Impact of Essential Oils on the Taste Acceptance of Tomato Juice, Vegetable Soup, or Poultry Burgers Laura Espina, Diego Garc´ıa-Gonzalo, and Rafael Pag´an
Abstract: Despite the vast body of available literature on the possibilities of essential oils (EOs) as food preservatives or functional ingredients, the sensory impact of their addition to foods has barely been approached. This work focuses on the hedonic taste acceptance of 3 food products (tomato juice, vegetable soup, and poultry burgers) when they are incorporated with potentially antimicrobial concentrations (20 to 200 μL/L) of 6 selected EOs (lemon, pennyroyal mint, thyme, and rosemary) and individual compounds (carvacrol, p-cymene). Although addition of 20 μL/L of pennyroyal mint or lemon EO did not change the taste acceptance of tomato juice, higher concentrations of these compounds or any concentration of the other 4 compounds did. In vegetable soup, the tolerance limit for rosemary EO, thyme EO, carvacrol, or p-cymene was 20 μL/L, while the addition of 200 μL/L of lemon EO was accepted. Tolerance limits in poultry burgers were established in 20 μL/L for carvacrol and thyme EOs, 100 μL/L for pennyroyal mint EO and p-cymene, and 200 μL/L for lemon and rosemary EOs. Moreover, incorporation of pennyroyal mint EO to tomato juice or poultry burgers, and enrichment of vegetable soup with lemon EO, could contribute to the development of food products with an improved sensory appeal. Keywords: antimicrobial activity, functional food, essential oil, food additives, sensory analysis
Introduction
scientific and popular interest because of (i) their GRAS (generally recognized as safe) status, (ii) their antimutagenic and antidiabetic properties that accounting for their addition as active compounds in functional foods (Sacchetti and others 2005; Bakkali and others 2008), (iii) their potential multipurpose use as food preservatives, exhibiting antibacterial, antimycotic, antitoxigenic, and antioxidant bioactivity (Burt 2004; Sacchetti and others 2005), and (iv) their synergy with other preservation techniques (Burt 2004; Raybaudi-Massilia and others 2009). These facts are opening up the possibility of producing minimally processed products treated at lower intensities. Addition of EOs or ICs has a usually strong impact on the flavor of foods. Since sensory appeal and taste acceptance strongly influence the repurchase of food products, it is necessary to diminish the sensory impact of added compounds by careful selection of EO or IC and by minimization of the concentration (Burt 2004). An intelligent approach to this problem is to select bioactive EOs or ICs extracted either from culinary fruits (such as lemon), or from plant material commonly used as seasonings or condiments in the form of spices or herbs (such as thyme, rosemary, mint, or oregano) (Tajkarimi and others 2010; Settanni and others 2012). Furthermore, the antimicrobial activity or the enhancement of the antimicrobial activity of other preservation techniques has been MS 20140314 Submitted 2/25/2014, Accepted 5/2/2014. Authors are with Dept. de Producci´on Animal y Ciencia de los Alimentos, Facultad de Veterinaria, Univ. de described in laboratory media or in foods for small concentrations Zaragoza, Miguel Servet 177, 50013 Zaragoza, Spain. Direct inquiries to author (200 μL/L at most) of these EOs (Citrus lemon L. EO, Thymus Pag´an (E-mail: [email protected]). algeriensis L. EO, Rosmarinus officinalis L. EO, Mentha pulegium L. EO) or the ICs carvacrol and p-cymene, which are major ICs
Convenience, health-related concerns and sensory appeal are listed as the most important factors that determine food choice (Steptoe and others 1995). The importance of convenience in eating behavior is reflected in a tendency to save time and effort by opting for ready-to-eat (RTE) or ready-to-cook (RTC) products with enhanced shelf life (Kilinc and others 2008). Recent consumers’ desire to maintain a healthy diet has led to changes in eating habits that include an increased intake of vegetables (Doyle and Erickson 2008), a preference shift toward fresh or minimally processed food whose organoleptic and nutritional properties remain as unaltered as possible (Abadias and others 2008), and an increased concern about the toxicological safety of chemicals and synthetic preservatives (Sacchetti and others 2005). This last concern about the synthetic origin of additives has resulted in a growing demand for natural alternatives such as bacteriocins or plant essential oils (EOs) (Sacchetti and others 2005). Many EOs and their individual constituents (ICs), which have been traditionally used as flavorings in drinks and food for their pleasant aroma and taste (Bauer and others 2001), are regaining
R C 2014 Institute of Food Technologists
doi: 10.1111/1750-3841.12529 Further reproduction without permission is prohibited
Vol. 79, Nr. 8, 2014 r Journal of Food Science S1575
S: Sensory & Food Quality
Practical Application: This work highlights the importance of sensory analyses when considering the use of EOs or their individual constituents (ICs) as food preservatives or functional ingredients in the food industry. It categorizes selected combinations of compounds and food products, and proposes the use of very low EO concentrations (20 to 200 μL/L). Moreover, it demonstrates that the selection of EOs or ICs based on their sensory compatibility with each food product can lead to the development of new, distinct products with a pleasant taste and/or with fresh-like quality.
Sensory impact of essential oils in food . . .
S: Sensory & Food Quality
in oregano EO (Ait-Ouazzou and others 2011a, 2011b, 2012a, 2012b; Espina and others 2012; Settanni and others 2012; Espina and others 2013). Even concentrations as low as 50 or 20 μL/L of carvacrol or citral (an IC present in lemon EO) have been applied in combination with heat to inactivate foodborne pathogens in fruit juices (Espina and others 2010; Ait-Ouazzou and others 2013). Despite the vast body of available literature on the possibilities of EOs and ICs as food preservatives, the sensory impact of their addition to foods has been barely approached. In fact, very few studies on the antimicrobial or antioxidant effect of EOs or ICs are complemented by preliminary analyses covering the organoleptic changes caused in food products by the incorporation of EOs or ICs (Omidbeygi and others 2007; Mitchell and others 2013; Espina and others 2014). To the best of our knowledge, however, a comprehensive study has yet to be conducted on consumers’ response to incorporation of several concentrations of different EOs or ICs affecting the hedonic acceptance of food matrices. For this investigation, therefore, 3 food products were selected according to the aforementioned criteria of convenience and healthy image: tomato juice, in representation of RTE tomato products, the tomato being the second most consumed vegetable in the U.S.A. (USDA 2010); vegetable soup meeting consumers’ expectations for a convenient, nutritious, and low-energy-dense product; and poultry burgers, representing widely consumed RTC meat products with a low fat content and a high unsaturation degree of fatty acids. Sensory analyses were performed to (i) assess the hedonic acceptance of these 3 food products infused with different concentrations of lemon EO, thyme EO, rosemary EO, pennyroyal mint EO, carvacrol, and p-cymene; (ii) select the taste limit concentration of each added compound for each food product; (iii) determine whether consumption frequency of each food product influenced on the acceptance of added compounds, and (iv) identify possible target groups of consumers of new food products.
Material and Methods Subjects and installations for sensory analyses A total number of 65 untrained panelists were recruited from the staff and students of the Veterinary School at the Univ. of Zaragoza, Spain. We selected the panel size according to statistical considerations stated by Hough and others (2006). Out of the 22 men, 32% were aged under 21, 54% were aged 21 to 30, and 14% were more than 30 y old; out of the 43 women, 21% were under 21, 65% were aged 21 to 30, and 14% were more than 30 y old (mean age = 24.5 y, SD = 7.6). All sensory work was performed in the sensory laboratory at the Pilot Plant of Food Science and Technology (Univ. de Zaragoza) under white fluorescent light. The panelists were situated in a total of 9 private booths to minimize distractions and possible interactions during the tests. Each booth was equipped with a personal computer to input data directly.
cited papers. Carvacrol (98%) and p-cymene (99%) were purchased from Sigma-Aldrich Chemie (Steinheim, Germany).
Preparation of food samples Both tomato juice from concentrate (Auchan, Murcia, Spain) and vegetable soup (Auchan, Boulogne-Billancourt, France) were UHT-treated, shelf-stable products, packaged in 1 L-TetraBrik carton package. These products were purchased in the same week of their evaluation, at a local supermarket and upon receipt were stored at 0 to 4 °C. Vegetable soup was composed of 16% of potato, 11% of carrot, and 12% of onion, tomato, leek, turnip, spinach, and celeriac. Tomato juice and vegetable soup were poured onto 1.5 L plastic bottles, and then the necessary quantity of each EO or IC was added to reach concentrations of 0, 20, 100, or 200 μL/L, and the whole was shaken vigorously to ensure even distribution of each added substance. These concentrations were chosen according to preliminary sensorial tests and to published results on their antibacterial activity when they were added to laboratory media and foods (Ait-Ouazzou and others 2011a, 2011b, 2012a, 2012b; Espina and others 2012). For each sample, 20 mL of juice or soup were offered in a transparent 10 cL coded glass cup at room temperature. Fresh minced chicken and turkey meat (Martinez especialidades, Valencia, Spain) was purchased in modified atmosphere trays containing 400 g each, and stored at 0 to 4 °C until used (3 d at most). When indicated, each EO or IC was added at 20, 100, or 200 μL/L before mixing and hand-forming the poultry minced meat into burgers. Afterward they were cooked with olive oil in a frying pan, it being ascertained that the center of the burger reached 60 °C as measured with a thermocouple (Almeco, mod. ZA 020-FS, Bernburg, Germany) and then was kept at 60 °C until cut into 20 g pieces and served on coded sectioned platters at room temperature. Procedure for sensory analyses Sensory tests were performed on 9 different days, with 3 d allocated for each food product and 2 EOs or ICs tasted each day. On each day, panelists were first offered a sample without EOs or ICs presented as a control, and then 2 groups of 3 samples each (with 20, 100, or 200 μL/L of either compound), so evaluating 6 samples in all. After the control sample, both groups of samples and each sample within both groups were offered in counterbalanced order (that is, in a different order for each judge). Water and palate cleansers taken between samples helped neutralize the EOs or ICs’ possible residual presence in the mouth. Breadsticks were used as the palate cleanser for burgers, and plain yogurt (Weidegl¨uck, Schwaben, Germany) was served as a better palate cleanser for liquid foods (tomato juice and vegetable soup) for its higher capacity of distribution in the mouth. On each day, panelists were asked to determine the hedonic acceptance of the 6 samples by ranking them in a 9-points scale going from −4 (much worse than the control) to +4 (much better than the control). Before testing the samples, panelists were also asked to evaluate the frequency of consumption of each food product by rating it 1 (every day), 2 (once per week or more often), 3 (once every 2 wk), 4 (once per month), or 5 (less than once per month).
Tested compounds The tested compounds were 4 EOs (lemon EO from Citrus lemon L., pennyroyal mint EO from Mentha pulegium L., thyme EO from Thymus algeriensis L., and rosemary EO from Rosmarinus officinalis L.) and 2 ICs present in EOs (carvacrol and p-cymene). Citrus lemon L. EO was obtained as described by Espina and others Statistical analyses (2011), while M. pulegium, T. algeriensis, and R. officinalis EOs were All data regarding age and sex of each 1 of the 65 panelists, obtained as described by Ait-Ouazzou and others (2011b, 2012b). as well as the frequency of consumption of each food product The composition of all these EOs was analyzed and reported in the and the hedonic value assigned to each 1 of the 54 noncontrol S1576 Journal of Food Science r Vol. 79, Nr. 8, 2014
Sensory impact of essential oils in food . . . Table 1–Mean, SD (standard deviation), P50 (median or 50th percentile), P90 (90th percentile), skew. (skewness), IQR (interquartile range), and PG (percentage of panelists rating the sample better than the control) values for each evaluated sample of tomato juice, vegetables soup, or poultry burgers containing 20, 100, or 200 μL/L of lemon EO, pennyroyal mint EO, thyme EO, rosemary EO, carvacrol, or p-cymene. All values are expressed in a 1 to 9 scale where 5.00 is the control score. Asterisks indicate statistically significant difference between the mean value and the control, according to Wilcoxon test (α = 0.05). Lemon EO 20
100
Carvacrol 200
20
Pennyroyal mint EO 20
100
200
Tomato juice Mean 5.03 3.48∗ 3.23∗ 3.97∗ SD 1.68 1.75 1.87 1.72 P50 5.00 3.00 3.00 4.00 P90 7.00 6.00 7.00 6.40 Skew. -0.11 0.56 1.00 0.39 IQR 2.00 2.00 2.00 2.00 PG 40.00 15.38 13.85 20.00
2.75∗ 1.49 2.00 5.00 0.50 3.00 3.01
2.25∗ 5.58∗ 4.22∗ 4.11∗ 3.55∗ 1.51 1.81 2.01 2.22 2.13 2.00 6.00 4.00 4.00 3.00 4.00 8.00 7.00 7.00 6.40 1.17 -0.05 0.42 0.26 0.57 2.00 3.00 3.00 4.00 3.00 4.62 50.77 35.38 30.77 23.08
Vegetables soup Mean 5.60∗ 5.15 4.78 4.91 SD 1.41 1.80 1.87 1.78 P50 6.00 5.00 5.00 5.00 P90 7.40 7.00 7.00 7.00 Skew. 0.17 −0.14 −0.26 −0.47 IQR 1.00 2.50 2.00 2.00 PG 52.31 49.23 41.54 38.46
3.58∗ 1.68 3.00 6.00 0.71 2.00 15.38
2.58∗ 4.26∗ 1.69 1.70 2.00 4.00 5.00 6.00 1.46 0.05 2.50 2.50 7.70 24.61
100
2.89∗ 1.28 3.00 4.00 0.44 2.00 4.62
200
Thyme EO 20
Rosemary EO
100
200
20
100
2.12∗ 1.46 2.00 5.00 1.73 1.00 6.15
2.02∗ 4.68 2.97∗ 1.56 1.99 2.02 1.00 4.00 3.00 5.00 8.00 7.00 1.62 0.40 1.02 2.00 3.00 3.00 6.15 32.31 15.38
200
p-cymene 20
100
200
2.23∗ 4.79 3.48∗ 2.74∗ 1.75 1.96 1.89 1.79 2.00 4.00 3.00 3.00 4.40 8.00 6.00 6.00 1.72 0.21 0.64 1.12 2.00 3.00 2.00 3.00 7.70 36.92 18.46 12.31
2.14∗ 5.06 3.83∗ 3.20∗ 5.38 3.86∗ 3.28∗ 4.98 1.25 1.74 1.71 1.63 1.55 1.81 1.79 1.51 2.00 5.00 4.00 3.00 5.00 4.00 3.00 5.00 4.00 8.00 6.00 6.00 7.00 6.00 6.00 7.00 0.87 −0.04 0.29 0.40 −0.42 0.20 0.29 −0.54 2.00 2.00 2.50 2.00 1.50 2.50 2.00 2.00 1.54 33.84 16.92 13.84 47.69 23.08 18.46 36.92
2.97∗ 1.52 3.00 4.40 0.99 2.00 7.69
2.38∗ 1.37 2.00 4.00 1.43 2.00 3.08
samples, were processed in SPSS Statistics 20.0 (IBM, Corp. Armonk, N.Y., U.S.A.), which was used to perform descriptive statistics and all statistical tests at a significance level of α = 0.05. The Kolmogorov–Smirnov test and observation of the Normal Q–Q plot were applied to determine the normality of the distributions; the Wilcoxon signed-rank test was applied to determine whether differences between each sample and the control sample were statistically significant; the Chi-square test with Monte Carlo simulation was applied to determine whether each food product’s consumption frequency had an influence on the acceptance of added compounds. Chi-square was also used to test other possible correlations, such as that between concentration and spread of the data. Additional tests, such as Kruskal–Wallis with Dunn’s multiple-comparison test, were used to test differences in the SD values between different groups.
Results and Discussion Determination of the hedonic acceptance of food products added with 20, 100, or 200 μL/L of each compound Table 1 contains the descriptive statistics (featuring measurements of centrality and spread) calculated from the hedonic scores shown in a 1 to 9 scale, with 5.00 being attributed to the control sample. Regarding the spread of the data, Table 1 shows the standard deviation (SD, a summary measure of the differences of each score from the total mean score) alongside the interquartile range (IQR, the difference between 75th and 25th percentiles). Taking into account that normally distributed data in a 9-point scale (that is, stanine scale) have a SD of 2, the score variability within each sample could be considered to be moderately low, with 1.76 as the average SD obtained from the 54 different samples.
SD or IQR did not vary as a function of the added compound or the concentration within each evaluated food product (Table S1), although SD values obtained from vegetable soup were significantly less dispersed than those from the other 2 food products, indicating a lower variability in the response to the addition of compounds to vegetable soup. The mean value (arithmetic average of the 65 scores for each sample) and the median (or 50th percentile, value below which half of the panelists scored each sample) are shown in Table 1 to define the centrality of each sample. The 90th percentile and the skewness coefficient, which characterizes the asymmetry of the distribution, are also presented to supply additional information on the acceptance of each sample. Since majority of data could not be fitted to a normal distribution, the nonparametric Wilcoxon test was performed to identify samples statistically different in their mean score from the control; those different means appear marked with an asterisk in Table 1. Furthermore, these mean values statistically different from 5.00 corresponded to those samples whose 50th percentile or median value was other than 5.00, with a few exceptions (tomato juice with 20 μL/L of rosemary EO or 20 μL/L of p-cymene; and poultry burgers with 20 μL/L of lemon EO, 100 μL/L of carvacrol, or 100 μL/L of rosemary EO). In this regard, we decided to use both criteria (Wilcoxon test and median values) to select conditions causing actual improvement or deterioration of each food product. Mean values are also depicted in Figure 1, which shows the average improvement or deterioration after addition of each concentration and compound, in terms of a percentage over or under the control set at 0. This percentage represented the proportion of variation of each sample mean in relation to the original 5.00 score. EOs or ICs are ranked in each figure according to the general acceptance for each food product, as a compendium of the 3 added concentrations. Vol. 79, Nr. 8, 2014 r Journal of Food Science S1577
S: Sensory & Food Quality
Poultry burgers Mean 5.63∗ 4.97 4.80 5.23 4.58 3.60∗ 5.31 4.82 4.09∗ 4.86 4.22∗ 3.45∗ 5.72∗ 4.52 5.00 5.17 4.94 3.43∗ SD 1.40 1.72 1.86 1.41 1.74 1.83 1.79 2.00 2.08 1.92 2.00 1.98 1.57 1.99 2.13 1.80 1.97 1.96 P50 5.00 5.00 5.00 5.00 4.00 3.00 5.00 5.00 4.00 5.00 4.00 3.00 6.00 4.00 5.00 5.00 5.00 3.00 7.00 7.00 7.40 7.00 7.00 6.00 8.00 7.40 7.00 8.00 7.00 6.40 8.00 7.00 8.00 8.00 8.00 7.00 P90 Skew. −0.15 0.18 −0.07 −0.25 −0.03 0.54 0.21 −0.04 0.18 0.09 0.33 0.87 −0.71 0.18 −0.08 −0.26 −0.12 0.72 IQR 2.00 2.00 2.50 2.00 2.00 3.00 3.00 3.50 4.00 2.00 3.00 2.00 2.00 3.00 3.50 2.00 2.00 2.00 PG 49.23 41.54 38.46 44.61 29.23 20.00 46.15 38.46 32.31 30.77 30.77 16.92 60.00 35.38 47.69 38.46 41.54 15.38
Sensory impact of essential oils in food . . . Several conclusions can be drawn from Table 1 and Figure 1. On the one hand, the ranking was different for each food product, meaning that acceptance of a determined EO or IC was dependent upon the food matrix they were added to. On the other hand, for all cases (except for addition of rosemary EO in burgers), increasing concentrations of compounds led to a decrease in the sensory acceptance. For example, tomato juice or vegetable soup obtained higher scores than the control samples when adding 20 μL/L of pennyroyal mint EO or rosemary EO, respectively, but the mean values were reduced by about 15% and 20% when adding 100 μL/L. Furthermore, a noteworthy decrease (by 50% to 60% when compared with the control) was discovered when adding 200 μL/L of rosemary EO, carvacrol, or thyme EO to tomato juice or the same concentration of carvacrol, p-cymene, or pennyroyal mint EO to vegetable soup. Moreover, these poorly appreciated samples were also characterized by strong positively skewed distributions, and by having achieved a maximum score of 2.00 and of 5.00 as evaluated by 50% and 90% of the panelists, respectively.
% variation in relation to control
A
20
Table 2 shows the tolerance limit for each EO or IC in the assayed conditions, defined as the maximum tested concentration whose addition did not modify the hedonic acceptance of the food product. Since for nonnormal distributions, the median is the most appropriate statistic of centrality (Choi 2005), limits were set to the maximum tested concentration that achieved a median value of at least 5.00. In tomato juice, only the lowest concentration of pennyroyal mint EO or lemon EO was tolerated, while the rest of the compounds would be dismissed. In fact, the addition of 20 μL/L of pennyroyal mint EO increased sensory acceptance. Vegetable soup tolerated the incorporation of the lowest concentration of rosemary EO, thyme EO, carvacrol, or p-cymene, but not any concentration of pennyroyal mint EO. However, the hedonic acceptance of vegetable soup increased after addition of 20 μL/L of lemon EO, and was not affected by higher concentrations. Consumers also evaluated positively the addition of 20 μL/L of rosemary EO or lemon EO in burgers, and tolerated higher concentrations: 20 or 100 μL/L of pennyroyal mint EO or p-cymene,
L
P
R
C
T
L
R
T
C
P
M
L
R
M
P
C
T
0 -20 -40 -60 -80
% variation in relation to control
B 20 0 -20 -40 -60 -80
C 20 % variation in relation to control
S: Sensory & Food Quality
M
0 -20 -40 -60 -80
S1578 Journal of Food Science r Vol. 79, Nr. 8, 2014
Figure 1–Percentage of variation of each mean score in comparison with the control sample, corresponding to tomato juice (A), vegetable soup (B), or burger meat (C) infused with 20 μL/L (white bars), 100 μL/L (light gray bars), or 200 μL/L (dark gray bars) of lemon EO (L), pennyroyal mint EO (M), thyme EO (T), rosemary EO (R), carvacrol (C), or p-cymene (P). Striped bars represent significant difference between the mean value and the control, according to the Wilcoxon test (α = 0.05).
Sensory impact of essential oils in food . . .
Determination of the influence of consumption frequency on the acceptance of added compounds In order to facilitate statistical analysis, for each food product, 2 groups of consumers were arranged: a group of frequent consumers (who used to have the product at least once per week, composed of those consumers who had chosen options 1 or 2 in the survey’s question on consumption frequency) and a group of occasional consumers, who had it less often and had answered options 3, 4, or 5. Table 3 shows the mean and median values obtained from both groups of consumers for each product added with any concentration of each EO or IC. According to the analysis of the mean values, there was no influence of the consumption frequency on the general acceptance of the samples, except for burgers incorporated with pennyroyal mint EO, thyme EO or p-cymene, showing significantly higher rates for frequent consumers. Although comparison of medians revealed higher rates in the frequent consumers group for some tomato juice and vegetable soup samples, the low number of frequent consumers of these products led to dismissal of further analyses. When considering concentrations in burgers, Chi-square tests revealed that score differences only happened when adding 200 μL/L of thyme EO or p-cymene, and 100 or 200 μL/L of pennyroyal mint EO (Table S2). Out of these samples, the effect of consumption frequency in the acceptance of samples added with pennyroyal mint EO was remarkable: while frequent consumers of burgers preferred samples with an added 100 or 200 μL/L over the control, 50% of occasional consumers rated them 4 or lower. Additionally, frequent consumers of burgers also ranked samples added with 20 μL/L of thyme EO or with 100 μL/L of p-cymene above the control. Identification of subgroups of consumers with a positive response toward incorporation of EOs or ICs Each range of 65 scores was represented in a histogram (Figure S1) to observe the distribution of data for each condition. Having acknowledged the general lack of normality and regardless of the central tendency, some represented data set seemed to resemble bimodal distributions (distributions with 2 modes), suggesting the presence of 2 different trends within the set of panelists. These distributions also belonged to samples with higher SD values than the average ones, which also concurs with the idea of having 2 different central tendencies. In order to help in the identification of samples provoking this divided response, the proportion of panelists that rated each sample as better than the control was calculated, and it is shown in
Table 2–Tolerance limits in μL/L of each EO or IC for each food product. Limits were defined as the maximum tested concentration whose addition did not modify the hedonic acceptance of the food product, since 50% of the panelists considered the sample with that added concentration at least as good as the control. Limits under 20 μL/L indicate that all the tested concentrations of an EO or IC decreased the hedonic acceptance of the food product.
Lemon EO Carvacrol Pennyroyal mint EO Thyme EO Rosemary EO p-cymene
Tomato juice
Vegetables soup
Poultry burgers
20
Abstract: Despite the vast body of available literature on the possibilities of essential oils (EOs) as food preservatives or functional ingredients, the sensory impact of their addition to foods has barely been approached. This work focuses on the hedonic taste acceptance of 3 food products (tomato juice, vegetable soup, and poultry burgers) when they are incorporated with potentially antimicrobial concentrations (20 to 200 μL/L) of 6 selected EOs (lemon, pennyroyal mint, thyme, and rosemary) and individual compounds (carvacrol, p-cymene). Although addition of 20 μL/L of pennyroyal mint or lemon EO did not change the taste acceptance of tomato juice, higher concentrations of these compounds or any concentration of the other 4 compounds did. In vegetable soup, the tolerance limit for rosemary EO, thyme EO, carvacrol, or p-cymene was 20 μL/L, while the addition of 200 μL/L of lemon EO was accepted. Tolerance limits in poultry burgers were established in 20 μL/L for carvacrol and thyme EOs, 100 μL/L for pennyroyal mint EO and p-cymene, and 200 μL/L for lemon and rosemary EOs. Moreover, incorporation of pennyroyal mint EO to tomato juice or poultry burgers, and enrichment of vegetable soup with lemon EO, could contribute to the development of food products with an improved sensory appeal. Keywords: antimicrobial activity, functional food, essential oil, food additives, sensory analysis
Introduction
scientific and popular interest because of (i) their GRAS (generally recognized as safe) status, (ii) their antimutagenic and antidiabetic properties that accounting for their addition as active compounds in functional foods (Sacchetti and others 2005; Bakkali and others 2008), (iii) their potential multipurpose use as food preservatives, exhibiting antibacterial, antimycotic, antitoxigenic, and antioxidant bioactivity (Burt 2004; Sacchetti and others 2005), and (iv) their synergy with other preservation techniques (Burt 2004; Raybaudi-Massilia and others 2009). These facts are opening up the possibility of producing minimally processed products treated at lower intensities. Addition of EOs or ICs has a usually strong impact on the flavor of foods. Since sensory appeal and taste acceptance strongly influence the repurchase of food products, it is necessary to diminish the sensory impact of added compounds by careful selection of EO or IC and by minimization of the concentration (Burt 2004). An intelligent approach to this problem is to select bioactive EOs or ICs extracted either from culinary fruits (such as lemon), or from plant material commonly used as seasonings or condiments in the form of spices or herbs (such as thyme, rosemary, mint, or oregano) (Tajkarimi and others 2010; Settanni and others 2012). Furthermore, the antimicrobial activity or the enhancement of the antimicrobial activity of other preservation techniques has been MS 20140314 Submitted 2/25/2014, Accepted 5/2/2014. Authors are with Dept. de Producci´on Animal y Ciencia de los Alimentos, Facultad de Veterinaria, Univ. de described in laboratory media or in foods for small concentrations Zaragoza, Miguel Servet 177, 50013 Zaragoza, Spain. Direct inquiries to author (200 μL/L at most) of these EOs (Citrus lemon L. EO, Thymus Pag´an (E-mail: [email protected]). algeriensis L. EO, Rosmarinus officinalis L. EO, Mentha pulegium L. EO) or the ICs carvacrol and p-cymene, which are major ICs
Convenience, health-related concerns and sensory appeal are listed as the most important factors that determine food choice (Steptoe and others 1995). The importance of convenience in eating behavior is reflected in a tendency to save time and effort by opting for ready-to-eat (RTE) or ready-to-cook (RTC) products with enhanced shelf life (Kilinc and others 2008). Recent consumers’ desire to maintain a healthy diet has led to changes in eating habits that include an increased intake of vegetables (Doyle and Erickson 2008), a preference shift toward fresh or minimally processed food whose organoleptic and nutritional properties remain as unaltered as possible (Abadias and others 2008), and an increased concern about the toxicological safety of chemicals and synthetic preservatives (Sacchetti and others 2005). This last concern about the synthetic origin of additives has resulted in a growing demand for natural alternatives such as bacteriocins or plant essential oils (EOs) (Sacchetti and others 2005). Many EOs and their individual constituents (ICs), which have been traditionally used as flavorings in drinks and food for their pleasant aroma and taste (Bauer and others 2001), are regaining
R C 2014 Institute of Food Technologists
doi: 10.1111/1750-3841.12529 Further reproduction without permission is prohibited
Vol. 79, Nr. 8, 2014 r Journal of Food Science S1575
S: Sensory & Food Quality
Practical Application: This work highlights the importance of sensory analyses when considering the use of EOs or their individual constituents (ICs) as food preservatives or functional ingredients in the food industry. It categorizes selected combinations of compounds and food products, and proposes the use of very low EO concentrations (20 to 200 μL/L). Moreover, it demonstrates that the selection of EOs or ICs based on their sensory compatibility with each food product can lead to the development of new, distinct products with a pleasant taste and/or with fresh-like quality.
Sensory impact of essential oils in food . . .
S: Sensory & Food Quality
in oregano EO (Ait-Ouazzou and others 2011a, 2011b, 2012a, 2012b; Espina and others 2012; Settanni and others 2012; Espina and others 2013). Even concentrations as low as 50 or 20 μL/L of carvacrol or citral (an IC present in lemon EO) have been applied in combination with heat to inactivate foodborne pathogens in fruit juices (Espina and others 2010; Ait-Ouazzou and others 2013). Despite the vast body of available literature on the possibilities of EOs and ICs as food preservatives, the sensory impact of their addition to foods has been barely approached. In fact, very few studies on the antimicrobial or antioxidant effect of EOs or ICs are complemented by preliminary analyses covering the organoleptic changes caused in food products by the incorporation of EOs or ICs (Omidbeygi and others 2007; Mitchell and others 2013; Espina and others 2014). To the best of our knowledge, however, a comprehensive study has yet to be conducted on consumers’ response to incorporation of several concentrations of different EOs or ICs affecting the hedonic acceptance of food matrices. For this investigation, therefore, 3 food products were selected according to the aforementioned criteria of convenience and healthy image: tomato juice, in representation of RTE tomato products, the tomato being the second most consumed vegetable in the U.S.A. (USDA 2010); vegetable soup meeting consumers’ expectations for a convenient, nutritious, and low-energy-dense product; and poultry burgers, representing widely consumed RTC meat products with a low fat content and a high unsaturation degree of fatty acids. Sensory analyses were performed to (i) assess the hedonic acceptance of these 3 food products infused with different concentrations of lemon EO, thyme EO, rosemary EO, pennyroyal mint EO, carvacrol, and p-cymene; (ii) select the taste limit concentration of each added compound for each food product; (iii) determine whether consumption frequency of each food product influenced on the acceptance of added compounds, and (iv) identify possible target groups of consumers of new food products.
Material and Methods Subjects and installations for sensory analyses A total number of 65 untrained panelists were recruited from the staff and students of the Veterinary School at the Univ. of Zaragoza, Spain. We selected the panel size according to statistical considerations stated by Hough and others (2006). Out of the 22 men, 32% were aged under 21, 54% were aged 21 to 30, and 14% were more than 30 y old; out of the 43 women, 21% were under 21, 65% were aged 21 to 30, and 14% were more than 30 y old (mean age = 24.5 y, SD = 7.6). All sensory work was performed in the sensory laboratory at the Pilot Plant of Food Science and Technology (Univ. de Zaragoza) under white fluorescent light. The panelists were situated in a total of 9 private booths to minimize distractions and possible interactions during the tests. Each booth was equipped with a personal computer to input data directly.
cited papers. Carvacrol (98%) and p-cymene (99%) were purchased from Sigma-Aldrich Chemie (Steinheim, Germany).
Preparation of food samples Both tomato juice from concentrate (Auchan, Murcia, Spain) and vegetable soup (Auchan, Boulogne-Billancourt, France) were UHT-treated, shelf-stable products, packaged in 1 L-TetraBrik carton package. These products were purchased in the same week of their evaluation, at a local supermarket and upon receipt were stored at 0 to 4 °C. Vegetable soup was composed of 16% of potato, 11% of carrot, and 12% of onion, tomato, leek, turnip, spinach, and celeriac. Tomato juice and vegetable soup were poured onto 1.5 L plastic bottles, and then the necessary quantity of each EO or IC was added to reach concentrations of 0, 20, 100, or 200 μL/L, and the whole was shaken vigorously to ensure even distribution of each added substance. These concentrations were chosen according to preliminary sensorial tests and to published results on their antibacterial activity when they were added to laboratory media and foods (Ait-Ouazzou and others 2011a, 2011b, 2012a, 2012b; Espina and others 2012). For each sample, 20 mL of juice or soup were offered in a transparent 10 cL coded glass cup at room temperature. Fresh minced chicken and turkey meat (Martinez especialidades, Valencia, Spain) was purchased in modified atmosphere trays containing 400 g each, and stored at 0 to 4 °C until used (3 d at most). When indicated, each EO or IC was added at 20, 100, or 200 μL/L before mixing and hand-forming the poultry minced meat into burgers. Afterward they were cooked with olive oil in a frying pan, it being ascertained that the center of the burger reached 60 °C as measured with a thermocouple (Almeco, mod. ZA 020-FS, Bernburg, Germany) and then was kept at 60 °C until cut into 20 g pieces and served on coded sectioned platters at room temperature. Procedure for sensory analyses Sensory tests were performed on 9 different days, with 3 d allocated for each food product and 2 EOs or ICs tasted each day. On each day, panelists were first offered a sample without EOs or ICs presented as a control, and then 2 groups of 3 samples each (with 20, 100, or 200 μL/L of either compound), so evaluating 6 samples in all. After the control sample, both groups of samples and each sample within both groups were offered in counterbalanced order (that is, in a different order for each judge). Water and palate cleansers taken between samples helped neutralize the EOs or ICs’ possible residual presence in the mouth. Breadsticks were used as the palate cleanser for burgers, and plain yogurt (Weidegl¨uck, Schwaben, Germany) was served as a better palate cleanser for liquid foods (tomato juice and vegetable soup) for its higher capacity of distribution in the mouth. On each day, panelists were asked to determine the hedonic acceptance of the 6 samples by ranking them in a 9-points scale going from −4 (much worse than the control) to +4 (much better than the control). Before testing the samples, panelists were also asked to evaluate the frequency of consumption of each food product by rating it 1 (every day), 2 (once per week or more often), 3 (once every 2 wk), 4 (once per month), or 5 (less than once per month).
Tested compounds The tested compounds were 4 EOs (lemon EO from Citrus lemon L., pennyroyal mint EO from Mentha pulegium L., thyme EO from Thymus algeriensis L., and rosemary EO from Rosmarinus officinalis L.) and 2 ICs present in EOs (carvacrol and p-cymene). Citrus lemon L. EO was obtained as described by Espina and others Statistical analyses (2011), while M. pulegium, T. algeriensis, and R. officinalis EOs were All data regarding age and sex of each 1 of the 65 panelists, obtained as described by Ait-Ouazzou and others (2011b, 2012b). as well as the frequency of consumption of each food product The composition of all these EOs was analyzed and reported in the and the hedonic value assigned to each 1 of the 54 noncontrol S1576 Journal of Food Science r Vol. 79, Nr. 8, 2014
Sensory impact of essential oils in food . . . Table 1–Mean, SD (standard deviation), P50 (median or 50th percentile), P90 (90th percentile), skew. (skewness), IQR (interquartile range), and PG (percentage of panelists rating the sample better than the control) values for each evaluated sample of tomato juice, vegetables soup, or poultry burgers containing 20, 100, or 200 μL/L of lemon EO, pennyroyal mint EO, thyme EO, rosemary EO, carvacrol, or p-cymene. All values are expressed in a 1 to 9 scale where 5.00 is the control score. Asterisks indicate statistically significant difference between the mean value and the control, according to Wilcoxon test (α = 0.05). Lemon EO 20
100
Carvacrol 200
20
Pennyroyal mint EO 20
100
200
Tomato juice Mean 5.03 3.48∗ 3.23∗ 3.97∗ SD 1.68 1.75 1.87 1.72 P50 5.00 3.00 3.00 4.00 P90 7.00 6.00 7.00 6.40 Skew. -0.11 0.56 1.00 0.39 IQR 2.00 2.00 2.00 2.00 PG 40.00 15.38 13.85 20.00
2.75∗ 1.49 2.00 5.00 0.50 3.00 3.01
2.25∗ 5.58∗ 4.22∗ 4.11∗ 3.55∗ 1.51 1.81 2.01 2.22 2.13 2.00 6.00 4.00 4.00 3.00 4.00 8.00 7.00 7.00 6.40 1.17 -0.05 0.42 0.26 0.57 2.00 3.00 3.00 4.00 3.00 4.62 50.77 35.38 30.77 23.08
Vegetables soup Mean 5.60∗ 5.15 4.78 4.91 SD 1.41 1.80 1.87 1.78 P50 6.00 5.00 5.00 5.00 P90 7.40 7.00 7.00 7.00 Skew. 0.17 −0.14 −0.26 −0.47 IQR 1.00 2.50 2.00 2.00 PG 52.31 49.23 41.54 38.46
3.58∗ 1.68 3.00 6.00 0.71 2.00 15.38
2.58∗ 4.26∗ 1.69 1.70 2.00 4.00 5.00 6.00 1.46 0.05 2.50 2.50 7.70 24.61
100
2.89∗ 1.28 3.00 4.00 0.44 2.00 4.62
200
Thyme EO 20
Rosemary EO
100
200
20
100
2.12∗ 1.46 2.00 5.00 1.73 1.00 6.15
2.02∗ 4.68 2.97∗ 1.56 1.99 2.02 1.00 4.00 3.00 5.00 8.00 7.00 1.62 0.40 1.02 2.00 3.00 3.00 6.15 32.31 15.38
200
p-cymene 20
100
200
2.23∗ 4.79 3.48∗ 2.74∗ 1.75 1.96 1.89 1.79 2.00 4.00 3.00 3.00 4.40 8.00 6.00 6.00 1.72 0.21 0.64 1.12 2.00 3.00 2.00 3.00 7.70 36.92 18.46 12.31
2.14∗ 5.06 3.83∗ 3.20∗ 5.38 3.86∗ 3.28∗ 4.98 1.25 1.74 1.71 1.63 1.55 1.81 1.79 1.51 2.00 5.00 4.00 3.00 5.00 4.00 3.00 5.00 4.00 8.00 6.00 6.00 7.00 6.00 6.00 7.00 0.87 −0.04 0.29 0.40 −0.42 0.20 0.29 −0.54 2.00 2.00 2.50 2.00 1.50 2.50 2.00 2.00 1.54 33.84 16.92 13.84 47.69 23.08 18.46 36.92
2.97∗ 1.52 3.00 4.40 0.99 2.00 7.69
2.38∗ 1.37 2.00 4.00 1.43 2.00 3.08
samples, were processed in SPSS Statistics 20.0 (IBM, Corp. Armonk, N.Y., U.S.A.), which was used to perform descriptive statistics and all statistical tests at a significance level of α = 0.05. The Kolmogorov–Smirnov test and observation of the Normal Q–Q plot were applied to determine the normality of the distributions; the Wilcoxon signed-rank test was applied to determine whether differences between each sample and the control sample were statistically significant; the Chi-square test with Monte Carlo simulation was applied to determine whether each food product’s consumption frequency had an influence on the acceptance of added compounds. Chi-square was also used to test other possible correlations, such as that between concentration and spread of the data. Additional tests, such as Kruskal–Wallis with Dunn’s multiple-comparison test, were used to test differences in the SD values between different groups.
Results and Discussion Determination of the hedonic acceptance of food products added with 20, 100, or 200 μL/L of each compound Table 1 contains the descriptive statistics (featuring measurements of centrality and spread) calculated from the hedonic scores shown in a 1 to 9 scale, with 5.00 being attributed to the control sample. Regarding the spread of the data, Table 1 shows the standard deviation (SD, a summary measure of the differences of each score from the total mean score) alongside the interquartile range (IQR, the difference between 75th and 25th percentiles). Taking into account that normally distributed data in a 9-point scale (that is, stanine scale) have a SD of 2, the score variability within each sample could be considered to be moderately low, with 1.76 as the average SD obtained from the 54 different samples.
SD or IQR did not vary as a function of the added compound or the concentration within each evaluated food product (Table S1), although SD values obtained from vegetable soup were significantly less dispersed than those from the other 2 food products, indicating a lower variability in the response to the addition of compounds to vegetable soup. The mean value (arithmetic average of the 65 scores for each sample) and the median (or 50th percentile, value below which half of the panelists scored each sample) are shown in Table 1 to define the centrality of each sample. The 90th percentile and the skewness coefficient, which characterizes the asymmetry of the distribution, are also presented to supply additional information on the acceptance of each sample. Since majority of data could not be fitted to a normal distribution, the nonparametric Wilcoxon test was performed to identify samples statistically different in their mean score from the control; those different means appear marked with an asterisk in Table 1. Furthermore, these mean values statistically different from 5.00 corresponded to those samples whose 50th percentile or median value was other than 5.00, with a few exceptions (tomato juice with 20 μL/L of rosemary EO or 20 μL/L of p-cymene; and poultry burgers with 20 μL/L of lemon EO, 100 μL/L of carvacrol, or 100 μL/L of rosemary EO). In this regard, we decided to use both criteria (Wilcoxon test and median values) to select conditions causing actual improvement or deterioration of each food product. Mean values are also depicted in Figure 1, which shows the average improvement or deterioration after addition of each concentration and compound, in terms of a percentage over or under the control set at 0. This percentage represented the proportion of variation of each sample mean in relation to the original 5.00 score. EOs or ICs are ranked in each figure according to the general acceptance for each food product, as a compendium of the 3 added concentrations. Vol. 79, Nr. 8, 2014 r Journal of Food Science S1577
S: Sensory & Food Quality
Poultry burgers Mean 5.63∗ 4.97 4.80 5.23 4.58 3.60∗ 5.31 4.82 4.09∗ 4.86 4.22∗ 3.45∗ 5.72∗ 4.52 5.00 5.17 4.94 3.43∗ SD 1.40 1.72 1.86 1.41 1.74 1.83 1.79 2.00 2.08 1.92 2.00 1.98 1.57 1.99 2.13 1.80 1.97 1.96 P50 5.00 5.00 5.00 5.00 4.00 3.00 5.00 5.00 4.00 5.00 4.00 3.00 6.00 4.00 5.00 5.00 5.00 3.00 7.00 7.00 7.40 7.00 7.00 6.00 8.00 7.40 7.00 8.00 7.00 6.40 8.00 7.00 8.00 8.00 8.00 7.00 P90 Skew. −0.15 0.18 −0.07 −0.25 −0.03 0.54 0.21 −0.04 0.18 0.09 0.33 0.87 −0.71 0.18 −0.08 −0.26 −0.12 0.72 IQR 2.00 2.00 2.50 2.00 2.00 3.00 3.00 3.50 4.00 2.00 3.00 2.00 2.00 3.00 3.50 2.00 2.00 2.00 PG 49.23 41.54 38.46 44.61 29.23 20.00 46.15 38.46 32.31 30.77 30.77 16.92 60.00 35.38 47.69 38.46 41.54 15.38
Sensory impact of essential oils in food . . . Several conclusions can be drawn from Table 1 and Figure 1. On the one hand, the ranking was different for each food product, meaning that acceptance of a determined EO or IC was dependent upon the food matrix they were added to. On the other hand, for all cases (except for addition of rosemary EO in burgers), increasing concentrations of compounds led to a decrease in the sensory acceptance. For example, tomato juice or vegetable soup obtained higher scores than the control samples when adding 20 μL/L of pennyroyal mint EO or rosemary EO, respectively, but the mean values were reduced by about 15% and 20% when adding 100 μL/L. Furthermore, a noteworthy decrease (by 50% to 60% when compared with the control) was discovered when adding 200 μL/L of rosemary EO, carvacrol, or thyme EO to tomato juice or the same concentration of carvacrol, p-cymene, or pennyroyal mint EO to vegetable soup. Moreover, these poorly appreciated samples were also characterized by strong positively skewed distributions, and by having achieved a maximum score of 2.00 and of 5.00 as evaluated by 50% and 90% of the panelists, respectively.
% variation in relation to control
A
20
Table 2 shows the tolerance limit for each EO or IC in the assayed conditions, defined as the maximum tested concentration whose addition did not modify the hedonic acceptance of the food product. Since for nonnormal distributions, the median is the most appropriate statistic of centrality (Choi 2005), limits were set to the maximum tested concentration that achieved a median value of at least 5.00. In tomato juice, only the lowest concentration of pennyroyal mint EO or lemon EO was tolerated, while the rest of the compounds would be dismissed. In fact, the addition of 20 μL/L of pennyroyal mint EO increased sensory acceptance. Vegetable soup tolerated the incorporation of the lowest concentration of rosemary EO, thyme EO, carvacrol, or p-cymene, but not any concentration of pennyroyal mint EO. However, the hedonic acceptance of vegetable soup increased after addition of 20 μL/L of lemon EO, and was not affected by higher concentrations. Consumers also evaluated positively the addition of 20 μL/L of rosemary EO or lemon EO in burgers, and tolerated higher concentrations: 20 or 100 μL/L of pennyroyal mint EO or p-cymene,
L
P
R
C
T
L
R
T
C
P
M
L
R
M
P
C
T
0 -20 -40 -60 -80
% variation in relation to control
B 20 0 -20 -40 -60 -80
C 20 % variation in relation to control
S: Sensory & Food Quality
M
0 -20 -40 -60 -80
S1578 Journal of Food Science r Vol. 79, Nr. 8, 2014
Figure 1–Percentage of variation of each mean score in comparison with the control sample, corresponding to tomato juice (A), vegetable soup (B), or burger meat (C) infused with 20 μL/L (white bars), 100 μL/L (light gray bars), or 200 μL/L (dark gray bars) of lemon EO (L), pennyroyal mint EO (M), thyme EO (T), rosemary EO (R), carvacrol (C), or p-cymene (P). Striped bars represent significant difference between the mean value and the control, according to the Wilcoxon test (α = 0.05).
Sensory impact of essential oils in food . . .
Determination of the influence of consumption frequency on the acceptance of added compounds In order to facilitate statistical analysis, for each food product, 2 groups of consumers were arranged: a group of frequent consumers (who used to have the product at least once per week, composed of those consumers who had chosen options 1 or 2 in the survey’s question on consumption frequency) and a group of occasional consumers, who had it less often and had answered options 3, 4, or 5. Table 3 shows the mean and median values obtained from both groups of consumers for each product added with any concentration of each EO or IC. According to the analysis of the mean values, there was no influence of the consumption frequency on the general acceptance of the samples, except for burgers incorporated with pennyroyal mint EO, thyme EO or p-cymene, showing significantly higher rates for frequent consumers. Although comparison of medians revealed higher rates in the frequent consumers group for some tomato juice and vegetable soup samples, the low number of frequent consumers of these products led to dismissal of further analyses. When considering concentrations in burgers, Chi-square tests revealed that score differences only happened when adding 200 μL/L of thyme EO or p-cymene, and 100 or 200 μL/L of pennyroyal mint EO (Table S2). Out of these samples, the effect of consumption frequency in the acceptance of samples added with pennyroyal mint EO was remarkable: while frequent consumers of burgers preferred samples with an added 100 or 200 μL/L over the control, 50% of occasional consumers rated them 4 or lower. Additionally, frequent consumers of burgers also ranked samples added with 20 μL/L of thyme EO or with 100 μL/L of p-cymene above the control. Identification of subgroups of consumers with a positive response toward incorporation of EOs or ICs Each range of 65 scores was represented in a histogram (Figure S1) to observe the distribution of data for each condition. Having acknowledged the general lack of normality and regardless of the central tendency, some represented data set seemed to resemble bimodal distributions (distributions with 2 modes), suggesting the presence of 2 different trends within the set of panelists. These distributions also belonged to samples with higher SD values than the average ones, which also concurs with the idea of having 2 different central tendencies. In order to help in the identification of samples provoking this divided response, the proportion of panelists that rated each sample as better than the control was calculated, and it is shown in
Table 2–Tolerance limits in μL/L of each EO or IC for each food product. Limits were defined as the maximum tested concentration whose addition did not modify the hedonic acceptance of the food product, since 50% of the panelists considered the sample with that added concentration at least as good as the control. Limits under 20 μL/L indicate that all the tested concentrations of an EO or IC decreased the hedonic acceptance of the food product.
Lemon EO Carvacrol Pennyroyal mint EO Thyme EO Rosemary EO p-cymene
Tomato juice
Vegetables soup
Poultry burgers
20
