A procedure for sensory detection of androstenone in meat and meat products from entire male pigs: Development of a panel training

Meat Science 122 (2016) 60–67

Contents lists available at ScienceDirect

Meat Science journal homepage: www.elsevier.com/locate/meatsci

A procedure for sensory detection of androstenone in meat and meat products from entire male pigs: Development of a panel training Mª. Dolores Garrido a,⁎, Macarena Egea a, Mª. Belén Linares a, Beatriz Martínez b, Ceferina Viera b, Begoña Rubio b, Francesc Borrisser-Pairó c a b c

Department of Food Science and Technology, Veterinary Faculty, University of Murcia, Espinardo, 30071 Murcia, Spain Estación Tecnológica de la Carne, Instituto Tecnológico Agrario, Junta de Castilla y León, Guijuelo 37770, Spain IRTA-Monells, Product Quality Program, Finca Camps i Armet, E-17121 Monells, Girona, Spain

a r t i c l e

i n f o

Article history: Received 11 December 2015 Received in revised form 22 July 2016 Accepted 25 July 2016 Available online 27 July 2016 Keywords: Boar taint Sensory Training Meat Meat products Androstenone

a b s t r a c t This study represents a proposal for training sensory panels in androstenone (AND) perception in meat and meat products. The procedure consists of four main parts: (1) selection and training of a sensory panel (11 panelists) using standards with Vaseline oil media as carriers of AND and skatole (SKA); (2) developing a training method AND detection in meat; (3) dry cured meat product and (4) cooked meat product. All candidates were able to distinguish between AND, SKA and AND + SKA in Vaseline oil, order AND solutions with different concentrations and classify them in the three categories: low, medium and high. The panel was able to differentiate the meat in the three categories, but only the high level in meat products. Due to the individual features in AND perception, specific training for each type of product is required. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Boar taint is an unpleasant odor mainly associated with the presence of two compounds: skatole (SKA) and androstenone (AND). Skatole (3methylindole) is a metabolite derived from the amino acid tryptophan produced in the lower gut by intestinal bacterial flora, and AND (5α androst-16-en-3-one) is a steroid produced in the testis (Lunde, Skuterud, Egelandsdal, et al., 2010). SKA is perceived by 99% of consumers and is regarded as unpleasant (Weiler, Fischer, Kemmer, Dobrowolski, & Claus, 1997), while the ability to perceive AND varies among subjects, since AND sensitivity is related with the genetic expression of the human odor receptor OR7D4 (Keller, Zhuang, Chi, Vosshall, & Matsunami, 2007). High levels of SKA in pig meat can be effectively reduced by diet and keeping animals free of faecal contamination (Whittington et al., 2011), while AND can be avoided by castration. In most EU countries male pigs are castrated to avoid boar taint in pork (Borrisser-Pairó et al., 2016). However, there is an increasing pressure to seek more humane alternatives to surgical castration, including marketing meat from entire animals with higher levels of AND (Bekaert, Hoerova, & Duca, 2013). Sensory analysis is one of the most common tools used in meat (Egea et al., 2014; Whittington et al., 2011) and meat products ⁎ Corresponding author. E-mail address: [email protected] (M.ªD. Garrido).

http://dx.doi.org/10.1016/j.meatsci.2016.07.019 0309-1740/© 2016 Elsevier Ltd. All rights reserved.

(Bañón, Costa, Gil, & Garrido, 2003; Lunde, Skuterud, Nilsen, & Egelandsdal, 2009) studies. If a trained panel is not able to recognize off-flavors or differences between products, it is assumed that consumers will not perceive them either (Hoehl, Schoenberger, Schwarz, & Busch-Stockfisch, 2013). The detection thresholds and perceived odor quality of AND greatly vary among individuals. Meier-Dinkel et al. (2013) studied how olfactory acuity affects trained assessors' perception of boar taint compounds and found substantial variance among individuals: to highly sensitive panelist could detect at least a ten-fold lower amount of AND (15 ng) on smell strips than less sensitive subjects (150 ng). Annor-Frempong, Nute, Whittington, and Wood (1997) reported that, in a ten member panel, the individual detection thresholds for AND varied from 0.2 to 1.0 μg g−1, adding that the selection and training of sensory assessors was essencial (Meier-Dinkel et al., 2013). Variability in test results can be reduced by selecting the most sensitive candidates, and then training and checking their performance continuously (Romero Del Castillo, Valero, Casanas, & Costell, 2008), for which purpose a key issue is the development of a well designed protocol. In addition, although the perception of AND is limited by its genetic nature, individual threshold differences between persons can be reduced, since previous research has demonstrated that repeated olfactory testing can increase the ability to correctly discriminate AND (Mörlein, Grave, Sharifi, Bücking, & Wicke, 2012). Although some studies have proposed training methods, most of them use various descriptors for AND and SKA (Bañón et al., 2003;

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67

Font-I-Furnols, Guerrero, Serra, Rius, & Oliver, 2000). However, differentiation between the components is more difficult when several attributes are used for the description of boar taint compounds (Dijksterhuis et al., 2000). Indeed, Lunde, Skuterud, Egelandsdal, et al. (2010) demonstrated that it is better to use a low number of attributes for boar taint studies. Furthermore, studies have demonstrated how the carrier influences odor and flavor perception in food products: for example, the studies of Baker and Ross (2014) in red wine, Cherdchu and Chambers (2014) in soy sauce and Lunde et al. (2009) in AND. Differences in boar taint perception between smell strips and fat were found in previous studies (Trautmann, Gertheiss, Wicke, & Mörlein, 2014). Pérez-Juan, Flores, and Toldrá (2008) found that volatile compounds of pork are influenced by binding with muscle proteins and other factors such as salt content and pH. Due to problems with tainted meat in the market place, there is a need to update on knowledge concerning processing opportunities for this raw material (Lunde et al., 2008) and the influence that the meat matrix has on AND perception. Therefore, it might be interesting to train a panel to detect boar taint in different meats and meat products. This paper suggests a protocol for training panels in AND perception for meat and meat products (dry cured and cooked) using a minimum number of descriptors (AND odor, AND flavor). 2. Materials and methods The present study consisted of four main parts (Tables 1 and 2): (Part 1) selection and training of a panel with standards using Vaseline oil media as carriers of AND and SKA; developing a method for training AND detection in (Part 2) meat, (Part 3) dry cured meat products (chorizo sausage) and (Part 4) cooked meat products (frankfurter sausage). All tests were performed in a room equipped with individual tasting booths according to ISO 8589 (2007). There were two sessions per day, one at 10:00 h and second at 15:00. All tests followed the ethical rules and regulations set by the University of Murcia, Spain. 2.1. Selection of sensory panelists Twenty four people, six men and eighteen women (24–55 years), all trained in tasting meat and meat products were tested for their ability to

61

detect AND and SKA crystals in pure form (Bañón et al., 2003). Anosmic candidates were rejected. 2.2. Part 1: Training to detect AND in Vaseline oil standard samples 2.2.1. Standard samples preparation for training For the preparation of standard solutions (Bañón et al., 2003), AND (5α-androst-16-en-3-one,M 272.43 g mol− 1, Sigma-Aldrich A8008, Saint Louis, USA) and SKA (3-methylindole, M 131.17 g mol−1, SigmaAldrich M51458, Saint Louis, USA) were diluted in Vaseline oil (Panreac, Castellar del Vallès, Spain) to obtain 5 mg kg−1 solution, from which working solutions were made. For the different tests, 37 ml of each solution were introduced into amber glass bottles containing 2.5 g of cotton (Scharlau, Sentmenat, Spain). Prior to the training session the bottles were maintained at 35 °C (Bañón et al., 2003) in an oven for 24 h. Samples were coded with random three-digit numbers. Previously, all the bottles were washed with 3 ml of a solution of 4% TWEEN 80 in water and then rinsed 10 times with water, 3 times with deionised water and once with ethyl alcohol (70%). To conclude the cleaning of the bottles, these were heated for 24 h in an oven at 100 °C. 2.2.2. AND and SKA odor description In the first session, odor descriptors and their relationship with AND and SKA were studied using two solutions of AND and SKA in Vaseline oil (Table 1). The AND concentrations used were based on the mean value of the higher levels obtained for entire male pigs in Spanish slaughterhouses (Borrisser-Pairó et al., 2016). The SKA concentrations obtained in the same study were below 0.1 mg kg−1. However, since the SKA perception threshold is 0.1 mg kg−1 (Bañón et al., 2003), it was decided to use a higher dose, 0.5 mg kg− 1, in the initial tests (“AND and SKA odor description” and “Identification of the boar taint compounds” tests) due to the difficulties involved in the differentiation of both products. The panelists were asked to describe the samples using all the terms they needed, unprompted and descriptive (Guerrero, Gou, & Arnau, 1997). 2.2.3. Identification of the boar taint compounds The capacity of candidates deemed sensitive to AND, SKA and AND + SKA was evaluated following the indications of Jellinek (1985) and Galán-Soldevilla and Ruiz Pérez-Cacho (2012). For this purpose

Table 1 Different tests used to develop a method to train panelists to detect AND in vaseline oil. Part

Scope of the experiment

Purpose

Method

Amount of odorant

Number of sessions

Pure crystals

Selection of analytical panelists, 24 potential panelists Training of panelist in AND and SKA detection in vaseline oil

Determination of AND and SKA sensitivity AND and SKA description

Sniff ANDand SKA pure crystalsa

Pure crystals of AND and SKA

1

Identification of the boar taint compounds

Ability to identify androtenone, SKA and AND and SKA in vaseline oilc

Ranking test

Order different solutionsc

Classification test

Three categories were fixed depending on the AND concentration: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1)d

1 Vaseline oil

AND: androstenone; SKA: skatole. a Bañón et al. (2003). b Guerrero et al. ( 1997). c Jellinek (1985), Galán-Soldevilla & Ruiz Pérez-Cacho (2012). d Font-I-Furnols et al. (2003), ISO 6658 (2005).

Free generation

b

−1

1.5 mg kg AND 0.5 mg kg−1 SKA 1.5 mg kg−1 AND 0.5 mg kg−1 SKA 1.5 mg kg−1 AND +0.5 mg kg−1 SKA 0.2, 0.7,1, 1.8 (set 1) mg kg −1 AND 0.4, 0.7, 0.9, 1.2 (set 2) mg kg−1 AND 0.8, 1.2, 1.6, 2.0 (set 3) mg kg−1 AND 0, 0.2, 0.7, 0.8, 1.6, 2.0 mg kg−1 AND (set 1) 0, 0.2, 0.4, 0.7, 1.5, 1.8 mg kg−1 AND (set 2) 0, 0.2, 0.7, 0.9, 1.6, 2 mg kg−1 AND+ 0.1 mg kg−1 SKA (set 3) 0, 0.2, 0.4, 0.7, 1.2, 1.8 mg kg−1 AND+ 0.1 mg kg−1 SKA (set 4)

1 4

4

8

62

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67

Table 2 The different tests used to develop a method to train panelists to detect AND in loin, dry cured meat product (chorizo) and cooked meat product (Frankfurt sausage).

Part

Scope of the experiment

Purpose

Method

Amount of odorant

2 Loin

Fix AND levels

Classification test

Sensory profiling of boar meat loins, 8 panelists

Descriptive test

Three categories were fixed depending on the AND concentration: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1)a Intensity rating (10 cm unstructured line), AND odor and flavorb

Fix AND levels

Classification test

Sensory profiling of boar dry cured meat products, 10 panelists Fix AND levels

Descriptive test

0 mg kg−1 AND+ 0.01 mg kg−1 SKA 0.4 mg kg−1 AND+ 0.04 mg kg−1 SKA 1.5 mg kg−1 AND+ 0.04 mg kg−1 SKA 0 mg kg−1 AND+ 0.02 mg kg−1 SKA 0.4 mg kg−1 AND+ 0.04 mg kg−1 SKA 0.8 mg kg−1 AND+ 0.04 mg kg−1 SKA 1.5 mg kg−1 AND+ 0.04 mg kg−1 SKA 0.4 mg kg−1 AND+ 0.09 mg kg−1 SKA 0 mg kg−1 AND+ 0.08 mg kg−1 SKA 0.4 mg kg−1 AND+ 0.08 mg kg−1 SKA 1.5 mg kg−1 AND+ 0.09 mg kg−1 SKA

3 Dry cured meat product (chorizo sausage)

4 Cooked meat product (frankfurter sausage)

Sensory profiling of boar cooked meat products, 10 panelists

Classification test

Descriptive test

Three categories were fixed depending on the AND concentration: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1)a Intensity rating (10 cm unstructured line), AND odor and flavorb Three categories were fixed depending on the AND concentration: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1)a Intensity rating (10 cm unstructured line), AND odor and flavorb

0 mg kg−1 AND+ 0.03 mg kg−1 SKA 0.4 mg kg−1 AND+ 0.03 mg kg−1 SKA 1.5 mg kg−1 AND+ 0.09 mg kg−1 SKA

Number of sessions 5

4

5

4

5

4

AND: androstenone; SKA: skatole. a Font-I-Furnols et al. (2003), ISO 6658 (2005). b ISO 4121 (2003).

three different solutions were used (Table 1). Twelve bottles were presented to each panelist. In six bottles the contents were known to the panelists (AND, SKA or AND + SKA) and the other six (as replicate) were simply assigned three digit number codes. The participants were instructed to open the bottles and sniff immediately. Coffee seeds were used to “clean the nose” between each sample. Each panelist had to match the bottles with the same content. This test was done twice in one session, and four sessions were carried out. At the end, each panelist had evaluated each solution sixteen times (AND, SKA or AND + SKA). Panelists were considered to have passed the test when they gave at least 70% of correct responses. 2.2.4. Ranking test with Vaseline oil The ranking test was carried out following the indications of Jellinek (1985) and Galán-Soldevilla and Ruiz Pérez-Cacho (2012). Panelists had to sniff and order different AND solutions –prepared from the working solutions– from minor to maximum concentration. In each session, each panelist had to order three sets with different concentrations of AND (Table 1). Four sessions were carried out. The panelist was considered to have passed the test if N 70% of answers were correct in four replicates. 2.2.5. Classification test with vaseline oil Three categories were fixed depending on the AND concentration: low (0–0.3 mg kg− 1), medium (0.4–0.9 mg kg−1) and high (1– 1.8 mg kg−1) (Font-I-Furnols, Gispert, Diestre, & Oliver, 2003). For the classification test (ISO 6658, 2005) panelists had to sniff and put each sample in the correct category (low, medium or high). For this purpose panelists had to classify four different sets: two sets of 6 samples with different concentrations of AND and two sets with AND in different concentrations + 0.1 mg kg− 1 SKA (Table 1). Two sets per session were tested (sets 1 and 2; sets 3 and 4). A total of eight sessions were carried

out, four for each combination of sets. Panelists were considered as “trained” when the percentage of correct responses reached 70%. 2.3. Training to detect AND in meat and meat product samples (Parts 2, 3 and 4) 2.3.1. Animals, meat and meat product samples Entire male pigs with boar taint were pre-selected at slaughterhouses by the human nose methodology (Borrisser-Pairó et al., 2016; Mathur et al., 2012). Then subcutaneous fat samples from the cervical region of the selected boars and castrates were analyzed for AND by gas chromatography–mass spectrometry and for SKA by high-performance liquid chromatography, thus providing carcasses with different levels of boar taint. Results are expressed in fat tissue. Then, meat samples of longissimus thoracis et lumborum muscle from pigs with b 0.1 mg kg−1 SKA and 0.01–2.9 mg kg−1 AND in fat tissue were selected for meat training (Part 2). Loins were frozen 24 h postmortem and stored at − 18 °C for subsequent analyses. For dry cured (chorizo) and cooked (frankfurter) sausages elaboration, meat and fat from the entire male pigs with AND concentrations of 0.5–0.8 μg kg−1 (medium level) and 1.10–2.75 μg kg−1 (high level) and SKA levels below 0.1 μg kg− 1 were used. The samples from castrated pigs had AND concentrations below 0.01 μg kg−1 and SKA concentrations below 0.1 μg kg− 1. Finally, meat product samples were analyzed for AND and for SKA content by the same way. 2.3.2. Preparation of meat samples Loins were thawed at 4 °C for 24 h before assessment. Subcutaneous fat was trimmed to ensure 1 cm thick fat in all samples. No seasoning was added to avoid masking the boar taint, as reported by Lunde et al. (2008). A double-sided grill (Media Liscia; Silanos Lavastoviglie Industriali, Pioltello, Italy) was pre-heated at 170 °C for 5 min, and

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67

both the top and the bottom of the grill were covered with aluminum foil before the steaks were cooked for 6 min, turning them every two minutes, to ensure uniform cooking, until the internal temperature reached 72 °C (portable T200 thermometer; Digitron Instrumentation Ltd., Hertford, United Kingdom). Samples with different levels of AND and b0.1 mg kg−1 SKA were cooked separately and aluminum foil was changed between samples. The samples were trimmed of any external connective tissue and cut into 2 × 2 cm pieces and then wrapped in aluminum foil coded with a random three digit number and stored in a sand bath at 60 °C until tasting. Sample presentation was balanced to account for order and carryover effects (Macfie, Bratchell, Greehoff, & Vallis, 1989). Mineral water and unsalted bread were provided for mouth rinsing between samples (Egea et al., 2014). 2.3.3. Dry cured meat product elaboration and preparation of samples Two batches of the different types of chorizo sausages (high, medium, and low AND level) were produced in a pilot plant according to a traditional formulation, which included 75% pork meat and 25% pork backfat. These were minced (P-32 FUERPLA, Valencia, Spain) to a particle size of about 6 mm and subsequently mixed in a vacuum mixer (A85 FUERPLA, Valencia, Spain) with the following common ingredients: sodium chloride (2%), Vera paprika (2%), garlic powder (0.15%), glucose (0.1%), polyphosphates (0.1%), sodium ascorbate (0.045%), sodium nitrite (0.015%), potassium nitrate (0.01%). The sausage mixture was then stuffed into 35–45 mm natural beef casings. The chorizo sausages were fermented in a drying chamber (Hermekit, Cenfrio, Spain) at 10 °C and 90–100% relative humidity (RH) for 15 h, 20–25 °C and 90% RH for 48 h, at 10–15 °C and 80–90% RH for 7 days, before reducing the RH slowly to 75% until the end of the ripening process. Weight losses reached 38.6 ± 2.1% at the end of ripening. The sausages were vacuumpackaged and stored at 4 °C until analysis. Chorizo sausages at the three AND levels were sliced (0.4 mm), and two slices of each were put in a Petri plate (90 × 14 mm) 2 h before each session. Samples were served at room temperature (23 °C). For the sensory test each panelist had to open the Petri plate and sniff the sample before tasting it. Unsalted bread and mineral water were used between samples. 2.3.4. Cooked meat product elaboration and preparation of samples Three types of frankfurter sausages (low, medium, and high AND level) were made. Two batches of the different types of frankfurter sausage were prepared in a pilot plant following normal industrial procedures. The basic recipe contained lean meat (37.5%), pork fat (37.5%), ice/water (25%), potato starch (2.5%), soybean protein (2%), sodium chloride (2%), kappa carrageenan (0.5%), sodium polyphosphate (0.3%), dextrose (0.25%), sodium ascorbate (0.05%) and sodium nitrite (0.015%). All additives were provided by Proanda S.L. (Sevilla, Spain). For each batch of sausages the ingredients were emulsified using a bowl cutter (CM-41, Mainca, Barcelona, Spain). The batters obtained were vacuum stuffed (Tecmaq Microwat, Barcelona, Spain) into 20 mm collagen casing (NB300, Edicas, Ripoll, Spain) and cooked to a 72 °C core temperature (Verinox Junior 1100, Vigolo Vattaro, Italy). After cooking, the sausages were cooled, vacuum-packaged and kept at 4 °C until the day they were analyzed. Frankfurter sausages from the three AND levels were sliced (10 mm), and two slices of each level were put in a glass Petri plate (50 × 14 mm). The samples were heated in a microwave (4 s, 800 W) until they reached an internal temperature of 72 °C and were served to the panelists immediately. Each panelist opened the Petri plate, sniffed the sample and then tasted it. Unsalted bread and mineral water were used between samples. 2.3.5. Classification test For the meat training (Part 2), the panelists tested four samples containing three different levels of AND in each session. The panelists smelled and tasted samples and then classified them in the three

63

categories. By the end of the test, each panelist had tasted at least six samples from each level. Five sessions were carried out. Panelists were considered as trained when the percentage of correct responses reached 70%. The test was carried out at the same way for dry cured (chorizo) and cooked (frankturter) sausages (Parts 3 and 4). 2.3.6. Descriptive test with meat samples Sensory analysis was carried out according to ISO 4121 (2003). During the first two sessions, each of the panelists evaluated separately five meat samples with known AND concentration (Table 2). Afterwards, the group discussed their findings openly with the intervention of the panel leader as a moderator. The selected descriptors were AND odor and AND flavor as described by Lunde, Skuterud, Egelandsdal, et al. (2010) and meat odor and meat flavor. The intensities were rated on an unstructured 10 cm line (0 = “not perceivable”; 10 = “extremely perceivable”). Meat samples with known concentrations were located by consensus on the unstructured line. This was done twice. In the second step, panelists evaluated five coded samples and made the sensory analysis individually. Again, this was done in duplicate (in two sessions). At the end of each session there was an open discussion modulated by the panel leader in order to improve panel training. When the training was concluded the descriptive test was carried out only with panelists who had passed the training step. Panelists 9, 10 and 11 were not available when loin sensory training was carried out. For dry cured (chorizo) and cooked (frankfurter) sausages (Parts 3 and 4), during the first session, the panelist tested the three reference samples (one from each level: low, medium and high) twice. Instead of meat odor and flavor, dry sausage and frankfurter odor and flavor, respectively, were evaluated. The panelists tested six coded samples from different levels in each session, making the sensory analysis individually. Four sessions were carried out. Panelist 8 was not available when dry cured and cooked product sensory training was carried out. 2.4. Statistical analyses All data were analyzed using the GLM procedure of the SSPS version 19.0 software (SPSS Inc., Chicago, IL) (1997). The statistical model investigated the main effects of AND level, panelist, and associated 2-way interactions. Data for the final sensory analysis of loin, dry sausage and cooked sausages were subjected to one-way ANOVA only for the treatment effect. In addition, when the effect was significant, a Tukey test was used to make pair-wise comparisons. All reported means are least square means, and the significance level was set at P b 0.05. 3. Results and discussion 3.1. Selection of taster and training with standard samples All candidates were sensitive to SKA pure crystals. From this group, there were two men and nine women who were able to perceive pure AND crystals, so these eleven people were recruited to start the training. In Spain Weiler et al. (2000) and Martínez et al. (2014) found that, 31– 34% of people tested were sensitive to AND, with a significantly higher proportion of women. Table 3 shows the results of the terms selected by panelists and their frequencies to describe the AND and SKA odors. The most repeated terms for AND were “sweat”, “stale urine”, “male pig” and “penetrating” and for SKA “naphthalene”, “farm-yard”, “urine”, and “feces-manure”. Font-I-Furnols et al. (2000) used similar terms to describe boar taint in meat samples: “urine”, “androstenone”, “chemical”, “skatole”, “sweat”, “stable”, and “sweet”. In a later study, Dijksterhuis et al. (2000), comparing panels from five different countries, found that AND was related mostly to urine attributes, while SKA was mostly related to manure and, to lesser extent, naphthalene. Although sensory panels in general were able to differentiate between AND and SKA, panels from different countries selected different descriptors with

64

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67

Table 3 Terms selected by panelists to describe androstenone (AND) and skatole (SKA) odors. Assessor 1 Stale urine Pig male Penetraiting Pungent Spicy Sweat Naphthalene Urine Feces-manure Farm-block Dirty

2

Total 3

4

5

6

7

x

8

9

x x

x

10

AND

SKA

x

3 2 3 1 1 4 0 0 0 0 0

0 0 0 0 0 1 6 3 4 6 1

x

x

x

x x •

x •

• • • • •

• •

• • • •

• •

x •

x •

• •

mg kg−1, high: 1–1.8 mg kg−1; repeated tests not included).

11

x

•

Table 5 Results from classification test of meat (loin) with different androstenone levels (low: 0–0.3 mg kg−1, medium: 0.4–0.9

•

Panelist

Classification (%)

1 2 3 4 5 6 7 8

100 100 75 75 75 75 100 100

3.2. AND detection in meat samples

x AND; •SKA.

different frequencies. Lunde, Skuterud, Egelandsdal, et al. (2010) also found a discrepancy in the terms used for AND odor in a study with four different trained panels, making differentiation between the boar taint terms AND and SKA difficult. The same authors showed that by using fewer descriptors, assessors were able to distinguish between AND and SKA to a higher degree. In this study we therefore used “AND odor” and “SKA odor” for the panel training. Table 4 represents the results obtained from the identification test. Panelists that gave N70% correct responses (Galán-Soldevilla & Ruiz Pérez-Cacho, 2012) were able to differentiate AND from SKA. Panelists who did not reach this percentage repeated the test once in an attempt to give the correct answer. Other studies, e.g. Etaio et al. (2010), described how candidates with a success rate of between 60 and 70% could retake the non-passed test once more in order to reach the established 70%. The results obtained in the ranking test are also shown in Table 4. The increased complexity of the test meant that the percentage of correct answers required was lower than in the identification test (Keelan, 2002). The probability of guessing at random is greater in the identification test than in the ranking test. Galán-Soldevilla and Ruiz Pérez-Cacho (2012) considered that a ranking test was passed when candidates gave N50% correct responses. Finally, as regards the classification test, all the panelists gave N70% of correct responses, both for AND and AND + SKA, and were considered as “capable” in the training phase. The percentages of correct responses for AND + SKA were in general lower than for only AND (mean values of correct responses of panel were 75 and 90 for AND + SKA and only AND, respectively). Although the SKA perception threshold in the panel was b0.1 mg kg−1, a slight influence of the presence of SKA presence in the solution on AND perception was observed when both compounds were presented together (AND + SKA) in the same container since the odor perception varied with respect to the AND solution (Annor-Frempong et al., 1997).

The results obtained from the classification test with meat samples are shown in Table 5. Eight of the initial eleven panelists were trained using loin. These eight panelists obtained N 70% of correct responses. The evaluation of panel performance involves analyzing individual differences among assessors. Table 6 presents the results obtained for the descriptive test training. The level effect was significant for the parameters meat odor, AND odor, meat flavor, AND flavor, meaning that so panelists were able to distinguish between different AND levels. For all parameters, the F values were higher for the treatment effect than for the panelist effect or their interaction, so these differences had more statistic weight. In spite of the training received, it is usual to find differences between panelist scores since panelists tend to use different parts of the scale. However the most relevant aspect is that these scores are proportional among themselves (Carbonell, Izquierdo, & Costell, 2002). This was the case for meat odor and meat flavor descriptors in the present study. These differences are sometimes difficult to homogenize due to individual genetic thresholds (Keller et al., 2007). Previous studies (Lunde, Skuterud, Hersleth, & Egelandsdal, 2010; Meier-Dinkel et al., 2013; Trautmann et al., 2014) indicated the relevance of individual thresholds in AND perception. In a work on training panelists for Gilthead Sea Bream, Carbonell et al. (2002) also found a panelist effect, which was explained by individual physiological sensibility. Fig. 1 shows that the terms AND odor and flavor depended on the different AND concentration used (low: 0, medium: 0.4, 0.8, and high: 1.5, 2.9). However for meat odor and meat flavor attributes, the differences between groups were slight when consecutive concentrations were

Table 6 Results from descriptive panel training with meat (loin) with different androstenone (AND) levels (low: 0–0.3 mg kg−1, medium: 0.4–0.9 mg kg−1, high: 1–1.8 mg kg−1), dry cured meat product (chorizo sausage) and cooked meat product (frankfurter sausage) with different androstenone (AND) levels (low: 0–0.3 mg kg−1, medium: 0.4–0.9 mg kg−1, high:1–1.8 mg kg−1). Panelist

Table 4 Percentage of final correct responses for identification, ranking and classification tests by each panelist (Repeated test not included).

Panelist

Identification (%)

Ranking (%)

Classification AND (%)

Classification AND + SKA (%)

1 2 3 4 5 6 7 8 9 10 11

72 72 78 72 81 72 75 78 89 72 75

68 64 67 77 64 50 77 73 71 75 71

81 100 100 100 94 87 100 94 83 71 83

72 75 72 75 75 72 72 75 94 69 81

AND: androstenone; SKA: skatole.

Loin

Dry sausage

Frankfurter sausage

Level

Panelist x level

Attribute

F

P-value F

P-value F

P-value

Meat odor AND odor Meat flavor AND flavor Dry sausage odor AND odor Dry sausage flavor AND flavor Frankfurter sausage odor AND odor Frankfurter sausage flavor AND flavor

4.70 1.55 6.49 1.33 20.35

0.000 0.218 0.001 0.299 0.000

0.004 0.000 0.001 0.000 0.000

0.550 0.428 0.561 0.337 4.51

0.915 0.976 0.906 0.995 0.000

15.34 14.78 9.73 19.16 47.35

7.07 0.000 15.52 0.000

63.61 0.000 117.05 0.000

4.25 3.94

0.000 0.000

2.85 0.004 17.49 0.000

198.17 0.000 97.26 0.000

5.17 3.00

0.000 0.000

8.01 0.000 19.06 0.000

78.53 89.78

0.000 0.000

3.21 2.61

0.000 0.000

1.34

91.56

0.000

2.48

0.001

0.228

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67

a a

a

a b

ab b

bc

ab bc c c

cd c

de

c d

c d

d

Fig. 1. Mean values of descriptive test (10 cm unstructured line) for pork loin with different androstenone (AND) levels: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1) (Font-I-Furnols et al., 2003) evaluated by panelists that passed the training step (n = 8). 0; 0.4; 0.8; 1.5; 2.9: AND concentration in meat samples (mg kg−1 in fat).

analyzed by Tukey's test. Meat odor and flavor seem to decrease when AND odor and flavor increase. It is possible that the volatile compounds responsible for meat odor and flavor were masked by AND. 3.3. AND detection in dry cured meat product (chorizo sausages) Table 7 shows the results of the classification test. Ten panelists were trained for AND detection in dry cured sausage. None of the panelist, except 9 and 11, was able to distinguish between meat samples with a low level, from samples which contained 0.6 mg kg− 1 AND, perhaps because of the volatile aromas of the spices added to dry sausage. Lunde et al. (2008) found that spices such as oregano can mask AND odor and flavor. On the other hand, all the panelists, except 6 and 10, gave N70% of correct responses for the high level of AND (1.5 mg kg−1). Both these panelists were able to detect AND in vaseline oil, but not in the dry cured meat product. Differences in boar taint perception between standards and meat have been found in previous studies (Trautmann et al., 2014). Perceptual interactions due to volatiles and non-volatiles in complex foods have already been discussed with respect to wine for example (Atanasova et al., 2005). One possible explanation is the higher fat content of sausages compared with meat: the addition of fat induces a significant retention of hydrophobic flavor compounds, resulting in noticeable effects on flavor perception (Guichard, 2002). Carrapiso, Jurado, Martín, and García (2007) found

65

in a study of the effect of fat content on flavor release from sausages that increasing the fat content caused a general increase in the volatile compound concentrations of the sausages (fat acted as a reservoir) but a decrease in the compound concentration in the headspace and in the nose space. In addition, varying moisture and fat contents might affect flavor release, as has been shown, for example, in sausages (Chevance & Farmer, 1999; Mörlein, Christensen, & Gertheiss, 2015). Mörlein et al. (2015) also suggested that the volatile compounds held responsible for boar taint (AND, SKA and indole) probably interact with proteins from the connective tissue matrix in back fat. On the other hand, the variability found between panelists in AND sensitivity could also be related with the genetic expression of the human odor receptor OR7D4 (Keller et al., 2007). Table 6 presents the result of the descriptive test. As can be seen, training, level, panelist, and level x panelist interaction had a significant effect (P b 0.01) meaning that panelists were able to detect high AND levels in dry sausages (“chorizo”). Fig. 2.a shows that the interaction could be due to panelists 6 and 10, who did not score the samples as the rest of the panelists did and were finally removed from the test for this kind of product. A descriptive test was repeated only with the selected panelists (Fig. 2.b) the level x panelist interaction disappeared. A Tukey test differentiated between the low and medium groups vs. high group (P b 0.05), which coincide with those found in the dry product training phase (Table 7). Panelists were able to distinguish low concentrations of AND (0–0.6) when the matrix used was vaseline oil or meat, but not when dry sausage was used. These results could be interesting for future studies about masking AND odor in entire male meat products. The literature on panel training in this kind of product is scarce. Some consumer studies, e.g. Chevillon et al. (2010), did not identify differences in preferences between dry sausages elaborated with meat from entire (1 g kg−1 of AND) or female and castrated animals.

Table 7 Results from classification test with dry cured meat product (chorizo sausage) and cooked meat product (frankfurter sausage) samples with different levels (low: 0–0.3 mg kg−1, medium: 0.4–0.9 mg kg−1, high:1–1.8 mg kg−1) of androstenone. Dry cured sausages

Frankfurter sausages

% correct responses

% correct responses

Panelist

Low

Medium

High

Low

Medium

High

1 2 3 4 5 6 7 9 10 11

100 67 67 60 38 70 50 100 28 86

63 53 67 50 45 21 42 78 11 88

70 90 91 80 88 48 100 100 63 100

60 100 46 100 83 38 38 100 33 89

40 100 33 50 80 25 37 86 43 100

50 100 78 75 80 83 75 100 50 88

Fig. 2. Descriptive test (10 cm unstructured line) of dry cured sausages (chorizo) with different androstenone (AND) levels: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1) (Font-I-Furnols et al., 2003) mean values of a: AND flavor of each panelist (n = 10) and b: of the entire panel that passed the training step (n = 8).

66

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67

Bekaert et al. (2013) found that boar taint detection depends on sample temperature. Panelists 9 and 11 were able to distinguish low from medium samples in both dry cured sausages and frankfurter. This fact could be related with their individual thresholds that were probably lower than for the rest of the panel. In this sense, Meier-Dinkel et al. (2013) found that highly sensitive assessors were capable distinguishing AND odor and flavor in low-fat boar loins (1.5 to 2.0 μg of AND per gram of melted back fat), whereas assessors of average sensitivity were less discriminating. Hence, we suggest selecting and training a sensory panel depending on the concentration of AND being used, since the AND perception threshold has an important genetic compound (Keller et al., 2007). Table 6 presents results of the descriptive training with frankfurter sausages. As can be seen panelist, level and level x panelist interaction showed differences (P b 0.01) for frankfurter sausage odor and flavor and AND odor. Panelists could differentiate between low and high levels, but used the scale differently, as can be seen on Fig. 3a. For AND flavor, there was no panelist effect, and the similar way in which the scale was marked can be seen in Fig. 3.b, although panelists 1 and 10 gave similar scores for the three levels. Panelists 1 and 10 did not pass the cooked product training, so they were rejected. In the case of dry cured sausage, the descriptive test was repeated only with panelists who passed the training; the results are shown in Fig. 3.c. The mean values for AND odor were slightly higher than for AND flavor, in contrast to the results obtained for the dry cured product. This could be related with the influence of serving temperature on AND perception (Bekaert et al., 2013; Font-i-Furnols, 2012). Frankfurter sausages were served immediately after being heated to 72 °C, while dry sausages samples were served at 23 °C (room temperature).

4. Conclusion The training procedure developed in this study resulted in a sensory panel able to differentiate meat with three levels of AND (low, medium and high), and meat products (dry cured, chorizo sausage and cooked frankfurter sausages) in high levels of AND when SKA levels were kept low. When the product analyzed involved other factors such as composition, processing and consumption conditions, as in meat products, the perception of AND could vary. In addition, not all panelists passed the specific training in meat products, which shows the individual abilities in the AND perception. So the importance of specific training for individual products is evident.

Acknowledgements

Fig. 3. Descriptive test (10 cm unstructured line) with cooked sausages (Frankfurter) with different androstenone (AND) levels: low (0–0.3 mg kg−1), medium (0.4–0.9 mg kg−1) and high (1–1.8 mg kg−1) (Font-I-Furnols et al., 2003) mean values of a: frankfurter flavor of each panelist, b: AND flavor of each panelist (n = 10) and c: of the entire panel that passed the training step (n = 8).

This study was financially supported by National Institute for Agronomic Research of Spain (INIA): Market potential and quality of meat and meat products from entire males. European perspectives of banning pig castration. BOARMARKET; RTA-2011-00027-C02-01 and by European Union (FEDER). F. Borrisser-Pairó is a recipient of a doctoral fellowship awarded by the INIA (Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria).

3.4. AND detection in cooked meat product (frankfurter sausages)

References

Ten panelists were trained for AND detection in cooked sausages. Table 7 presents the results of the classification tests with frankfurter with different levels of AND (low, medium, high). Only panelists 2, 5, 9 and 11 were able to differentiate low to medium levels. For the high level, only panelists 1 and 10 had b 70% correct responses. This greater range of responses could be related with the fact that frankfurter sausages contain more fat and fewer spices that dry sausages. In addition, the frankfurters sausages were heated at 72 °C and were tasted hot, while dry sausage was consumed at room temperature (23 °C).

Annor-Frempong, I. E., Nute, G. R., Whittington, F. W., & Wood, J. D. (1997). The problem of taint in pork: 1. Detection thresholds and odour profiles of androstenone and skatole in a model system. Meat Science, 46(1), 45–55. Atanasova, B., Thomas-Danguin, T., Chabanet, C., Langlois, D., Nicklaus, S., & Etievant, P. (2005). Perceptual interactions in odour mixtures: Odour quality in binary mixtures of woody and fruity wine odorants. Chemical Senses, 30(3), 209–217. Baker, A. K., & Ross, C. F. (2014). Sensory evaluation of impact of wine matrix on red wine finish: A preliminary study. Journal of Sensory Studies, 29(2), 139–148. Bañón, S., Costa, E., Gil, M. D., & Garrido, M. D. (2003). A comparative study of boar taint in cooked and dry-cured meat. Meat Science, 63(3), 381–388. Bekaert, G., Hoerova, M., & Duca, M. L. (2013). Risk, uncertainty and monetary policy. Journal of Monetary Economics, 60(7), 771–788.

M.ªD. Garrido et al. / Meat Science 122 (2016) 60–67 Borrisser-Pairó, F., Panella-Riera, N., Zammerini, D., Olivares, A., Garrido, M. D., Martínez, B., ... Oliver, M. A. (2016). Prevalence of boar taint in commercial pigs from Spanish farms. Meat Science, 111, 177–182. Carbonell, I., Izquierdo, L., & Costell, E. (2002). Sensory profiling of cooked gilthead sea bream (Sparus aurata): Sensory evaluation procedures and panel training. Food Science and Technology International, 8(3), 169–177. Carrapiso, A. I., Jurado, A., Martín, L., & García, C. (2007). The duration of the outdoor rearing period of pigs influences Iberian ham characteristics. Irish Journal Agricultural Food Research, 46, 105–115. Cherdchu, P., & Chambers, E. (2014). Effect of carriers on descriptive sensory characteristics: A case study with soy sauce. Journal of Sensory Studies, 29(4), 272–284. Chevance, F. F., & Farmer, L. J. (1999). Release of volatile odor compounds from full-fat and reduced-fat frankfurters. Journal of Agricultural and Food Chemistry, 47(12), 5161–5168. Chevillon, P., Bonneau, M., Le Strat, P., Guingand, N., Courboulay, V., Quiniou, N., & Atlantique, C. A. (2010). Acceptabilité par les consommateurs des viandes de porc mâle entier transformées en saucisse, lardon, saucisson sec et jambon cuit. Journées de la Recherche Porcine, 42, 227–228. Dijksterhuis, G. B., Engel, B., Walstra, P., I Furnols, M. F., Agerhem, H., Fischer, K., ... Bonneau, M. (2000). An international study on the importance of androstenone and skatole for boar taint: II. Sensory evaluation by trained panels in seven European countries. Meat Science, 54(3), 261–269. Egea, M., Linares, M. B., Garrido, M. D., Villodre, C., Madrid, J., Orengo, J., ... Hernández, F. (2014). Crude glycerine inclusion in Limousin bull diets: Animal performance, carcass characteristics and meat quality. Meat Science, 98(4), 673–678. Etaio, I., Albisu, M., Ojeda, M., Gil, P. F., Salmerón, J., & Elortondo, F. P. (2010). Sensory quality control for food certification: A case study on wine. Method development. Food Control, 21(4), 533–541. Font-i-Furnols, M. (2012). Consumer studies on sensory acceptability of boar taint: A review. Meat science, 92(4), 319–329. Font-I-Furnols, M. F., Gispert, M., Diestre, A., & Oliver, M. A. (2003). Acceptability of boar meat by consumers depending on their age, gender, culinary habits, and sensitivity and appreciation of androstenone odour. Meat Science, 64(4), 433–440. Font-I-Furnols, M. F. I., Guerrero, L., Serra, X., Rius, M. A., & Oliver, M. A. (2000). Sensory characterization of boar taint in entire male pigs. Journal of Sensory Studies, 15(4), 393–409. Galán-Soldevilla, H., & Ruiz Pérez-Cacho, P. (2012). Panel training programme for the protected designation of origin “Aceituna Aloreña de Malaga”. Grasas y Aceites, 63(1), 109–117. Guerrero, L., Gou, P., & Arnau, J. (1997). Descriptive analysis of toasted almonds: A comparison between expert and semi-trained assessors. Journal of Sensory Studies, 12(1), 39–54. Guichard, E. (2002). Interactions between flavor compounds and food ingredients and their influence on flavor perception. Food Reviews International, 18(1), 49–70. Hoehl, K., Schoenberger, G. U., Schwarz, K., & Busch-Stockfisch, M. (2013). Is perception of sucrose and caffeine affected by training or experience? Monitoring training effects in female subjects over a half year period. Journal of Sensory Studies, 28(1), 1–13. ISO 4121 (2003). Sensory analysis. Guidelines for the use of quantyitative responses scales. Geneva, Switzerland: International Organization for Standardization. ISO 6658 (2005). Sensory analysis. Methodology. General guidance. Geneva, Switzerland: International Organization for Standardization. ISO 8589 (2007). Sensory analysis. General guidance for the design of test rooms. Geneva, Switzerland: International Organization for Standardization. Jellinek, G. (1985). Sensory evaluation of food. Theory and practice. Ellis Horwood Ltd.

67

Keelan, B. (2002). Handbook of image quality: Characterization and prediction. Cap 7. Calibrated psychometrics using quality rulers. New York, United States: CRC Press (Ed:). Keller, A., Zhuang, H., Chi, Q., Vosshall, L. B., & Matsunami, H. (2007). Genetic variation in a human odorant receptor alters odour perception. Nature, 449(7161), 468–472. Lunde, K., Skuterud, E., Egelandsdal, B., I Furnols, M. F., Nute, G. R., Bejerholm, C., ... Hersleth, M. (2010a). The importance of the recruitment method for androstenone sensitivity with respect to accurate sensory evaluation of androstenone tainted meat. Food Quality and Preference, 21(6), 648–654. Lunde, K., Skuterud, E., Hersleth, M., & Egelandsdal, B. (2010b). Norwegian consumers' acceptability of boar tainted meat with different levels of androstenone or skatole as related to their androstenone sensitivity. Meat Science, 86(3), 706–711. Lunde, K., Egelandsdal, B., Choinski, J., Mielnik, M., Flåtten, A., & Kubberød, E. (2008). Marinating as a technology to shift sensory thresholds in ready-to-eat entire male pork meat. Meat Science, 80(4), 1264–1272. Lunde, K., Skuterud, E., Nilsen, A., & Egelandsdal, B. (2009). A new method for differentiating the androstenone sensitivity among consumers. Food Quality and Preference, 20, 304–311. Macfie, H. J., Bratchell, N., Greehoff, K., & Vallis, L. V. (1989). Designs to balance the effect of order of presentation and first order carry over effect in hall tests. Journal of Sensory Studies, 4(2), 129–148. Martínez, B., Rubio, B., Vieira, C., Panella-Riera, N., Egea, M., Linares, M. B., ... Oliver, M. A. (2014). Sensory preference for fermented products manufactured with meat from entire male pigs with different levels of androstenone. 2nd International Symposium on Fermented Meat. Valencia: CSIC Consejo Superior de Investigaciones Científicas. Mathur, P. K., ten Napel, J., Bloemhof, S., Heres, L., Knol, E. F., & Mulder, H. A. (2012). A human nose scoring system for boar taint and its relationship with androstenone and skatole. Meat Science, 91(4), 414–422. Meier-Dinkel, L., Sharifi, A. R., Tholen, E., Frieden, L., Bücking, M., Wicke, M., & Mörlein, D. (2013). Sensory evaluation of boar loins: Trained assessors' olfactory acuity affects the perception of boar taint compounds. Meat Science, 94(1), 19–26. Mörlein, D., Christensen, R. H. B., & Gertheiss, J. (2015). Validation of boar taint detection by sensory quality control: Relationship between sample size and uncertainty of performance indicators. Meat Science, 100, 232–236. Mörlein, D., Grave, A., Sharifi, A. R., Bücking, M., & Wicke, M. (2012). Different scalding techniques do not affect boar taint. Meat Science, 91(4), 435–440. Pérez-Juan, M., Flores, M., & Toldrá, F. (2008). Effect of pork meat proteins on the binding of volatile compounds. Food Chemistry, 108(4), 1226–1233. Romero Del Castillo, R., Valero, J., Casanas, F., & Costell, E. (2008). Training, validation and maintenance of a panel to evaluate the texture of dry beans (Phaseolus vulgaris L.). Journal of Sensory Studies, 23(3), 303–319. Trautmann, J., Gertheiss, J., Wicke, M., & Mörlein, D. (2014). How olfactory acuity affects the sensory assessment of boar fat: A proposal for quantification. Meat Science, 98(2), 255–262. Weiler, U., Fischer, K., Kemmer, H., Dobrowolski, A., & Claus, R. (1997). Influence of androstenone sensitivity on consumer reactions to boar meat. In M., K., & B. (Eds.), Boar taint in entire male pigs: EAAP publication, 92, . Weiler, U., I Furnols, M. F., Fischer, K., Kemmer, H., Oliver, M. A., Gispert, M., ... Claus, R. (2000). Influence of differences in sensitivity of Spanish and German consumers to perceive androstenone on the acceptance of boar meat differing in skatole and androstenone concentrations. Meat Science, 54(3), 297–304. Whittington, F. M., Zammerini, D., Nute, G. R., Baker, A., Hughes, S. I., & Wood, J. D. (2011). Comparison of heating methods and the use of different tissues for sensory assessment of abnormal odours (boar taint) in pig meat. Meat Science, 88(2), 249–255.