weight)	284.05^a ± 17.10	360.26^b ± 18.40	348.52^b ± 3.70
Grape skin^1 (fresh weight)	307.94^a ± 4.49	619.56^b ± 8.15	734.39^c ± 14.20
Grape skin^1 (lyophilized)	785.18^a ± 13.74	1099.86^b ± 47.60	1056.47^b ± 20.46
Grape marc^2 (fresh weight)	138.92^a ± 6.36	349.48^b ± 10.89	548.94^c ± 14.93
Grape marc^2 (lyophilized)	488.37^a ± 2.11	875.86^b ± 6.99	955.40^c ± 11.55
Note: ^1Cultivar black magic; ^2Cultivar Sangiovese. Different letters in row indicate significant differences at P < 0.001 based on Tukey HSD test.

Biscuit-production trials and effect of the proposed changes on the nutritional quality

Six different biscuit-making trials were carried out, according to formulations reported in Table 1. The starting point (trial 1) was the recipe of a plain biscuit (without chocolate or other ingredients potentially interfering with the results) with a fat content in the range of the commonly marketed types (Caponio et al., 2006). Besides the addition of anthocyanin extract, at increasing amounts, the aims of the successive trials were (i) to replace soft wheat flour with semolina (trials 2–6); (ii) to replace butter with extra virgin olive oil (trials 5–6); and (iii) to reduce as much as possible the amount of fat and sugar (trials 4–6). Nutritional improvements had to be balanced with technological features. Trials 1–6 did not encounter technological drawbacks, while incohesive dough, cracked biscuits, and surface defects were observed at higher levels of enrichment or lower amounts of fat than those adopted in trial 6 (data not shown).

Replacing soft wheat flour (trial 1) with semolina (trials 2–6) led to a significant increase (P < 0.05) of carotenoid pigments in the final product (Table 3). The levels of carotenoids observed in semolina-based biscuits varied from 3.57 × 10⁻³ to 4.27 × 10⁻³ g β-carotene/kg biscuits. These amounts were in the range of other durum wheat-derived products such as pasta (Borrelli, Troccoli, Di Fonzo, & Fares, 1999) and durum wheat bread (Pasqualone, Caponio, & Simeone, 2004). A positive contribution of carotenoids from the added lipids, either olive oil (Bilancia, Caponio, Sikorska, Pasqualone, & Summo, 2007) or butter (Nozière, Graulet, Lucas, Martin, Grolier, & Doreau, 2006), was observed. Biscuits obtained in trials 2 and 3 showed significantly higher levels of total carotenoid pigments (P < 0.05) than those from trial 4, characterized by a formulation less rich in butter. Trial 4 led to biscuits with similar carotenoid contents to those from trials 5 and 6, where butter was replaced by the same amount of extra virgin olive oil.

Table 3. Average nutritional values (per kg product) of biscuits derived from different production trials.
Tabla 3. Promedio de valores nutricionales (por kg de producto) de las galletas, derivado de distintas pruebas de producción.

Parameter	Trial
	1	2	3	4	5	6	Control
Moisture (g)	49 ± 2	50 ± 2	48 ± 1	52 ± 2	49 ± 1	49 ± 1	50 ± 1
Proteins (g)	58 ± 1	63 ± 2	63 ± 2	72 ± 3	71 ± 2	71 ± 2	71 ± 2
Total carbohydrates (g)	703 ± 7	699 ± 9	697 ± 8	749 ± 9	750 ± 10	749 ± 9	749 ± 10
Total fats (g)	162.1 ± 5.9^a	161.8 ± 6.5^a	162.2 ± 5.2^a	99.9 ± 3.8^b	101.0 ± 4.0^b	101.6 ± 3.3^b	101.3 ± 4.1^b
Saturated fats (g)	109.0 ± 3.7^a	108.2 ± 4.2^a	108.9 ± 3.6^a	65.5 ± 2.8^b	14.6 ± 0.4^c	14.9 ± 0.5^c	14.6 ± 0.3^c
Unsaturated fats (g)	49.6 ± 2.3^a	49.8 ± 3.0^a	50.0 ± 2.9^a	30.1 ± 1.4^b	83.6 ± 2.8^c	84.6 ± 3.0^c	82.8 ± 2.4^c
Total carotenoids (g β-carotene)	1.97 ± 0.06 × 10^−3a	4.27 ± 0.08× 10^−3b	4.25 ± 0.10 × 10^−3b	3.57 ± 0.08 × 10^−3c	3.60 ± 0.11 × 10^−3c	3.62 ± 0.09 × 10^−3c	3.60 ± 0.12 × 10^−3c
Total anthocyanins (g CyCl)	42 ± 3 × 10^−3a	42 ± 2 ×10^−3a	59 ± 4 × 10^−3b	60 ± 4 × 10^−3b	60 ± 4 × 10^−3b	103 ± 5 × 10^−3c	0.0 ± 0.0^d
Energy value (kJ)*	18.8 ± 0.9 × 10^3a	18.8 ± 1.0 × 10^3a	18.8 ± 0.8 × 10^3a	17.5 ± 0.8 × 10^3b	17.5 ± 1.0 × 10^3b	17.6 ± 0.9 × 10^3b	17.5 ± 0.8 × 10^3b
Note: *Determined by Atwater general conversion factors. Different letters in row indicate significant differences at P < 0.05 based on Tukey HSD test.

The substitution of butter with extra virgin olive oil was made to lower the content of saturated fatty acids, and in general, the reduction of fat and sugar was aimed at decreasing the energy value. The average nutritional values of biscuits (Table 3) indicated a significant improvement of the nutritional quality from trials 1 to 6, lowering fat content to 100 g/kg, that is about half the level of the majority of marketed biscuits. A notable decrease of saturated fatty acids, balanced by an increase of unsaturated fatty acids, was also achieved (Table 4).

Table 4. Fatty acid composition of biscuits derived from different production trials.
Tabla 4. Composición de ácidos grasos de galletas derivada de distintas pruebas de producción.

Fatty acid (g kg^−1 fat)	Trial
	1	2	3	4	5	6	Control
≤C14:0	110.5 ± 3.5	111.9 ± 4.1	115.5 ± 4.8	113.3 ± 3.0	n.d.	n.d.	n.d.
C14:0	115.3 ± 4.1	116.1 ± 5.0	116.4 ± 4.9	111.3 ± 3.1	n.d.	n.d.	n.d.
C16:0	327.3 ± 6.9	326.5 ± 8.1	323.3 ± 7.2	322.4 ± 7.8	125.2 ± 4.2	126.1 ± 4.8	124.9 ± 4.7
C18:0	119.5 ± 3.8	117.8 ± 3.3	116.9 ± 4.1	114.8 ± 2.9	19.8 ± 0.6	19.5 ± 0.4	19.2 ± 0.3
C18:1	238.5 ± 5.7	239.1 ± 6.1	239.6 ± 5.5	233.4 ± 5.2	735.7 ± 9.9	737.4 ± 9.6	732.1 ± 9.3
C18:2	52.9 ± 1.8	54.3 ± 2.1	53.6 ± 2.0	58.9 ± 2.6	85.6 ± 2.9	86.3 ± 2.8	82.1 ± 2.0
C18:3	15.0 ± 0.4	15.7 ± 0.4	15.2 ± 0.3	12.1 ± 0.3	6.1 ± 0.1	5.8 ± 0.1	5.5 ± 0.1
Note: n.d. = not detected.

Regarding the level of anthocyanin enrichment achieved, biscuits derived from trial 6 contained 103 × 10⁻³ g CyCl/kg product. Considered that the mean weight of biscuits was about 8 g, five biscuits supply for about 4 × 10⁻³ g anthocyanins, similarly to the most marketed food supplements, that contain 2 × 10⁻³ – 18 × 10⁻³ g anthocyanins per dose (Rubino, 2011).

The effect of the various changes in formulation on the color of biscuits was analyzed (Table 5), to take into account that color is one of the main factors influencing food quality and acceptability. The color of biscuits was influenced by the content of yellow and red pigments in the starting flour, namely carotenoids and anthocyanins, as well as by the Maillard and caramelization reactions that occur during the baking process. On the whole, biscuits from trials 1 to 5 appeared from pale to dark orange, those from trial 5 were reddish, and control biscuits showed a more evident yellow hue. On the whole, the index b* varied from 18.10 to 31.33, indicating a variable contribution of yellowness. The highest value of b* was observed in control biscuits, where anthocyanins were absent. This index was significantly correlated (P < 0.01) with total carotenoid pigments (Table 6), as already reported in literature (Digesù, Platani, Cattivelli, Mangini, & Blanco, 2009; Pasqualone et al., 2004). Besides, b* appeared to be negatively correlated with anthocyanin content (P < 0.01). The index a* significantly increased from trials 1 to 6 with values from 3.01 to 8.94, in the field of redness. It resulted to be correlated with anthocyanin level (P < 0.001), as expected due to the red color of anthocyanin extracts (Bechtold et al., 2007), while it was not correlated with carotenoids (Table 6). The values of 100 − L* (BI) resulted to be correlated with anthocyanin content (P < 0.001) but not with carotenoids. The anthocyanin-enriched biscuits from trial 6 showed significantly higher 100 − L* than all the other samples. The chromatic attribute C* (chroma or saturation) was significantly higher in the control, denoting a more vivid color in these samples than in all the other biscuits. It was negatively correlated with anthocyanin content (P < 0.001). The hue (h°) ranged from 67.10 to 86.13, with the lowest value (indicating the highest redness) in the anthocyanin-enriched biscuits from trial 6, while the highest value (indicating the highest yellowness) was observed in the control biscuits. Hue was negatively correlated with anthocyanin level (P < 0.001) and positively correlated with carotenoids (P < 0.05) (Table 6).

Table 5. Colorimetric indices of biscuits derived from different production trials.
Tabla 5. Índices colorimétricos de galletas derivados de distintas pruebas de producción.

Index	Trial
	1	2	3	4	5	6	Control
100−L*	31.66 ± 1.03^a	30.59 ± 1.10^a	34.31 ± 1.02^b	35.24 ± 1.22^b	36.18 ± 0.80^b	38.31 ± 0.77^c	20.26 ± 0.62^d
a*	3.01 ± 0.08^a	3.18 ± 0.10^a	4.47 ± 0.20^b	4.58 ± 0.21^b	4.51 ± 0.18^b	8.94 ± 0.51^c	2.12 ± 0.11^d
b*	18.10 ± 0.99^a	27.40 ± 0.98^b	29.34 ± 0.56^b	18.26 ± 0.82^a	20.36 ± 1.13^ac	21.17 ± 1.03^c	31.33 ± 0.97^d
C*	18.34 ± 0.87^a	27.58 ± 1.21^b	29.68 ± 1.36^bc	18.82 ± 0.93^a	20.85 ± 1.26^ad	22.98 ± 1.21^d	31.40 ± 1.38^c
h°	80.55 ± 1.43^a	83.38 ± 2.91^ad	81.34 ± 2.30^a	75.91 ± 1.54^b	77.51 ± 1.02^b	67.10 ± 1.13^c	86.13 ± 3.48^d
Note: Different letters in row indicate significant differences at P < 0.05 based on Tukey HSD test.

Table 6. Correlations between total anthocyanin content, total carotenoid content, and colorimetric indices of biscuits derived from different production trials.
Tabla 6. Correlaciones entre el contenido de antocianinas total, el contenido de carotenoides total y los índices colorimétricos de galletas, derivadas de distintas pruebas de producción.

	a*	b*	100−L*	C*	h°
Total anthocyanins	0.934***	−0.641**	0.932***	−0.762***	−0.920***
Total carotenoids	0.283	0.656**	0.134	0.281	0.468*
Note: *, **, *** = significant correlation at P = 0.05, 0.01, and 0.001, respectively.

Paired-preference test involving anthocyanin-enriched biscuits versus control plain biscuits

The biscuits derived from trial 6 and control biscuits (i.e. biscuits obtained in the same conditions except for the addition of anthocyanins) were the object of a paired-preference test aimed to assess product acceptability and the willingness of consumers to purchase (Figure 1). The obtained results (Table 7) show that differences in taste and color were evident, as stated by a large majority of consumers (P < 0.001). The differences in color detected by the consumers confirmed the results of biscuit color analysis. These differences, however, did not affect the acceptability of anthocyanin-enriched biscuits. In fact, both in color and taste, as well as at a global evaluation, the number of consumers who preferred the enriched biscuits was not significantly different from the number of people who preferred control biscuits. Finally, consumers were asked if they would also buy the non-preferred biscuit. In the case of a global preference for the anthocyanin-enriched biscuits (42 people), no significant difference was observed between the number who expressed a willingness to buy the non-preferred product (20 people) and those who would not buy it (22 people). On the contrary, the majority of consumers who expressed a preference for the control biscuits did not express a willingness to buy the non-preferred product (P < 0.05).

In conclusion, on the whole, the obtained results demonstrated the feasibility of the production of anthocyanin-enriched durum wheat biscuits, containing extra virgin olive oil, with low levels of fats and sugar. The proposed functional biscuits showed adequate acceptability, and their realization could give a new valorization to an abundant by-product of wine-making industry, such as grape marc.

Table 7. Results of the paired-preference test aimed to assess biscuit acceptability and willingness to buy among 103 biscuit consumers (A = anthocyanin-enriched biscuits derived from trial 6; B = control biscuits).
Tabla 7. Resultados de las pruebas de preferencia realizadas entre 103 consumidores de galletas con el fin de evaluar la aceptabilidad y la disposición de comprar las galletas, aplicadas a dos muestras de galletas (A = galletas enriquecidas con antocianinas derivadas de la prueba 6; B = galletas de control).

Question	Type and number of responses		Level of significance
Regarding taste, which biscuit do you prefer?	A, 42	B, 61	n.s.
Was the difference in taste evident?	Yes, 101	No, 2	0.001
Regarding color, which biscuit do you prefer?	A, 50	B, 53	n.s.
Was the difference in color evident?	Yes, 102	No, 1	0.001
On the whole, which biscuit do you prefer?	A, 42	B, 61	n.s.
Would you buy the less preferred biscuit? (Preference for A)	Yes, 20	No, 22	n.s.
Would you buy the less preferred biscuit? (Preference for B)	Yes, 17	No, 44	0.05
Note: n.s. = not significant. Friedman test was applied to verify if responses in each pair differed significantly from each other.
Nota: n.s. = no significativo. Se aplicó la prueba de Friedman para determinar si las respuestas obtenidas frente a cada par de galletas mostraban variaciones significativas entre ellas.

Acknowledgements

The authors gratefully acknowledge Bob Pasqualone for linguistic revision.