Factors affecting boar taint in pigs

Transcription

1 Factors affecting boar taint in pigs Dario Zammerini A dissertation submitted to THE UNIVERSITY OF BRISTOL In accordance with the requirements for award of the degree of Doctor of Philosophy (PhD) In the Faculty of Medical and Veterinary Sciences Department of Clinical Veterinary Sciences Division of Farm Animal Sciences May, ,774

2 Abstract Boar taint is an off odour/flavour that can potentially become manifest when meat and meat products from entire male pigs are consumed. Boar taint is due to high concentrations in the fat of at least one of the main compounds involved, skatole and androstenone. The mechanisms controlling androstenone and skatole formation and deposition in the fat are largely known; however, a definitive solution to reduce boar taint incidence has yet to be found. This work investigates the factors affecting boar taint and looks at possible solution to reduce its incidence in pork. There were no significant differences in the levels of both compounds in the fat from different parts of the animal, suggesting that the cheapest part of the carcass can be used for the measurement of boar taint. This study showed that there is a wide variability in the incidence of boar taint in the British pig population mainly due, it is thought, to differences in the genetics of the pigs and in the managerial aspects of the farms. Growth rate was studied as a possible factor in boar taint variation. It was found that faster growth (>0.7 kg/d) significantly increased concentrations of both taint compounds giving no support to the view that the older, slowest growing pigs are likely to be more tainted. In confirmation of earlier studies, it was found that skatole has the biggest impact scores for abnormal odour and flavour and that skatole enhances the sensory perception of androstenone. So it is possible that reducing skatole alone may be an effective way to control boar taint. An effective way to reduce skatole levels in the fat is the use of diets rich in fibre or in fermentable carbohydrates. In this study feeding chicory (Cichorium intybus L.), a source of a fermentable carbohydrate know as inulin, had a significant effect in reducing skatole levels in the fat. Chicory fed at the level of 9% for 2 weeks was successful in reducing skatole to a level well below the threshold for this compound (0.2µg/g), with only 1 pig with a skatole value over the threshold and 55% with levels typical of castrated pigs. Another solution to the boar taint problem could be the exclusion of tainted carcasses on the slaughterline. The most efficient cooking methods to detect boar taint are microwave and hotwire; however, the wide variability between assessors limits the use of these heating methods as standard tests.

3 Ai miei genitori, Gianfranco e Veronica. Per il loro infinito amore e appoggio.

4 Acknowledgments Firstly, I would like to acknowledge Prof. Jeff Wood for his valuable advice and guidance thorough these three years and for his faith in me without which this dissertation would not have been accomplished. A special thank to Fran Whittington for all the help, time and patience offered to this work and my academic development. Special thanks to those people who helped me during these years and collaborated during the analytical procedures Dr. Peter Sheard, Kathy Hallet, Geoff Nute and Sue Hughes. I am very grateful to BPEX for the financial support. Finally I want to thank all my family for their support during these three years away from home and all the friends (it is impossible to name you all) that I met here in the UK who made this experience unique.

5 I declare that the work in this dissertation was carried out in accordance with the requirements of the University s Regulations and Code of Practice for Research Degree Programmes and that it has not been submitted for any other academic award. Except where indicated by specific reference in the text, the work is the candidate s own work. Work done in collaboration with, or with the assistance of, others, is indicated as such. Any views expressed in the dissertation are those of the author. SIGNED:... Date:...

6 Index List of Figures... ix List of Tables... xii 1 General Introduction Literature review Boar taint Contribution of compounds to boar taint Androstenone Skatole Sensory perception of boar taint Control of boar taint Rearing and management procedures, environments Breeds and genetics Surgical castration Immunocastration Effect of dietary composition Sugar beet pulp Raw potato starch Inulin and chicory roots Methods for detection of boar taint Online detection on carcasses Laboratory techniques Androstenone and skatole levels in pigs on British farms Summary Introduction Materials and methods Results and discussion Conclusions Androstenone and skatole levels in different fat tissues Summary Introduction Materials and methods Results and discussion Conclusions vi

7 5 Effects of pig growth rate and health status on boar taint and meat eating quality Summary Introduction Materials and methods Experiment Experiment Results and discussion Experiment Experiment Conclusions Effect of dietary chicory on boar taint Summary Introduction Materials and methods Preliminary study First feeding trial Main feeding trial Results and discussion Preliminary study First feeding trial Main feeding trial Conclusions Comparison of heating methods for sensory assessment of boar taint in pig meat Summary Introduction Materials and methods Heating methods used for sensory assessment of fat samples Microwave Test Melting test Hotwire test Boiling test Composite samples Sensory assessment Results and discussion vii

8 7.5 Conclusions General conclusions Appendix Determination of skatole and indole concentration in fat Determination of androstenone in fat Appendix 2. Tables and figures Androstenone and skatole levels in pigs on British farms Androstenone and skatole levels in different fat tissues Effect of dietary chicory on boar taint Comparison of heating methods for sensory assessment of boar taint in pig meat Appendix 3. Published papers References viii

9 List of Figures Figure 3-1. Concentrations of skatole and androstenone in the 63 farms Figure 3-2. Skatole and androstenone levels (μg/g) in the 63 farms Figure 3-3. Distribution and correlation between skatole and androstenone levels (μg/g) in the 33 farms of Abattoir Figure 3-4. Distribution and correlation between androstenone levels (μg/g) and carcass weight (kg) in the 63 farms Figure 4-1. Concentrations of skatole in the 20 pigs Figure 4-2. Concentrations of androstenone in the 20 pigs Figure 4-3. Linear regression (red line) of skatole (μg/g) in backfat and neck along with the line of equality (black line) Figure 4-4. Linear regression (blue line) of androstenone (μg/g) in backfat and neck along with the line of equality (black line) Figure 4-5. The plots of differences between skatole levels from backfat and neck Figure 4-6. The plots of differences between androstenone levels from backfat and neck Figure 5-1. Levels of taint compounds in the 2 pulls Figure 5-2. Levels of skatole in the 4 farms Figure 5-3. Levels of androstenone in the 4 farms Figure 6-1. Skatole and androstenone concentrations in the 30 farms Figure 6-2. Skatole and androstenone concentrations in the 13 farms Figure 6-3. Skatole levels in the 4 groups during the 3 weeks of sampling (horizontal line representing the threshold level) ix

10 Figure 6-4. Distribution of skatole concentrations (µg/g of fat) in the 120 pigs fed different levels of chicory at 2 weeks Figure 6-5. Comparison of skatole levels (µg/g of fat) of all control pigs from week 0 with those fed 9% chicory for 2 weeks Figure 6-6. Concentrations of androstenone in the groups fed 9% chicory for 0, 1 and 2 weeks Figure 6-7. Dendrogram with word linkage and correlation coefficient distance. 122 Figure 1-1. Classification of samples into 4 groups differing in levels of taint compounds (µg/g fat) Figure 7-2. Effect of adding 10% each of muscle and salivary gland tissue on skatole concentration in composite sample expressed as a proportion of the backfat value Figure 7-3. Effect of adding 10% each of muscle and salivary gland tissue on androstenone concentration in composite sample expressed as a proportion of the backfat value Figure 7-4. Relationships between abnormal odour score and skatole concentration in backfat in the different heating methods Figure 7-5. Effect of sample type on the inverse relationship between pork odour and abnormal odour using the microwave test (40 samples) Figure 1-6. Charts representing the percentage classification of samples into degrees of undesirability for backfat using the microwave method of cooking (HA = 1.0 µg, HS = 0.20 µg, LA = 0.99 µg, LS = µg) Figure 1-7. Charts representing the percentage classification of samples into degrees of undesirability for composite samples using the microwave method of cooking (HA = 1.0 µg, HS = 0.20 µg, LA = 0.99 µg, LS = µg) x

11 Figure Linear regression (red line) of skatole (μg/g) in backfat and cheek along with the line of equality (black line) Figure Linear regression (red line) of skatole (μg/g) in neck and cheek along with the line of equality (black line) Figure Linear regression (blue line) of androstenone (μg/g) in backfat and cheek along with the line of equality (black line) Figure Linear regression (blue line) of androstenone (μg/g) in neck and cheek along with the line of equality (black line) Figure The plots of differences between skatole levels from backfat and cheek Figure The plots of differences between skatole levels from neck and cheek Figure The plots of differences between androstenone levels from backfat and cheek Figure The plots of differences between androstenone levels from neck and cheek xi

12 List of Tables Table 3-1. Concentrations of taint compounds in the 63 farms (µg/g fat) Table 3-2. The correlation of taint compounds with carcass weight Table 3-3. The correlation of taint compounds with P Table 3-4. Differences in taint compounds, weight and P2 between abattoirs Table 4-1. Concentration of the taint compounds in the 20 pigs (µg/g fat) Table 4-2. The correlation of taint compounds with carcass weight Table 4-3. The correlation of taint compounds with P Table 4-4. The effect of the site of adipose tissues on taint compounds Table 4-5. The correlation of taint compounds between backfat and fat from the neck Table 4-6. The correlation of taint compounds between backfat and fat from the cheek Table 4-7. The correlation of taint compounds between fat from the neck and the cheek Table 4-8. Descriptive statistics for differences (backfat - neck) and crude 95% limits of agreement (n=19) Table 4-9. Descriptive statistics for differences (backfat - cheek) and crude 95% limits of agreement (n=19) Table Descriptive statistics for differences (neck - cheek) and crude 95% limits of agreement (n=19) Table 5-1. Growth rate and P 2 fat thickness in growth category groups Table 5-2. Differences between the 2 sexes (all 225 pigs) in growth rate and P 2 fat thickness xii

13 Table 5-3. Differences between the 2 sexes in growth rate and P 2 fat thickness in the Fast group Table 5-4. Effects of growth category on meat quality in the loin muscle Table 5-5. Effects of growth category on meat quality in the loin muscle in 90 and 110 kg carcass weight groups Table 5-6. Effects of growth category on skatole and androstenone concentrations in males Table 5-7. Effects of carcass weight group on skatole and androstenone concentrations in males Table 5-8. Effects of growth category on eating quality of griddled loin steaks (1 to 8 scales) in all pigs Table 5-9. Eating quality of griddled loin steaks (1 to 8 scales). Comparison of male and female pigs Table Concentrations of taint compounds in the 30 farms (µg/g fat) Table 6-2. Effect of feeding chicory on skatole levels (µg/g) Table 6-3. Sensory results after 2 weeks Table 7-1. Concentrations of taint compounds in backfat of 120 entire male pigs (µg/g fat) Table 7-2. Concentrations of taint compounds present in backfat (mean µg/g fat ± standard deviation) Table 7-3. Concentrations of skatole, androstenone, androstene-α-ol and androstene-β-ol in composite samples containing fat, muscle and sub-maxillary salivary gland in proportions of 8:1: xiii

14 Table 7-4. Scores for backfat odours in HS/HA and LS/LA using different heating methods (values with different superscripts on the same row within method are significantly different, P < 0.05) Table 7-5. Scores for backfat odours in HS/LA and LS/HA using different heating methods (values with different superscripts on the same row within method are significantly different, P < 0.05) Table 7-6. Scores for backfat odours in the 4 taint groups using the hotwire method (means derived from analysis of variance with method and assessor as factors, with 10 replicates) Table 7-7. Correlations between odour scores and boar taint compounds Table 7-8. Effect of adding 10% muscle and 10% salivary gland to 80% head fat (Comp) compared with back fat (BF) on odour scores for the microwave test Table 7-9. Effect of adding 10% muscle and 10% salivary gland to 80% head fat (Comp) compared with backfat (BF) on odour scores for the Boiling test 75 0 C Table Scores for highly undesirable in HS/HA and negative in LS/LA for the 3 assessors when evaluating backfat Table Concentrations of taint compounds in the 30 farms from A Table Concentrations of taint compounds in the 33 farms from A xiv

15 1 General Introduction Rearing entire male pigs for meat production is more profitable than the use of castrated males. However, the use of entire males is limited in most of European countries because of the problem known as boar taint. Boar taint is an unpleasant odour and flavour which is due to high concentrations in the fat of androstenone and skatole. The poor eating experience linked with boar taint has subsequent economic implication for the pork industry but it is still not clear how big it is the impact on the market. In the UK no males are castrated at all, mainly because of the use of lighter animals compared to other European countries, in fact, although it is commonly accepted that light weights can reduce the risk of boar taint this has not been clearly demonstrated and the use of boars also limits the exportation of pork and pork products to other European countries. Since the isolation of the two main compounds, the scientific community has concentrated its attention on the mechanisms regulating the development and accumulation of androstenone and skatole. Despite the fact that in the last thirty years much progress has been achieved on the understanding of boar taint, this problem has not been solved yet. This study analysed the impact of boar taint in the British pig population, testing the levels of these compounds in commercial farms. It investigated different factors that could have a significant effect on the final levels in the fat of boar taint compounds, and ways to control it ante- and post-mortem. 15

16 Literature review 2 Boar taint Rearing entire males instead of castrated male pigs for meat production is more profitable due to their enhanced feed conversion and higher proportion of lean meat in the carcase (Walstra 1974). Entire males also have the advantage of a lower output of nitrogen to the environment (Desmoulin et al., 1974) and reduction of suffering for the animal associated with the practice of surgical castration. However, non-castration is associated with social stress and fighting, resulting in skin lesions and ultimately carcase damage. It is also recognised that the quality of meat from some entire males is influenced by its off odour and taste, better known as boar taint, which many individual consumers find unpleasant. Boar taint is a distinctive and unpleasant taint perceived through a combination of sensory odour, flavour and taste in pork and pork products during cooking and eating. It has been described as animal, urine, faecal and/or sweat like in character. It is mainly caused by accumulation in the fat of at least one of the two compounds, androstenone or skatole, and to a much lesser degree by accumulation of indole and other compounds. According to current European legislation (EU Regulation 854/2004) carcases with a pronounced sexual odour should not be considered fit for human consumption, although there are practical difficulties in detecting taint in pig carcases on slaughter lines. The rearing of entire male pigs is avoided in most countries as castrating male pigs reduces, if not eliminates, boar taint. However, animal welfare concerns are increasing the pressure on pig producers to stop castration. 16

17 Castration of male pigs has been generally abandoned in a number of countries including the UK, Ireland and Australia, and has been partially abandoned in Portugal and Spain, where it is mostly practised in pigs destined for the production of high quality cured products. In Denmark, about 5% of males are left entire and the carcases are tested for skatole levels. Since the isolation in the late 60s of what are considered the two main compounds of boar taint, numerous studies have investigated the causes and factors affecting its development. The scientific community has concentrated its attention on the ways androstenone and skatole are accumulated, with the aim to understand how to control, reduce and prevent the problem of boar taint. However, until the present day, there is not a perfect solution that it is able to guarantee the absence of boar taint in pork, keeping at the same time the economic advantages of rearing entire males and avoiding welfare problems. 2.1 Contribution of compounds to boar taint Two main compounds have been found to be responsible for boar taint. Androstenone (5α-androst-16-ene-3one), a pheromone produced in the testis and exhibiting a urine-like and perspiration odour, was isolated from boar fat by Patterson (1968). Skatole (3-methylindole), a breakdown product of the amino acid tryptophan in the large intestine, exhibiting a faecal-like and naphthalene odour, was isolated from boar fat by Vold (1970). While in numerous subsequent studies it has been confirmed that androstenone and skatole are the main contributors to boar taint other substances with an off-odour/off-flavour have been identified in boar fat that may contribute to a minor degree. Among these, García-Regueiro and Diaz (1989) found significant the presence of indole and other 16-androstene 17

18 steroids. Indole (2,3-benzopyrrole), like skatole, is formed in the large intestine of the pig by intestinal flora. Because of the common pathway, indole is generally accepted to be involved in boar taint even if its smell is not strong and easily detected as in the case of skatole (Annor-Frempong et al., 1997a; Angels Rius et al., 2005). These authors found that indole was associated with boar taint perception but its contribution was minor compared to skatole. Opposite results have been found by Wiseman et al. (1999) with indole not being significantly correlated with boar taint. Angels Rius et al. (2005), using a trained test panel, identified also aldehydes and short chain fatty acids, which are common cause of undesirable aromas in food products, as the main classes of substances in fat samples classified as tainted in their study but with low levels of androstenone and skatole. However, the concentrations of most of these substances in tainted samples compared to untainted ones were not significant. In this work the presence of two contaminants, styrene and 1,4 dichlorobenzene, showed a high concentration in samples classified with boar taint, suggesting that these compounds could have been responsible for the development of some off-flavours. In a precedent study by Angels Rius Solé and Garcia-Regueiro (2001) 4-phenyl-3-buten-2-one was the main compound identified in samples low in androstenone and skatole classified as tainted. Its presence could contribute to boar taint, promoting the perception of androstenone and skatole. In addition to androstenone, other C 19-Δ16-steroids have been isolated by Claus (1979) and considered to contribute to boar taint to a lesser degree: 5βandrost-16-ene-3one, α-androstenol (5α-androst-16-en-3α-ol) and β-androstenol (5α-androst-16-en-3β-ol). Although all these other compounds appear to be less important to the perception of boar taint compared to androstenone and skatole, they probably contribute significantly to the global odour and taste of pork. 18

19 Overall, it is accepted that androstenone and/or skatole are the main contributors to the problem of boar taint as has been confirmed in numerous studies; however no common results have been reached on the respective importance of them to boar taint. There are some publications that underline the importance of the level of androstenone for tainted pork, others that suggest skatole as the main contributor and finally other studies found that there are interactions between skatole and androstenone that affect the sensory response. Squires et al. (1992) and Babol et al. (1996) concluded that androstenone is a better indicator of boar taint, as its perception in tainted meat assessed by a taste panel was correlated to high levels of androstenone. In their review also Claus et al. (1994) considered androstenone the most important cause of boar taint, an idea that was supported consequently by the work of Wiseman et al. (1999). However, not all the studies support the view of androstenone as the main compound of boar taint. Different papers show the opposite result with skatole as the main contributor (Lundström et al., 1988; Bejerholm and Barton Gade, 1993). In a consumer survey carried out in seven countries Matthews et al. (2000) found that a high level of skatole has a bigger impact on consumer acceptability compare to androstenone. It also seems that there are interactions between skatole and androstenone that affect the sensory response. For example, skatole has been demonstrated to enhance the sensory perception of androstenone (Annor- Frempong et al. 1997b), supporting the importance of both compounds in boar taint perception. However, it is clear that meats from animals that have elevated concentrations of both these compounds is associated with an unpleasant experience for the 19

20 consumer. The lack of consistency between the results obtained in the various studies leads to a controversial situation that is damaging for the pig industry, limiting free trade in pig meat between countries, and leaving the research community, that looks at ways to reduce the boar taint problem or to find a valid method of detection of boar taint on the slaughter line, without a clear objective for studies Androstenone Androstenone or 5α-androst-16-ene-3one is a male sex steroid pheromone, exhibiting a urine-like odour, which is produced in the Leydig cells of the testes in parallel with biosynthesis of other steroids during steroidogenesis (Claus and Hoffmann, 1980), and it belongs to the group of compounds termed androstenes. Androstenone biosynthesis is controlled by the neuro-endocrine GnRH-LH axis (Claus et al., 1983). The release of the gonadodropin-realising hormone (GnRH) from the hypothalamus induces the pituitary gland to secrete follicle stimulating hormone (FSH) and luteinizing hormone (LH). The rise in LH levels induces spermatogenesis, with a resulting synthesis of androgens, oestrogens and androstenes. The activation of the GnRH-LH axis mainly occurs during the pubertal development of the pig, which explains the increased levels of androstenone at the onset of the puberty stage in boars. Androstenone is transported via the blood stream to the target tissue, the submaxillary salivary glands (Gower, 1972), where it binds to a specific binding protein, pheromaxein (Booth, 1984). The primary function of pheromaxein is to trap the pheromonal steroids from the blood and to transport them in the aqueous medium of saliva, since the androstenes are lipophilic. But in the salivary glands 20

21 portions of androstenone is converted to α-androstenol and to a lesser extend to β- androstenol (Claus and Hoffmann, 1980). When a mature boar is aroused by the presence of female pigs, it produces a copious amount of saliva that it is also deposited in the environment by the rubbing action of the boar s snout. Booth (1987) discovered a direct effect of the pheromonal steroids emanating from the saliva foam on the boar s mouth on female pigs as well as an effect over time due to saliva deposited in the environment, due to binding with pheromaxein. Different studies showed that the odour of androstenone and the other androstenes facilitate the adoption of the mating stand in oestrous pigs (Melrose et al., 1971; Perry et al., 1980), and Mattioli et al. (1986) also found that the smell of androstenone elicits oxytocin release in sows in oestrous. The androstenone that is not accumulated in the salivary glands is catabolised by the liver. The enzyme 3β-hydroxysteroid dehydrogenase (3β-HSD), also responsible of the production of androstenone in the testis, is responsible for the oxidation of androstenone, the first step in its hepatic catabolism (Doran et al., 2004). The resulting products of the first phase, mainly β-androstenol, are then conjugated by sulfotransferases (SULTs) along with other specific enzymes, increasing the water solubility of the steroids allowing their excretion through the kidney (Sinclair et al., 2005 and 2006). The variation in androstenone levels in adipose tissue between pigs could be due to differences in the biosynthesis or in the catabolism of the steroid in the liver, linked to a low activity and/or expression of the enzymes controlling them. It is still not clear what is more important. Babol and Squires (1999) found no significant relationship between the oxidative metabolism of androstenone in the liver and the levels of androstenone in fat. 21

22 Other authors (Doran et al., 2004; Nicolau-Solano et al., 2006) found that an excessive accumulation of androstenone in the adipose tissue is related to the level of expression of hepatic 3β-HSD. It has been demonstrated that the expression of some these metabolizing enzymes is breed specific. Doran et al. (2004) observed a lower expression of 3β-HSD in liver microsomes from Meishan pigs with high androstenone levels compared with Large White pigs with low androstenone levels. Moe et al. (2007) found a reduced expression of the same enzyme in Norwegian Landrace with high levels of androstenone while in Duroc boars high levels of androstenone were linked to an inhibited expression of the sulfotransferases. It is generally accepted that a high production rate of androstenone will lead to an accumulation in the adipose tissue because of the incapacity of the liver to metabolise all the androstenone. The androstenone which has not been metabolized in the liver is easily transferred from plasma to adipose tissue, inside the adipocytes due to its lipophilic structure. Andresen (1976) found that high concentrations of androstenone in the fat are linked to the level of the steroid in peripheral plasma over the value of 15ng/ml, plasma level which has later been confirmed by Sinclair et al. (2001). However, Malmfors et al. (1976) showed a dynamic relationship between the levels of androstenone in the plasma and the concentration of androstenone in the fat, consequently when the concentration of androstenone in peripheral blood decreases from high levels, the concentration in adipose tissue also gradually decreases. It is not possible to define a cut-off level for androstenone in the fat for sorting the carcasses, due to the complexity of boar taint perception and the resulting 22

23 discrepancy between studies. The most commonly used cut-off levels to sort out tainted meat are 0.5 μg and 1.0 μg of androstenone per g of fat, as respectively suggested in the studies of Claus et al. (1994) and Rhodes (1971). A big European study (Walstra et al., 1999), that involved 6 countries, showed that at usual slaughter weights fat androstenone levels are very variable between animals. These authors measured androstenone levels by chemical analysis in 4313 entire males and found that 29% of them where over the threshold of 1.0 μg/g of fat. This study also showed the variability of androstenone concentrations in different pig populations (6 countries x 2 seasons), as the proportion of entire males with androstenone levels higher than 1.0 μg/g of fat varied from 12% to 43%. As we could expect this variability is also present in the UK, in the same study 400 pigs were analysed and 22.3% of the samples were over the threshold; results that are in contrast with a precedent survey in the UK by Babol and Squires (1995) in which the percentage of boars over the threshold of 1.0 μg/g was 10%. Fat androstenone levels are mostly affected by genetic factors controlling its production and excretion as well as by the degree of sexual maturity. It is well known that the peack in the production of sexual steroids occurs at different ages in pigs due to the variability in the sexual maturation, whereas Bonneau at al. (1987a, b) demonstrated that the genetic potential for androstenone production is responsible for the magnitude of the elevation in the single individual. A boar to exhibit high androstenone level at the commercial slaughter weight needs a combination of early sexual maturation and high potential for androstenone production. It is commonly accepted that increasing slaughter weight results in animals being more sexually mature and consequently presenting high levels of 23

24 androstenone. However, Walstra et al. (1999) showed in their study a low correlation between fat androstenone concentration and carcass weight. Indeed, when we compare animals that have reached sexual maturity, the individual variability in potential synthesis and elimination of androstenone plays a central role in the final fat androstenone concentration. Fat androstenone levels are not greatly influenced by management factors affecting the sexual maturation of the pigs. However, birth to slaughter systems result to be the best way to breed pigs to obtain lower androstenone levels (Fredriksen et al., 2003). Different studies showed that the fighting associated with the mixing of unfamiliar pigs; the formation of new hierarchy and the food competition are responsible of increasing the androstenone levels in plasma and fat (Claus et al., 1994; Giersing et al., 2000). Other management aspects could influence androstenone levels in a limited way; Andersonn et al. (1998) studied the effect of the photoperiod, decreasing the day length, with the result of an earlier sexual maturity and consequently slightly higher androstenone levels. While Øverland et al. (1995) and Zeng et al. (2002) demonstrated that feeding a high energy diet leads to an acceleration of the pubertal development and increases androstenone levels Skatole The indolic compound skatole or 3-methylindole is a breakdown product of the amino acid tryptophan produced in the large intestine of the pig by intestinal bacterial flora (Walstra and Maarse, 1970; Jensen et al., 1995a), and it is commonly described as having a faecal-like odour. These with other studies have shown that the bacterial metabolism of tryptophan in the pig large intestine mainly lead to the production of two volatile lipophilic compounds, indole and skatole; with a third 24

25 product of the tryptophan metabolism, indolic acetic acid (IAA) that is main precursor of skatole in the hind guts of pigs (Knarreborg et al., 2002). While many types of intestinal bacteria are responsible of the conversion of tryptophan to indole and IAA, the strains of only two of the genera containing common intestinal bacteria, the genera Clostridium and Lactobacillus, are capable of further degradation of IAA to skatole (Jensen and Jensen, 1998). Yokoyama and Carlson (1981) isolated from the bovine rumen a Lactobacillus strain able to produces 3- methylindole by decarboxylating IAA, but is not able to form 3-methylindole directly from tryptophan. During the years different bacteria have been found to produce 3-methylindole: a strain of Lactobacillus helveticus (Kowalewska et al., 1985), Clostridium scatologenes, and Clostridium nauseum (Fellers and Clough, 1925; Rosenberger, 1959). In contrast with these bacteria Jensen et al (1995a) found that C. scatologenes DSM 757 was able to generate 3-methylindole from tryptophan, but Deslandes et al. (2001) concluded that the organism producing skatole and causing boar taint in pigs is Lactobacillus sp. strain Skatole is normally not detected in the stomach or small intestine and the skatole-producing bacteria are normally present in the colon of the pigs but they represent less than 0.01% of the total intestinal microflora (Jensen and Jensen, 1993). These bacteria to produce skatole mainly utilise the tryptophan originates from the diet, but also the one that became available with the degradation of the intestinal mucosa. This mucosa is characterized by a very high turnover and the resulting cell debris is a source of tryptophan for the IAA-producing bacteria (Claus et al. 1994). Agergaard and Jensen (1993) found no differences in skatole production in the gut between sexes. 25

26 The skatole produced in the gut can remain in the intestine, can be excreted through faeces or it can be metabolised in the liver and the degeneration products, as androstenone, are excreted with the urine (Claus et al., 1994). Frydman et al. (1972) found that the skatole that is not excreted with the faeces is rapidly absorbed by the intestinal mucosa by the venous blood vessels and then it circulates through the peripheral blood stream with a half life of approximately 60 minutes (Agergaard and Laue, 1993). Even if Agergaard and Laue (1998) have demonstrated that the liver is potentially capable to extract skatole from blood in quantities that greatly exceed what is found under physiological conditions, in some boars a proportion of skatole is accumulated in the adipose tissue because the liver is not able to metabolise all of it. The reason must be found in the influence of the sex steroids, androstenone in particular, on the skatole metabolism in the liver. As for androstenone, skatole metabolism in the liver reckon on two phases, the oxidation and the conjugation. The enzymes cytochrome P4502E1 (CYP2E1) (Babol et al, 1998a) and the cytochrome P4502A6 (CYP2A6) (Diaz and Squires, 2000) are the main enzymes in the first step of the skatole metabolism; while Babol et al. (1998b) found that the second stage is regulated by the sulfotransferease A (SULT1 A), a hydroxisteroid sulfotransferases (HST). As the activities of these key enzymes are positively correlated with the skatole metabolism, a pig producing high levels of these liver enzymes will present low level of skatole in the fat. High levels of skatole in the fat can be related to high levels of androstenone, due to its effect on skatole metabolism. Babol et al. (1999) demonstrated the competitive inhibition of androstenone on the formation of some skatole metabolites in liver microsomes. This antagonising effect of androstenone was then confirmed by Doran et al. (2002) studying primary cultured hepatocytes. These authors found 26

27 that while skatole induced the expression of the protein CYP2E1, androstenone antagonised this effect with a consequent accumulation of skatole in the fat due to a reduced metabolism. This was later explained by Tambyrajah et al. (2004) whom found that CYP2E1 promoter is activated by the transcription factors COUP-TF1 and HNF-1 and that the promoter activity is decreased by androstenone, which inhibit the binding of COUP-TF1 to the promoter. The inhibiting effect of testicular steroids on the hepatic metabolism of skatole was further supported by other studies. In Whittingotn at al. (2004) study, the decreased levels of sex steroids, in particular androstenone, following castration was associated with an increased expression of P4502E1 with a consequent reduction in skatole levels in the fat. The increase in testicular steroid concentration initiates an increase in skatole levels at young age (Zamaratskaia et al., 2004) and there is a positive correlation between the levels of free oestron in fat and skatole level (Zamaratskaia et al., 2005a). Skatole, like androstenone, also seems to be easily transferred from plasma to adipose tissue (Babol and Squire, 1999) and there is a correlation between plasma and fat levels of skatole (Tuomola et al., 1996; Agergaard and Laue, 1998; Zamaratskaia et al., 2004), but it is yet not completely understood which is the physiological function of skatole and if it has any target tissues. Skatole has a toxic effect on many microorganisms; it has a bacteriostatic effect on gram-negative enterobacteria (Tittsler et al., 1935) and it is toxic for rumen protozoa (Yokoyama and Carlson, 1979). Thornton-Manning et al. (1993) demonstrated that skatole is the etiologic agent causing acute bovine pulmonary edema and emphysema (ABPE) in cattle, acting as pneumotoxin that causes the degeneration of certain lung tissues. Skatole has also the ability to affect the production of serotonin and in high 27

28 concentration to cause the haemolysis of bovine erythrocytes. However, skatole has not toxic effect on pigs as it has in ruminants or microorganisms (Deslandes et al., 2001). As for androstenone, it is not easy to define a precise limit to the level of skatole which is acceptable in the meat. In the literature, some threshold values for skatole concentrations in fat have been used in sensory studies with consumers or for sorting of carcasses; values above which consumers or trained panels would negatively react to meat from entire males. The most common used threshold values are 0.2 and 0.25 μg/g of fat (e.g. by Desmoulin et al., 1982; Mortensen et al., 1986; Bonneau et al., 1992; Claus, 1995, Walstra et al., 1999). In the European study of Walstra et al. (1999), with 6 countries involved, the average of entire males with values between 0.2 and 0.25 μg/g was of 4% while 11% of the pigs were over 0.25 μg/g. This study showed a variability of skatole levels between countries and population of pigs as for androstenone, with values over the threshold of 0.25 μg/g that varied between 2% and more than 20%. The variability of skatole values also involved the pigs population of the UK. In the same study the 8.9% of 400 pigs breed in the UK were over the value of 0.25 μg/g, a value not much different from those reported by Babol and Squires (1995) between 7 and 10%, but in contrast with two precedent studies conducted by MLC (Meat and Livestock Commission, 1989 and 1992) in which the boar over the threshold value were respectively 5% and 11.4%. Managerial strategies are really important for skatole levels because they can have an impact on the production of it in the large intestine. One approach to solve the problem of boar taint caused by skatole would be the limitation of the microbial 28

29 fermentation of tryptophan. Choosing the right diet composition is one essential strategy for controlling and manipulating the microbial activity and therefore the production of skatole in the gastro-intestinal tract (Jensen, 1999). In the last two decades different studies looked at the effect of different kind of diets, obtaining in some cases controversial results; with a special attention to the protein content of the diet, the effect of high fibre diets and the use of high fermentable carbohydrates (Lin et a., 1992; Jensen et al., 1995b; Zamaratskaia et al., 2003; Whittington et al. 2004; Hansen et al., 2006). Feeding strategies (e.g. dry diet vs. wet feed) and feed intake could also affect the microbial population of the guts and the levels of skatole. Anderson et al. (1997) reported that wet feeding with whey reduce fat skatole level compare to dry feeding, and Jensen and Mikkelsen (1998) had a marked effect on the gastrointestinal bacteria with liquid feeding. Mikkelsen et al. (2004) discovered an effect of different structures of the feed (pellets vs. nonpelleted /fine vs. coarse) on the microbial population of the gut, but the effect of feed structure on skatole has never been investigated. Kjelsen (1998) studied the effect of different feeding levels during the growth period and found that the general feed intake had no effect, whereas withdrawing feed for 12 hours prior to slaughter reduced the skatole level. However, this result on 12 hours withdrawal of feed was opposite to that obtained by Anderson et al. (1997). Environmental condition such as stocking density, air temperature and cleanliness and floor type are also important for the skatole levels in the fat (Walstra et al., 1999). Hansen et al. (1994) showed that pigs raised in high stock density, with high air temperatures and on dirty concrete floor have a significantly higher skatole level than those kept clean on wholly slatted floors (Hansen et al., 1994). 29

30 There are strong indications of genetic influences on skatole levels in pigs. Pedersen (1998) showed that the heritability of the level of skatole in the fat is quite high, and Lundström et al. (1994) demonstrated the presence of a major gene affecting its levels. This gene, as the others involved in boar taint, has not yet been isolated due to the strong effect of breed on boar taint that has been largely demonstrated in different studies (Squire and Lou, 1995; Xue et al., 1996; Pedersen, 1998; Doran et al., 2002; Babol et al. 2004). In the study by Babol et al. (2004) four breeds have been compared showing a big variability in skatole levels. The percentages of entire male pigs over the threshold were respectively: 25.5% for Yorkshire; 31.6% for Landrace; 20.3% for Hampshire and 61.1% for Duroc. However, the levels of skatole in a particular breed are not consistent between studies and countries; like Duroc pigs had low levels of skatole in precedent studies (Squire and Lou, 1995; Xue et al., 1996). The only breed that has consistently shown to have high levels of skatole at slaughter age is the Meishan, most probably due to the early sexual maturation that characterise this breed (Hortos et al., 2000; Doran et al., 2002). Therefore the differences in skatole levels between breeds should be considered in pratical pig production, when using entire males, but more studies on commercial crossbreeds are needed. 2.2 Sensory perception of boar taint Boar taint has been described by various researchers as having distinctive sensory characteristics described as urine-like, animal-like, sweat-like and faecal-like (Disjksterhuis et al., 2000). It is widely recognised and accepted that the contribution of androstenone to the unpleasant experience of boar taint is associated with the urine-like perception. However, in the literature androstenone 30

31 is described in other many different ways ranging from unpleasant odour like ammonia, sweaty, dirty, acrid, silage smell, parsnip to a more pleasant sweet floral scent (Patterson, 1968; Claus et al., 1994; Disjksterhuis et al., 2000). It is well demonstrated that the perception and the ability to detect androstenone is variable between people, with differences in the sexes. Amoore and Buttery (1978) reported that 50% of the people tested in their study were anosmic to androstenone, while about 15% detected a subtle odour, describing it in some cases as pleasant, and the remaining 35% were strongly sensitive to androstenone. This variability in the human population to perceive androstenone is well demonstrated in later studies. Baydar et al. (1993) reported a percentage of 21.5% anosmic to androstenone, with a difference between sexes with the 15.8% of the women and 26.8% of the men. A similar difference was previously reported in an extensive survey by Gilbert and Wysocki (1987) in which 15.8% of the women and the 24.1% of the men were unable to detect androstenone. However, Wysocki et al. (1989) demonstrated that after a long period of training a portion of people presenting a specific anosmia to androstenone can induce the ability to perceive androstenone. The participants of the study sniffed androstenone for 3 minutes, three times a day for 6 weeks, and afterwards 50% of them were able to perceive androstenone. Unlike androstenone there is no evidence of anosmia linked to the smell of skatole. The vast majority of people, as high as 99% (Weiler et al., 1997), are able to detect skatole and find it unpleasant mainly associated with a faecal-like perception (Bonneau, 1998). Skatole at low concentrations is normally used in the making of fragrances and perfumes as it is described as pleasant and with flowery odour, but at high concentrations is associated with unpleasant odours (Hansson et al. 1980). In the literature skatole is described as having an odour like faecal, 31

32 mothball, musty and naphthalene (Annor-Frempong et al. 1997a; Disjksterhuis et al., 2000). The use of trained sensory panels is very common in the literature, as people can be trained to distinguish between androstenone and skatole in boar meat and to detect between different levels of these substances (Annor-Frempong et al., 1997a; Disjksterhuis et al., 2000). However, the use of a trained taste panel for sensory evaluation is not widely accepted as the members of the panel are frequently exposed to boar taint compounds, a reason that could lead to an overestimation of the tainted meat in comparison with a consumer panel. Another problem, present also with the use of taste panels, is the difficulty in comparing the results from different studies, because of the differences in the training received by the panellist and in the sample preparation and presentation (Babol and Squires, 1995; Disjksterhuis et al., 2000). But the use of trained panels in the past years has been fundamental to define the best descriptors for both androstenone and skatole, and to discover the thresholds levels for these compounds. In the work by Annor- Frempong et al. (1997a), the first one trying to define some descriptors, the trained panel generated seven descriptors for androstenone: sweaty, ammonia, parsnip, nose-feel, silage, acrid and dirty; and three for skatole: mothball, musty and body reaction. While in the study by Font I Furnols et al. (2000) samples with high androstenone levels were associated to urine, sweat, chemical and rancid odour and flavour, turpentine, viscera, pig/animal and naphthalene odour, and piquant flavour. The samples with high skatole levels were characterised by sweat odour and flavour, stable, manure and naphthalene odour and pig/animal flavour. In an international study, using a standardised sensory procedure and involving trained 32

33 sensory panels in seven European countries, Disjksterhuis et al. (2000) found that between 8 predefined attributes androstenone related mostly to urine whereas skatole was mostly associated with manure and, to a lesser extent to naphthalene. The results of studies with consumer surveys on the acceptability of boar meat can be quite inconsistent because of the large variation in the incidence of boar taint, and because of the variety of culinary habits between countries (Malmfors and Lundström, 1983). Overall the consumers reaction to cooked meat from boars is of repulsion and rejection; however, there are populations, more used to the consumption of entire males, that consider the taint to be an integral characteristic of pork meat. The majority of the variations in the sensory description of boar taint in consumer surveys is due to the individual variation and to the consumer sensitivity to the main compounds, but is also caused by the variations between porcine carcasses in their levels of boar taint compounds (Wysocki and Beauchamp, 1984; Weiler et al., 1997). For all these reasons it has proved challenging in the past years to quantify the percentage of consumers dissatisfied with pork because of boar taint. In their review Malmfors and Lundström (1983) considered nine studies performed in six European countries comparing meat from entire males, gilts and castrates, and reported dissatisfaction that ranges from 5% to 35% for meat from boars compare to a range of 3-10% for gilts and castrates. This variation in the results between meat from boars and gilts has been also reported by Matthews et al. (2000) in their consumer survey conducted in seven European countries. Overall 34% of the consumers were dissatisfied with the odour and 22% with the flavour of entire male pork compare with respectively 28% and 19% for the meat from gilts. Large 33

34 variations were observed between countries, the range of dissatisfaction for odour was from 28 to 42% and from 16 to 31% for flavour. The aim of this study was also to discover the contribution of the two compounds to boar taint. The results showed that skatole contributes more than androstenone to the proportion of consumers which are dissatisfied with the odour of entire male pork, while they have a similar and additive contribution to flavour. In this survey the consumers from the UK had the lowest percentage of dissatisfaction for both odour and flavour. This may be explained by the high percentage of anosmic people compared with other country reported by Gilbert and Wysocki (1987) and by the possibility that UK consumers have been eating meat from boars for many years and have got accustomed to the taint. However, in the study by Matthews et al. (2000) non-pork eaters were excluded from the survey and this rules out the people who had previous poor eating experiences with boar tainted meat and have ceased to consume pork. It is important to consider that as boar taint is mainly evidenced during cooking, the people who prepare the meal are most likely to detect it and if these individuals are also the main meat purchasers then negative experiences with boar tainted pork will influence their repeat purchasing decisions with negative economic repercussions for the pork market (Bryhni et al, 2002, 2003). 2.3 Control of boar taint In the last decades the scientific community, besides studying the mechanisms behind the problem of boar taint, has also examined the several possible methods to reduce the incidence of boar taint in slaughter pigs. The problem of boar taint must be controlled before the utilisation of entire male pigs for meat production can be allowed in most of the countries. At the moment there is no single solution 34

35 to the problem as some methods have significant effects, but are not economically advantageous, and others seem to have only marginal effects as skatole and androstenone do not always respond to the same measures. From the results in the literature we can assume that skatole levels can be reduced by modulating management aspects of pig farming such as diet, feeding and rearing conditions, whereas the genetic selection seems more efficient at lowering androstenone content (Bonneau, 2006; Jensen, 2006). It is commonly accepted that surgical castration eliminates the problem of boar taint; therefore measures that delay or suppress sexual development can reduce both compounds (Babol and Squires, 1995; Babol et al., 1996; Doran et al., 2003). The ways to achieve a delay in the sexual development include restricted feeding or nutrient supply, control of the social environment and alteration of the photoperiod; while the suppression of sexual development can be achieve with immunisation methods Rearing and management procedures, environments The amount of boar taint due to high levels of androstenone and skatole is affected by husbandry practices and different studies have been conducted with particular attention on housing conditions. The social stress, due to higher level of aggression and mounting behaviour, in groups containing entire males leads in most cases to increased levels of boar taint, with a particular increase of androstenone levels. Giersing et al. (2000) reported higher levels of androstenone in the carcasses of aggressive and dominant boars compared with other boars in the same group. Precedent papers (Bonneau and Desmoulin, 1980; Narendran et al., 1980) showed that the differences in androstenone levels was related to the stimulatory effect of group rearing compared to the individual rearing and not to the presence of 35

36 females (Lundström et al., 1987; Andersson et al., 1999, 2005). In pig farming there is the common practise of regrouping pigs during production to maintain uniform groups of similar age and weight, so that pigs can be slaughtered pen-wise. This has been practised even if it has been showed that mixing leads to fighting (Andersson et al., 2003); that the fights are higher in groups of uniform age (Rushen, 1987) and that the degree of unfamiliarity also affects the level of aggression (Rundgren and Löfquist, 1985; Arey and Franklin, 1995). The increase of aggressive behaviour will accelerate the initiation of puberty and provoke testicular activity with a consequent increase in androstenone levels. This has been confirmed in the study by Fredriksen et al. (2003) in which the birth to slaughter systems was tested. The pigs were kept in their litter groups from birth to slaughter, and were not mixed with other pigs, resulting in lower levels of androstenone. The floor areas, the type of flooring, high stocking density and high temperatures in a pig pen have particular consequences for the development and levels of skatole and indole in pig adipose tissue. Different authors (Kjeldsen, 1993; Hansen et al., 1994, 1995) shown that fat skatole and indole levels can be increased in entire males, females and castrates by keeping the pigs on heavily soiled concrete floors with a mixture of faeces and urine, compared with rearing pigs on clean concrete or clean slatted floors. This is especially a problem in warm summer periods when the temperature of the faeces and urine on the piggery floor may be higher than the ambient temperature, particularly when the pigs were lying in the excreta (Hansen et al., 1994). Friis (1993) showed that skatole can be partially re-absorbed through the pig skin, as around 40% of skatole used in the study was absorbed through it in the belly region of the pigs. Usually in intensive production systems the pigs have 36

37 only the possibility to wallow in the dunging area, which can lead to increased levels of skatole and indole in fat. This means that pen floors should be kept clean, or that floors should be well drained, and that pigs should have other means of thermoregulation than wallowing in excreta. As mention before, decreasing the day-length results in earlier sexual maturity and thus androstenone production levels may rise (Andersson et al., 1998). However, in practical conditions the effect seems minor and it is also not known for how long a period or within which photoperiod limits, day length should be controlled. An effective strategy, although expensive, to reduce the incidence of boar taint seems to be the slaughtering at light weights as most studies show that the concentrations of boar taint compounds increase with carcass weight, increasing the proportion of abnormal odours detected by taste panellists or consumers. This could be explained by the positive correlation between the rate of androstenone synthesis and live weight of boars and the onset of puberty (Prunier et al., 1987). However, the European study by Walstra et al. (1999) shows that this relationship between taint compounds and carcass weight across the commercial slaughter weight range in the EU is quite small and correlations between skatole and androstenone concentrations on the one hand and carcass weight on the other are low. The overall correlation was about 0.1 for both compounds and there were small differences between countries; the highest correlation being 0.20 for androstenone and 0.13 for skatole. The range of carcass weight in that study was kg (average 76 kg). In a comparison study in the UK by MLC (Meat and Livestock Commission, 1989) the carcass weights of 65 and 80 kg were considered; skatole did not increase as weight increased and only 5% of samples exceeded the 37

38 threshold of 0.25 μg/g of fat at both weights. However, androstenone concentrations exceeding the threshold of 1.0 μg/g increased from 3% at the lighter weight to 8% at the heavier weight. In a German study by Claus et al. (1994) the carcass weights of 75 and 85kg were compared in Landrace boars. Around 21% of the 75kg boars had levels of androstenone over the threshold of 0.5 μg/g compared to 49% of the 85kg pigs; and the pigs with levels of skatole above the threshold of 0.25 μg/g were 4% and 11% respectively in the 75 and 85kg groups. However, it is difficult to specify a particular weight where boar taint could occur as the genetic aspect of the breeds has an important role in the onset of puberty and consequently on boar taint levels. Moreover the commercialisation of light weight pig carcasses is not economically advantageous as entire males are more profitable at higher pig weights (Walstra, 1974) and a lighter weight cannot guarantee a tainted-free meat Breeds and genetics There are distinct differences in both androstenone and skatole levels between breeds reported in the literature, but the variation between individual animals within the breed suggests also an important effect of genetic factors on the levels of these compounds, which have been identified in a number of studies. Babol et al. (2004) studied the levels of skatole in the most common breeds used in pig production in Europe and North America: Landrace, Duroc, Yorkshire and Hampshire, obtaining some different results compared to precedent works (Squires, 1992; Xue et al., 1996; Pedersen, 1998; Hortos et al., 2000) probably due to different in ages of the pigs between these studies. Babol et al. (2004) reported the Hampshire pigs to be the lowest in skatole levels, while previously it seemed to 38

39 be the breed with the highest levels; Duroc male pigs resulted with highest levels while in the past studies this breed is reported to have low levels of skatole. From these studies it was also shown that between five and eight percent of purebred Hampshire, Yorkshire and Landrace boars have high concentrations of androstenone in fat, whereas around or over 50% of Duroc intact males have high concentrations. The only breed that shows consistently high levels of both compounds at slaughter age is the Chinese breed Meishan, probably due to the fact that these pigs are characterised by an early onset of puberty (Hortos et al. 2000; Doran et al., 2002). From the literature estimating the genetic parameters for androstenone and skatole it results that the biosynthesis of androstenone is governed strongly by genetic component while skatole is not. In the literature, the estimated heritability value of androstenone levels in adipose tissue in commercial pigs is high (h²=0.56), ranging from 0.25 to 0.88 (Sellier, 1998; Selllier et al., 2000; Varona et al. 2005). While skatole levels in fat show medium heritability values ranging from 0.19 to 0.54 (Pedersen, 1998; Tajet et al. 2006), but Lundström et al. (1994) suggested that a major gene is affecting the skatole level. Tajet et al. (2006) have also reported a positive genetic correlation between skatole and androstenone levels ( ). Androstenone can be potentially reduced by controlling the genes that affect the degree of sexual maturity as those responsible for the activity of the enzymes responsible for androstenone testicular synthesis or liver metabolism. In a recent study Grindfleck et al. (2010) selected to analyse fifteen candidate genes potentially affecting androstenone levels in boars, which have been isolated in precedent 39

40 studies, and suggested that several are important in the regulation of androstenone level. It is possible to select for pigs which do not have boar taint but work is still necessary to identify the right genetic markers for low boar taint. Before the selection for pigs with low incidence of boar taint starts it is important to isolate the correct genes as a wrong selection strategy could have economic disadvantages if the selection reduces the benefit of the natural anabolic hormones or the eating quality of the pork Surgical castration The surgical castration (gonadectomy) of young male pigs is a very old procedure widely practised in pig farming to reduce the aggressive male behaviour and to prevent the occurrence of boar taint. In the 25 European Union (EU) countries about 100 million pigs are castrated every year, representing approximately the 80% of the male pig population (Migdał et al. 2009). In the EU the Directive 2001/93/EC stipulates that piglets can be castrated without anaesthesia within the first week of life, after it only a veterinarian should perform the castration with the use of anaesthesia and additional prolonged analgesia. In Switzerland since 2009, regardless piglet s age, surgical castration without anaesthesia is prohibited and in Norway there is the same restriction with the objective to fully ban surgical castration by This practise is very effective against boar taint as it prevents the synthesis of androstenone, however the castrate pigs do not benefit of the anabolic effect of the other androgens produced in the testis, with resulting worst production performances (Walstra, 1974). 40

41 Although this method is effective against boar taint, the welfare concerns brought up by the scientific community related to the immediate, short-term and long-term effects of the surgical castration on the animals have pushed the European countries to look for valid alternatives with the prospective of a future ban Immunocastration The term immunocastration commonly refers to the practise that uses a vaccine against GnRH with the aim to inhibit testicular development and consequently reduce the incidence of boar taint. However, in the past thirty years different hormonal treatments have been tested as a possibility to reduce the incidence of boar taint. A first attempt has been made against androstenone itself, trying to suppress its production in the testis, but a large variability in the responses to the treatment has made it unreliable (Shenoy et al., 1982; Daniel et al., 1984). The use of the porcine growth hormone (pgh) has been tested by Hagen et al. (1991). This treatment improved feed efficiency in boars, reduced the levels of subcutaneous fat and marbling fat, but did not have any incidence on boar taint compounds. Bonneau et al. (1994) found that the immunization with anti-luteinizing hormonereleasing hormone (LH-RH) was effective in reducing the levels of androstenone in the fat but this treatment had no positive effect on skatole concentration and on the performance of the pigs. As GnRH is too small to be immunogenic, its vaccine involves the injection of GnRH conjugated to a foreign protein, and combined with an adjuvant, to induce anti- GnRH antibody formation. Two major factors to be considered in the development of vaccines against GnRH for commercial use in farm animal species are the adjuvant used, and the number of immunizations needed for effective 41

42 immunocastration. At present a commercial vaccine has been developed and launched with the name of Improvac (Pfizer Animal Health, Louvain-la-Neuve, Belgium) and used worldwide on boars. This vaccine needs to be administered twice intramuscularly with the second booster injection suggested by the producer to be given 4 5 weeks prior to slaughter. In immunocastration the timing of the vaccination is really important. Early immunization will cause a complete castration with unambiguous results on testes weight, making differentiation on the slaughter line very easy. However, most of the economic advantages typical of entire males are lost in these animals (Turkstra et al., 2002; Zeng et al. 2002). The alternative, later immunisation, concentrates on maintaining most of the performance advantages of intact male pigs in immunised animals. The disadvantage is that some measurements would have to be performed on the carcasses in order to check the effectiveness of the treatment. In this procedure, an optimum time interval between the booster injection and slaughter needs to be established as the sexual development of the animal can differ between different breeds. The challenge is to keep testicular secretion of anabolic steroids at a high level until as late as possible and still allow enough time for immunocastration to decrease fat skatole and androstenone concentrations to acceptable levels before slaughter. Different studies have shown that immunisation of young intact male pigs against GnRH is effective at inhibiting genital tract development and reducing plasma LH, FSH and testosterone concentrations (Falvo et al., 1986; Meloen et al., 1994; Dunshea et al., 2001). It is also effective in decreasing androstenone and skatole levels in the fat and hence the incidence of boar taint (Fuchs et al., 2009; Schmoll et 42

43 al., 2009). Both the mean levels and the variability of androstenone and skatole levels are sharply reduced in immunocastrated pigs compared with entire males. The levels of both compounds, as well as boar taint intensity, are similar in immuno- and surgical castrates (Dunshea et al., 2001). However, this vaccination faces a concern from the industry and the consumers as there is variability in the immune response of the animals, resulting in some pigs not being immunocastrated and with the risk of consequently high androstenone levels (Bonneau et al., 1994; Turkstra et al., 2002). Owing to the very high variability of androstenone and skatole levels, the non-respondent pigs do not necessarily exhibit high levels of boar taint-related compounds; however, due to the absence of a satisfactory method for the assessment of boar taint on the slaughter line, it is presently not possible to guarantee the absence of boar taint in all individual immunocastrated pigs. Another concern related to immunocastration is the risk of self-injection for the person carrying out the treatment. The effects of immunocastration on performance, carcass characteristics and meat quality have been widely studied, but it is yet not clear if immunocastration is better than surgical castration. Zeng et al. (2002) found that immunocastrated pigs had better food conversion ratio (FCR) than surgical castrated pigs but worst than entire males, while Fàbrega et al. (2010) found no differences between immunocastrated pigs and boars with the first ones having higher values of growth rate and feed intake. However, Zeng et al. (2002) reported that the immunocastrated pigs had a tendency for better growth performance than castrated. It seems that generally meat percentage and lean percentage are higher in boars than in immunocastrated and surgical castrated pigs (Zeng et al., 2002; 43

44 Fàbrega et al., 2010; Font i Furnols et al., 2010), but surgical castrated pigs are fatter than the immunocastrated ones (Fuchs et al., 2009). Font i Furnols et al. (2009) found that from a sensory perspective meat from vaccinated males is not different from meat from females or surgically castrated males, whereas meat from entire males is distinguishable from all other types Effect of dietary composition In the last decades the nutritional factors have been the most studied management aspects of pig farming. This is mainly because the skatole production is strongly linked to the diet composition. Skatole is a product of bacterial activity in the large intestine and the diet can influence its levels in fat, possibly through altering the bacterial activity or availability of the substrate, tryptophan (Hansen et al., 1997; Jensen et al., 1997). However, the level of tryptophan in the feed per se does not influence skatole production (Pedersen et al., 1986), as tryptophan is easily absorbed in the small intestine and thus does not reach the caecum and colon where skatole is produced (Jensen and Jensen, 1998). Different diets may have quantitative and qualitative effects on intestinal microflora and consequently influence the rate of skatole synthesis; while a change in the diet seems to have no effect on androstenone concentration in the fat. The level of androstenone in the fat is not influenced by the ingredients of the diet itself but it can be increase with an high energy diet (Claus et al., 1994; Øverland et al., 1995; Zeng et al., 2002), because it is able to accelerate the pubertal development in entire male pigs, which may in turn result in an earlier increase in the levels of androstenone (Einarsson et al., 1979). 44

45 On the contrary, according to Jensen and Jensen (1998), the are several ways to modify the diet that can reduce the concentration of skatole. Reducing the protein fermentation is one of the solutions, either by lowering the amount of protein reaching the hindgut, or by changing microbial metabolism by using carbohydrates that are not absorbed in the small intestine and which are preferentially fermented by those microbiota which not produce skatole. Another way is to increase the bulk of digesta in the hindgut, by the use of water holding dietary fibres such as lignin or pectin. This could also lead to a decrease in the gastro-intestinal transit time, allowing time for less skatole to be absorbed in the hindgut. Another way to change the composition of the microbiota in the hindgut is by stimulating the production of indole instead of skatole. It has been showed that a liquid feed instead of dry feed (Jensen and Mikkelsen, 1998), or a higher intake of dietary fibres (Agergaard et al., 1998) leads to a relatively higher amount of indole production. In an in vitro study, Jensen and Jensen (1993) showed that the proportion of indole vs. skatole was 20, 60 and 85% at ph 5.0, 6,5 and 8.0, respectively. Later, Claus et al. (1994) demonstrated that the addition of bicarbonate in the diet increases the ph-value in the intestine with a consequent increase in the proportion of indole vs. skatole. As for androstenone the use of diets with high energy content increase the production of skatole because of hormonal regulation (Claus and Raab, 1999). Different studies demonstrated that high energy diets increase the levels of Insulin-like growth factor 1 (IGF-1) with a consequent elevated cell mitosis and apoptosis rate. This has a strict correlation with skatole production, as the resulting dead cells from the elevated apoptosis are a major source of tryptophan for the skatole producing bacteria (Hansen et al. 1995; Claus and Raab, 1999; Claus et al., 2003). 45

46 Several studies have focused on the effect of feed ingredients on the backfat skatole content of entire male pigs. Most of the studies have investigated the effects of diets rich in fibre, such as sugar beet pulp, or rich in fermentable carbohydrates that escape digestion in the small intestine, such as raw potato starch and chicory Sugar beet pulp The inclusion of fibre-rich feedstuffs in the diet influences the backfat skatole content, but not all types of dietary fibre seem able to increase hindgut fermentation. For example Jensen et al. (1997) found that lignin, a component in wheat bran, had no effect on fermentation whereas pectin, present in sugar beet pulp, readily increased fermentation. Contradictory hypotheses are available to explain the effect of high fibre diet on fat skatole content. Some authors suggested that a diet with extra content of fibre allows more undigested protein to reach the large intestine, with consequently more degradation of tryptophan to skatole (Just et al., 1983; Lin et al., 1992). Jensen et al. (1995a) found that carbohydratefermenting bacteria in the large intestine get more active if a diet with more fermentable sources like dietary fibres is supplied to pigs on the expense of the proteolytic bacteria, which produce skatole. Moreover, the carbohydratefermenting bacteria produce large amounts of short-chain fatty acids (SCFA) that cause a decrease in ph, antagonising the proteolytic bacteria that have an optimally function at neutral or alkaline ph. Finally, Jensen et al. (1995b) demonstrated that extra dietary fibre results in more bulky material in the large intestine that increases water binding capacity, leading to a dilution of skatole in the large intestine. As consequence they obtained a decrease in skatole absorption, due to a in less contact of skatole with the intestinal wall. 46

47 Sugar beet pulp (SBP) is a by-product of the sugar refining, characterised in high content of water soluble carbohydrates and mostly rich in crude fibre of high digestibility. The use of sugar beet pulp in pigs is controversial, as it is not completely clear if it has negative effects on productivity and meat quality aspects, but most important the reducing effects on skatole level reported in the literature are variable. Wood et al. (1993) tested the effect of different levels (15, 30 and 45%) of SBP on the eating quality of pig meat and on skatole concentration in the fat. As the concentration of SBGP increased, the P2 fat thickness decreased and water concentration in fat increased. From the sensory test both pork odour intensity and abnormal odour intensity tend to decline as SBP increased, but not with a significant effect. Finally, the skatole concentration in the backfat decrease as the percentage of SBP increased. However, later works had opposite results on skatole concentration in the fat. Øverland et al. (1995) and Van Oeckel et al.(1998) found no effect on skatole when a 15% diet was tested; whereas Knarreborg et al. (2002) and Whittington et al. (2004) observed significant reduction in skatole levels in pigs using respectively 10 and 20% of sugar beet pulp. The different results on the use of SBP could be due to numerous aspects that differed between studies, like breeds and managerial factors Raw potato starch Chicory and potato starch are diet ingredients reach in low digestible carbohydrates that have been found to have the most pronounced effect in lowering skatole levels. Drochner (1993) found that raw potato starch (RPS) is not fully digestible in the upper gastrointestinal tract; the resistant starch affects the microflora of the gastrointestinal tract and influences intestinal functions with an 47

48 effect on skatole levels (Jensen et al., 1995b; Kleessen et al., 1997; Claus et al., 2003; Le Blay et al., 2003; Wang et al., 2004). The reduction of skatole production is mainly due to the production of considerable amounts of butyrate that is consequence of the RPS fermentation. Has been demonstrated that butyrate is able to reduce cells apoptosis in the large intestine of the pigs (Mentschel and Claus, 2003) with a reduction of cell debris that is the main substrate for skatole formation in the gut (Claus et al., 2003). The inclusion of RPS in the diet could also reduce the absorption of skatole from the large intestine, as the faecal wet and dry weight is increased by RPS (Wang et al., 2004), while the intestinal transit time is decreased (Drochner, 1993). In the last decade different studies have demonstrated that a supplementation of RPS to the normal commercial diets reduce significantly the levels of skatole in faeces, blood and fat (Lösel and Claus, 2005; Zamaratskaia et al., 2005b; Lösel et al. 2006; Chen et al. 2007; Pauly et al., 2008). Lösel and Claus (2005) tested different levels of RPS in the diet (20, 30 and 40%) for a period of 2-3 weeks before slaughter, and found that a minimum of 30% is required to have a significant reduction in skatole levels. Different percentages and different feeding period have been tested; however, most recently Pauly et al. (2008) reported that feeding 30% of RPS for 7 days prior to slaughter was able to significantly reduce the skatole concentration in entire males when compare to other entire males used as control but not when compared to castrated Inulin and chicory roots Inulin, as all the fructo-oligosaccharides, is a non-digestible carbohydrate (Cumming et al., 1997; Roberfroid and Slavin, 2000) and this characteristic is due 48

49 to the beta configuration of the anomeric C2 in its fructose monomers that form β 2 1 glycosidic linkages. Alles et al. (1996) found that as inulin resists the digestion in the small intestine of the human digestive enzymes it finishes to serve as substrates for the microflora in the large intestine. The presence of inulin in the hind gut favourites the growth of those bacteria that are considered beneficial at the expenses of those potential harmful, and the end products of the microbial fermentation are lactic acid and SCFA (Buddington et al., 1996). Chicory (Cichorium intybus L.) is a bioactive crop that has been subject of numerous studies as it has been considered potentially able to improve the meat eating quality of intact males, reducing the concentration of skatole and consequently boar taint. The chicory roots are rich in fructo-oligosaccharides, especially inulin, and to a lesser extent in sesquiterpene lactones (bitter compounds), that could also have a positive effect (Bais and Ravishankar, 2001). Chicory roots also contain glucose and sucrose (De Leenheer, 1996). Inulin, extracted from the chicory roots, is a mixture of oligo- and polymers of GpyFn (α-d-glucopyranosyl-[β-dfructofuranosyl]n-1-d-fructofuranoside) and FpyFn (β-d-fructopyranosyl-[β- Dfructofuranosyl]n-1-D-fructofuranoside); however both the compounds have to be considered under the same nomenclature. Inulin can differ in the number of fructose unites, degree of polymerization, that varies from 2 to 60, with an average of 12 (Roberfroid, 2000). Claus et al. (1994) first and then Jensen and Jensen (1998) demonstrated that the addition of non-digestible oligosaccharides like fructo-oligosaccharides to the diet decrease skatole levels in faeces, backfat and blood; consequently the use of chicory roots as supplement of the diet as been considered to be a solution for boar taint. In 49

50 the last decade the use of chicory roots and pure inulin have been tested using different levels in the diet and fed for different period of time, but it still not clear which are the minimum amount of chicory and the time required to have a positive effect on skatole. A first positive result is reported by Rideout et al. (2004) who found that a supplementation of 5% chicory inulin extract for 2 weeks significantly decreased the skatole excretion in the faeces. However, in this study only six animals have been tested and no measurements of skatole concentrations in the fat were been taken, leaving the doubts of the effect on boar taint. Later different Danish studies looked at the effects of chicory on skatole levels in the fat and also on the sensory aspects of the meat. A first study with 2.5, 5, 10 and 20% dried chicory have demonstrated that 10% dried chicory or more in the feed reduced skatole in the blood and backfat of entire male pigs significantly after 7, 14 and 21 days of feeding, resulting in a significant reduction in perceived boar taint and also improving the flavour and taste of meat produced from entire male pigs (Byrne and Hansen, 2005; Hansen, 2005). The reduction in skatole level both in blood and backfat was dose-dependent so that the reduction was most pronounced with 20% chicory in the feed and the effect was already seen after 1 week feeding (Hansen, 2005). Further studies partially confirm these results. Hansen et al. (2006) tested the effect of a supplementation for different period of time of crude chicory (25%) or dry chicory (25%) or inulin (14%) on skatole levels in the fat, in three different trials with a total of 80 pigs. It was found that 25% crude chicory fed for 4 weeks before slaughter significantly reduced the levels of skatole in backfat, but not when fed for only 2 weeks. Also the supplementation of dry chicory and inulin were able 50

51 to significantly reduce skatole concentrations after 6 weeks on the diet. Byrne et al. (2008) looked at effect of chicory on the sensory aspect of the meat using the same animals from the previous study. The long period feeding treatments (4 and 6 weeks) significantly reduced perceived sensory boar taint in the cooked pork meat on entire males and had no negative effect on the other sensory characteristics. In the latest study by Hansen et al. (2008) used a supplementation of 10 and 13.3% dry chicory roots fed for 1 and 2 weeks, finding no significant effect of all the treatments on skatole levels in the fat. 2.4 Methods for detection of boar taint Online detection on carcasses On-line detection of carcasses with unacceptable levels of boar taint compounds, which may make them unsuitable for production of high quality products, would be one solution to the boar taint problem. In the European Union the Directive 91/497/EEC allows the trade of meat from entire male pigs with no restrictions for carcass weight under 80 kg; male carcasses over 80 kg may be passed fit for human consumption provided they have free from boar taint after been assessed with an objective method. However, the main problem is that this Directive does not state which is the best method to detect boar taint on the slaughterline. This is why at present, in the EU, there is no harmonized method for detecting boar taint. In the UK some abattoirs test the carcasses using a boiling test, where a sample of fat is cooked and assessed by an operator (personal observation). Jarmoluk et al. (1970) suggested an alternative method consisting in the use of soldering iron applied to the exposed backfat of the carcass to cause the volatilisation of androstenone and skatole which can be detected by an operator. Both methods are successful in 51

52 detecting boar taint; however the effective detection differs between operators and the fatigue of the sensory response develops quickly. Bonneau and Russeil (1984) suggested that the measurement of the length and size of the bulbourethral glands could be use to predict the incidence of boar taint in the carcasses. Later Babol et al. (1995) tried to improve this strategy correlating these measurements with skatole quantification in the fat. However, even if this strategy is fast and of easy application on the slaughter-line, it results not completely efficient and reliable. At present the only technology applied on the slaughter line is an automatic assay, based on the spectrophotometric method developed by Mortensen and Sorensen (1984), used in Danish slaughter plants to detect skatole. A backfat sample is removed from the carcass, analysed in a laboratory and the results used to remove from the production line the tainted carcasses. However, also this method presents limitations as not only skatole but also other indole-related compounds are measured, with the risk of an overestimation of boar taint, while androstenone is not measured, and a maximum of 180 samples per hour can be tested. In the last decade, there has been a rapid development of chemical sensor and gas sensor technologies for analysis of volatile compounds. Different studies showed that chemical sensor arrays combined with multivariate data processing methods could be potentially used for rapid non-destructive analysis of meat (Haugen and Kvaal, 1998; Blixt and Borch, 1999). Boar taint could also be detected in the vapour phase by non-specific gas-sensor arrays; this method would allow the measurement of the odour of the meat instead of analysing the specific compounds 52

53 responsible for boar taint (Gunn et al., 2004). However, as this technique requires training and calibration against sensory analysis or some valid reference method, it cannot completely replace reference methods like the use of sensory panels. The use of electronic nose is a promising technology to be use on the slaughter line to detect boar taint, but at the present time this technology is still under development (Vestergaard et al., 2006). The methods described above for detecting boar taint are time consuming and costly and it would therefore be useful to have objective rapid methods in order to sort out the boars on the slaughter line based on both chemical and sensory criteria. New methods should allow a high number of samples to be analysed within a short period of time with a sufficient reproducibility and accuracy Laboratory techniques In the last thirty years different objective techniques to determine the levels of boar taint compounds in the fat have been developed and improved, many of which are time consuming and expensive and so not applicable for a daily use on the slaughterline. The most used techniques rely either on immunological or chromatographic approaches. The most effective immunological method for detection of both skatole and androstenone is based on the enzyme-linked immunosorbent assay (Abouzied et al., 1990; Walstra et al., 1999). This technique seems to be sensitive and reliable but requires sophisticated equipment and a long time preparation, and also the production of large quantities of antibodies proves to be a limit. 53

54 The chromatographic methods are the most diffuse techniques to quantify both androstenone and skatole concentrations, that are identified with the use of a chromatogram after been separated from other volatile compounds while they pass through a chromatography column. The headspace Gas Chromatography Mass Spectrometry is an expensive and technically difficult method that resulted to be more suitable to detect skatole and indole than androstenone (García-Regueiro et al., 1995; Rius et al., 2005). Another technique developed by De Brabander et al. (1985) is the High Resolution Gas Chromatography. The preparation of the sample required a long procedure in which the taint compounds are extracted from the sample after derivatization and saponification, to be then ready to be quantify in an open tubular capillary gas chromatography. The most popular and widely used chromatography technique is the High Performance Liquid Chromatography, that is able to quantify rapidly and simultaneously both levels of androstenone and skatole in a sample (García-Regueiro and Diaz, 1985; Hansen-Moller, 1994; Banon et al., 2003). The chromatographic-based techniques are considered in the scientific community to be accurate and reliable for the determination of androstenone and skatole concentrations in the fat. However, for all of them the preparation of the sample is a long procedure, technically difficult to perform, with the added risk of equipment failure, and most important they are also expensive. 54

55 Experimental work 3 Androstenone and skatole levels in pigs on British farms 3.1 Summary Male pigs are not castrated in Britain and boar taint, an offensive odour/flavour, is a potential problem for the consumers when the meat is cooked. Boar taint is due to high concentrations in the fat of at least one of the main compounds involved, skatole and androstenone. These compounds are derived from different metabolic processes, but both are metabolised in the liver with an important interaction for their final levels in the fat (Doran et al., 2002). Past studies have found that for most of the pigs produced in the UK, around 85% (Walstra et al., 1999), the levels in the fat are below the thresholds of 0.2 μg for g of fat for skatole, and 1.0 μg for g of fat for androstenone, above which taint problems can arise (Bonneau et al. 1992). There is no recent survey on the levels in the fat of skatole and androstenone in British pig herds. In this study 63 farms serving two commercial abattoirs in the East of England were tested for levels of both compounds using a composite sample, each one produced from the backfat of 50 entire males. The overall means for the 63 farms were 1.65 μg/g of fat for androstenone and 0.14 μg/g of fat for skatole. There were 52/63 (83%) of farms with a cumulative average concentration of androstenone above the threshold of 1.0 μg/g but only seven farms (11%) with a composite skatole concentration above the 0.2ug/g threshold for this compound. Overall 53/63 (84%) farms had at least one of the compounds over the threshold level and six (9.5%) farms had both values over the thresholds. 55

56 If compared with past studies, the results for the androstenone levels are unusual in the overall mean value and for the number of farms resulting over the threshold. However, the skatole results are more similar to the past surveys, with only 11% of the farms over the threshold value. This could mean a reduced problem with boar taint as in the UK this compound seems to be more associated with boar taint than androstenone (Annor-Frempong et al., 1997b). 3.2 Introduction In the United Kingdom in the last 40 years the castration of pigs has not been widely practised and is currently prohibited under the Assured British Pigs Farm Assurance Standards (Red Tractor Farm Assurance Pigs Standards, 2010). A potential problem consequent to the use of entire male pigs for meat production is boar taint, an offensive odour/flavour that is mainly perceived during cooking. The perception of boar taint is due to the high concentrations in the fat of at least one of the two compounds widely accepted to be responsible for tainted pork, androstenone and skatole. These compounds are derived from different processes; androstenone is a pheromone produced in the testis and skatole is a product of the fermentation in the large intestine. Both compounds are metabolised in the liver to be then excreted by the kidneys with the urine. The amount of androstenone and skatole that the liver is not able to metabolise goes into the peripheral blood stream and it is accumulated in the fat. The sexual steroids, in particular high levels of androstenone, are able to influence negatively the skatole metabolism with a consequent accumulation in the fat (Doran et al., 2002). For this and other reasons, not all completely clear, some pigs show levels in the fat of these compounds over the threshold where taint problems can arise. These thresholds are 0.2 µg/g of fat 56

57 for skatole and 1.0 µg/g of fat for androstenone (Bonneau et al, 1992). It is not possible to give a definitive percentage of pigs that present boar taint as this is influenced by many factors. In a European study by Walstra et al. (1999) conducted in six countries, the amount of pigs presenting levels over the thresholds were 29% and 15% respectively for androstenone and skatole. In particular in the UK the pigs over the threshold for androstenone were around 22%, and around 9% for skatole. Older studies in the UK have reported different percentages showing the variability that characterises the boar taint problem. In a study by MLC (Meat and Livestock Commission, 1989) around 5% of the animals were reported to be over the threshold for androstenone and another 5% for skatole, while later Babol and Squires (1995) reported that in UK 10% of the pigs were over the threshold for androstenone and around 8% for skatole. In the literature after the study of Walstra et al. (1999) it is not possible to find other works that can give us a realistic idea of the relevance of the boar taint problem in the British pig population. This study investigated the frequency in herds of pigs of skatole or androstenone above the threshold concentrations. 3.3 Materials and methods A total of 63 farms supplying two commercial abattoirs of a major pig processing company in the East of England were screened for levels of androstenone and skatole in the fat. Entire male pigs from 30 farms were sampled in one abattoir (A1) and 33 farms were represented in a second site (A2). The aim was to establish a boar taint benchmark representative of the British pig population. For each farm, samples of subcutaneous fat from the dorsal neck region were collected from a minimum of 29 to a maximum of 50 entire male pigs, of unknown breeds, 57

58 slaughtered on the same day. The sample collected from each pig was an area of skin plus attached fat all the way down to the intersection with the meat of max 15x5 cm, taken from the cervical region close to where the head is removed. After an initial training visit to the abattoir, all samples were taken over several weeks by the staff of the two abattoirs. Samples per farm were individually labelled and bagged and stored at -20ºC before being sent in batches to University of Bristol (UoB), and hot carcass weight and P2 fat thickness were recorded for each pig. For each farm a single sample was obtained from all the pigs. After removing the skin and the subcutaneous glands, 2 g of fat, including all the layers, was obtained from each pig. The 100 g of fat obtained for each farm (when 50 pigs were sampled) was then blended together using a food processor Moulinette S (Moulinex ) and subsampled for analysis. In the obtained sub-samples of fat, skatole and indole concentrations were measured using the simultaneous distillation-extraction procedure followed by GC analysis, according to the methodology described by Annor-Frempong, et al. (1997b) (see Appendix 1). Androstenone concentrations were measured using a modification of the high resolution gas chromatographic procedure of De Brabander and Verbeke (1986) (see Appendix 1). The threshold values used were 0.2 and 1.0 µg/g fat for skatole and androstenone respectively. Statistical analysis was performed using SPSS computer software. 3.4 Results and discussion Analysis of the sub-samples of fat from the 63 farms obtained from the two commercial abattoirs revealed an overall average of 0.14 µg/g of fat for skatole and 1.65 µg/g of fat for androstenone (Table 3-1). The average value for indole was 0.07 ± 0.03 µg/g of fat. 58

59 Table 3-1. Concentrations of taint compounds in the 63 farms (µg/g fat). Skatole Androstenone Mean ± Standard Dev ± ± Geometric mean Median Minimum Maximum The number of entire males collected from each farm varied from a minimum of 29 to a maximum of 50, with an average of 48 pigs sampled for each farm. The mean value for hot carcass weight was 83.2 ± 3.7 kg (mean ± standard deviation), and the P2 value was 11 ± 1.1 mm. The results for skatole and androstenone levels involving all the 63 farms are shown in Figure 3-1, in which each bar represents the mean value for one farm whose backfat samples have been minced together. The threshold level for skatole is represented by the blue horizontal line while the red one is the androstenone threshold. There was wide variability between farms, as shown in other works on boar taint. There were 52/63 (83%) of farms with a cumulative average concentration of androstenone above the threshold of 1.0 μg/g but only seven farms (11%) with a composite skatole concentration above the 0.2 μg/g threshold for this compound. Overall 53/63 (84%) farms had at least one of the compounds over the threshold level and six (9.5%) farms had both values over the thresholds. If compared with the past studies cited previously, the results for the androstenone levels are unusual in the overall mean value of 1.65 µg/g compared with 0.81 µg/g found by Walstra et al. (1999) and for the number of farms that are over the threshold, while the skatole values are similar to those obtained before (0.14 vs µg/g). The 59

60 µg/g of fat variability in the concentrations is presumably due to the different genetics and production systems on the different farms Skatole Androstenone Figure 3-1. Concentrations of skatole and androstenone in the 63 farms. The farms were using both indoor and outdoor production systems; it is not known if there is any influence of this aspect on boar taint levels but other managerial factors do have an influence, like cleanness of the pen floor or day-length (Friis, 1993; Andersson et al., 1998). The breeds used on the farms in this study are not known but are presumed to be composite commercial crosses carrying the genotypes of several breeds. Different studies have shown differences between breeds in the levels of both compounds and have shown how much genetics can influence the individual levels within breeds (Squires, 1992; Xue et al., 1996; Pedersen, 1998; Hortos et al., 2000; Babol et al., 2004). 60

61 There was no correlation between the levels of skatole and androstenone in the whole population of 63 farms, but there was a significant (P<0.05) correlation between these values in the 33 farms of abattoir A2 (Figure 3-2 and 3-3). Although many studies show positive correlations between the compounds (Bonneau et al., 1992; Zamaratskaia et al., 2004), this is not always the case. Figure 3-2. Skatole and androstenone levels (μg/g) in the 63 farms. 61

62 Figure 3-3. Distribution and correlation between skatole and androstenone levels (μg/g) in the 33 farms of Abattoir 2. The relationships between androstenone and skatole concentrations and carcass weight and P2 fat thickness are shown in Tables 3-2 and 3-3. These show that for androstenone there was a significant correlation (r=0.35; P<0.01) with the carcass weight of the pigs (Figure 3-4) but not with the P2. So the higher values of androstenone found here compared with previous studies can be partly explained by the higher weight of the carcasses (83.2kg) compared with the national mean in 2009 of 78kg (BPEX Annual Technical Report 09-10). For skatole there were no significant correlations with both weight and P2 fat thickness. 62

63 Figure 3-4. Distribution and correlation between androstenone levels (μg/g) and carcass weight (kg) in the 63 farms. Table 3-2. The correlation of taint compounds with carcass weight. Correlation Coefficient (r ) P-value Skatole Androstenone < 0.01 Indole < 0.05 Table 3-3. The correlation of taint compounds with P2. Correlation Coefficient (r ) P-value Skatole Androstenone Indole Analysis of variance showed a significant difference between the pigs collected in the two abattoirs for levels of skatole (P<0.001) and androstenone (P<0.001) as 63

64 shown in Table 3-4. The overall mean value for skatole was lower in A1 (0.10 μg/g) compared with A2 (0.17 μg/g), while androstenone showed an opposite result (A1=1.99 μg/g; A2=1.33 μg/g). Table 3-4. Differences in taint compounds, weight and P2 between abattoirs. Abattoir p-value Variable A1 A2 Skatole (μg/g) <0.001 Androstenone (μg/g) <0.001 Indole (μg/g) Weight (kg) <0.05 P2 (mm) In A1 all the 30 farms had androstenone concentrations above the threshold value of 1.0 µg/g and only one farm (3.3%) had skatole concentrations above the threshold of 0.2µg/g fat. In A2 there were 22/33 (67%) of farms with a cumulative average concentration of androstenone above the threshold but only six farms (22%) with a composite skatole concentration above the threshold for this compound. Values for individual farms are given in the Appendix 2. The lower carcass weight of the pigs in A2, along with different genetics and production systems on the different farms may explain these variations. The skatole values are similar to the mid-range to higher end of the values found in other studies in the UK; however, the values for androstenone are all higher than in these previous studies (Meat and Livestock Commission, 1989; Babol and Squires, 1995; Walstra et al., 1999). Whilst the high numbers of farms with composite sample androstenone concentrations above the threshold might indicate a high prevalence of boar taint, this may not be so as British assessors tend to associate this with 64

65 skatole rather than androstenone. The sensory assessment of fresh pork is an important aspect of the eating quality; in the last twenty years many studies have attempted to relate tissue concentrations of androstenone and skatole to the sensory attributes of the meat without reaching a common result on the respective importance of them to boar taint (Lundström et al., 1988; Calus et al. 1994; Annor- Frempong et al. 1997b; Matthews et al. 2000). In this survey only 7/63 samples had skatole concentrations above the threshold. However, to get a composite value just over the threshold of 0.2 means that some pigs must have had skatole concentrations in their backfat much greater than this with a consequent problem of tainted meat in these individual pigs. 3.5 Conclusions This work showed a wide variability in the levels of androstenone and skatole in the fat of the British pig population. This variability was also underlined by the significant difference (P<0.001) in the levels of the compounds between the farms supplying the two abattoirs used for the sampling, presumably due to different genotypes and husbandry practises. There was a positive correlation between the hot carcass weight and the androstenone levels in the fat, but the high carcass weight (83.2 ± 3.7 kg) would not explain the high percentage of farms (83%) over the threshold for androstenone. These results are unusual in that the concentrations of androstenone are much higher than expected and more farms were above the threshold than expected. However, the skatole results are more encouraging as only 11% of the farms were over the threshold and in the UK this compound seems to be more associated with boar taint than androstenone. 65

66 4 Androstenone and skatole levels in different fat tissues 4.1 Summary Boar taint is the major problem of meat quality related with the production of entire male pigs. The main compounds associated with boar taint are androstenone, a pheromone produced in the testis, and skatole, a product of the microbial fermentation in the hind gut, that mainly accumulate in adipose tissue as they are both lipophilic molecules. The most common adipose tissue used to analyse the levels of these compounds in most previous studies on boar taint is the subcutaneous fat from the dorsal mid-loin site (refer to as P2) in the carcass, usually referred to as backfat. The concentrations of taint compounds in different parts of the carcass have not been widely documented, as most studies looked at alternatives to adipose tissue (Claus et al., 1993; Tuomola et al. 1996). In this study three different sites of the carcass (backfat, neck and cheek) of 20 entire male pigs from a commercial abattoir were sampled to compare the levels of the taint compounds and to see if alternative sites to backfat can be use for the study of boar taint. The overall means for the 20 pigs were 0.11 µg/g of fat for skatole and 1.16 µg/g of fat for androstenone. There were no significant differences between the mean values of either taint compounds between the three sites. For all the taint compounds analyzed the measurements from the three different carcass sites were positively (P < 0.01). However, analysis suggests that measurements of skatole in the fat from the neck seem to slightly overestimate its concentration compared with backfat and cheek; while the measurements of androstenone in backfat seem 66

67 to slightly overestimate its concentration compared with those of the fat from neck and cheek. Wood et al. (1986) have reported different percentages of fat in adipose tissue from different part of the carcass, which could explain the differences in the concentrations of the taint compounds. The analysis of the difference of the mean between the sites showed a good agreement between the measurements for skatole and indole and to a less degree also for androstenone. Because of the small number of pigs studied it cannot definitely concluded that the measurement of taint compounds in the backfat are equal to those in the fat from the neck and the cheek of the pig. 4.2 Introduction Meat and products from entire male pigs often have a higher incidence of odours and flavours found unpleasant by some consumers, known as boar taint. Boar taint is due to an excessive accumulation in the adipose tissue of the lipophilic compounds skatole and androstenone, and to a much lesser degree of indole. Skatole and indole, both characterised by a faecal-like odour, have a common origin as they are synthesized by bacteria in the hind gut and their levels have been shown to be more susceptible to dietary and environmental conditions (Claus et al, 1994; Hansen et al., 1994). Androstenone is a sexual pheromone, usually described as having a urine-like odour, produced in the testis in parallel with the biosynthesis of other steroids during steroidogenesis, and its level is highly dependent on the sexual maturity of the pig (Bonneau et al, 1987b). Both the main compounds, androstenone and skatole, are lipophilic molecules that can be easily transferred from plasma to adipose tissue. Their levels in the fat are characterised by a dynamic relationship with those in the plasma; as their levels in the plasma 67

68 increase a heavy accumulation in the fat usually follows but as the concentrations in the peripheral plasma decrease also the concentrations in adipose tissue gradually decrease (Malmfors et al., 1976; Agergaard and Laue, 1993). As both compounds are lipophilic in the past studies their levels have been primarily analyzed from adipose tissue samples. The backfat is the most widely sampled adipose tissue to determine androstenone and skatole levels and is considered to be representative of the whole pig. In the literature what is referred as backfat is subcutaneous fat taken from the P2 site, approximately 65 mm from the dorsal mid-line at the level of the last rib. This has been commonly used because subcutaneous fat accounts for 60-70% of the total body fat (Kouba et al., 1999) and sampling from the P2 site can be easily done. Some studies have looked at the possibility of using different kinds of tissue samples to determine the levels of skatole and indole; for example serum, plasma, submaxillary salivary gland or faeces (Claus et al., 1993; Tuomola et al. 1996), but there is not a good alternative for androstenone. It has been demonstrated that the chemical composition and the development of fat tissues vary in the pig (Wood et al., 1986) and for this reason it could be speculated that skatole and androstenone levels vary in different adipose tissues. Hansen et al. (1994) found significantly different concentrations of skatole and indole between the inner and the outer layer of backfat, while in the study by Rius and García-Regueiro (2001) the levels of skatole and indole in backfat and the fat covering Longissimus dorsi were significantly correlated. There is no documentation that shows if there is any differences in the levels of androstenone and skatole in the fat from different sites of the carcass. In a first study conducted in our University (data not published) the levels of androstenone and skatole from 68

69 backfat and minced fat (77:23 subcutaneous: intermuscular fat) were significantly correlated (r = and respectively; P 0.001). This study was therefore conducted to investigate whether the concentrations of boar taint compounds vary in the different adipose tissues of the carcass and if different sites can be consequently used for sampling. 4.3 Materials and methods A total of 20 entire male pigs, of unknown breeds, from 8 different farms were sampled in a commercial abattoir near Bristol during a working day. For each pig, samples of subcutaneous fat from the P2 area (referred to as backfat), dorsal neck region and cheek were collected, individually labelled and bagged. The samples then were taken to the University of Bristol and stored at -20ºC until analysis. For each pig hot carcass weight and P2 fat thickness were also recorded. The backfat sample collected from each pig was an area of skin plus attached fat all the way down to the intersection with the meat of max 15x5 cm, taken near the P2 area. The fat samples from the neck were of the same type and size of the samples of backfat, taken from the cervical region close to where the head is removed. The last fat sample collected from each pig was a square area of 10x10 cm of skin plus attached fat taken from the centre of one cheek. This site was examined because in some EU countries, cheek tissue is used in abattoir heating tests to generate odours to determine if carcasses are suitable for human consumption. From each sample skatole and indole concentrations were measured using the simultaneous distillation-extraction procedure followed by GC analysis, according to the methodology described by Annor-Frempong, et al. (1997b) (see Appendix 1). Androstenone concentrations were measured using a modification of the high 69

70 resolution gas chromatographic procedure of De Brabander and Verbeke (1986) (see Appendix 1). The data of the compounds levels were subjected to simple analysis, analysis of variance (ANOVA), with the sampling site as a factor to determine whether there appeared to be any significant differences between the mean values for the compounds concentrations, and correlation analyses between the 3 sampling sites using SPSS computer software. Also, to determine if there was good agreement between the compounds levels in the different sites, the comparison method of Bland and Altman (1986) was used, and the differences between the taint measurements and the 95% limits of agreement were calculated using Microsoft Excel The calculations of crude 95% lower and upper limits are as follows: Lower crude limit of agreement = Mean of differences 1.96 x Standard deviation. Upper crude limit of agreement = Mean of differences x Standard deviation. The analysis of variance was conducted using also the logarithms of the data to reduce the variability, but as the results show no differences with those of the arithmetic values the remaining analysis were conducted without the use of logarithmic transformation. 70

71 4.4 Results and discussion Analysis of the mean values obtain from the 3 samples from each pig reveal an overall average for the 20 pigs of 0.11 µg/g of fat for skatole and 1.16 µg/g of fat for androstenone (Table 4-1). The average value for indole was 0.04 ± 0.03 µg/g of fat (mean ± standard deviation). Table 4-1. Concentration of the taint compounds in the 20 pigs (µg/g fat). Skatole Androstenone Mean ± Standard Dev ± ± Geometric mean Median Minimum Maximum There were 9/20 (45%) pigs with an average concentration of androstenone above the threshold value of 1.0 µg/g, and 5/20 (25%) pigs with an average concentration of skatole above the threshold of 0.2µg/g fat. The values for skatole are not so different from the mean values that can be found in the literature and from the results presented in the previous chapter (0.14 µg/g), but the androstenone levels were higher than expected, but lower than the previous study (1.65 µg/g). There was wide variability in the compounds levels between the pigs, as shown in previous work on boar taint, and the individual results from the 3 sites of sampling for skatole and androstenone are shown in Figure 4-1 and 4-2. In these, each pig is represented by 3 bars, and the thresholds levels for the compounds are represented by the red horizontal line. Values for individual pigs are given in the Appendix 2. 71

72 µg/g of fat µg/g of fat Back Fat Neck Fat Cheek Fat Figure 4-1. Concentrations of skatole in the 20 pigs Backfat Neck fat Cheek Fat Figure 4-2. Concentrations of androstenone in the 20 pigs. 72

73 The mean for hot carcass weight was 76.8 ± 7.5 kg, slightly lower than the UK national mean of 78 kg (BPEX Annual Technical Report 09-10), and the P2 value was 12 ± 1.6 mm. There was a significant positive correlation (r=0.462; P<0.05) between the skatole and androstenone levels, as many studies have already shown (Bonneau et al., 1992; Zamaratskaia et al., 2004). However, the concentrations of skatole and indole were not correlated, although many studies have reported positive correlations between these compounds (Hansen et al., 1994; Tuomola et al., 1996; Rius and García-Regueiro, 2001). The relationships between androstenone and skatole concentrations and carcass weight and P2 fat thickness are shown in Tables 4-2 and 4-3. These show that for skatole there were no correlations with both weight and P2 of the pigs. For androstenone there was a significant positive correlation (r=0.45; P < 0.05) with the carcass weight of the pigs but not with P2. These results are similar to those reported in Chapter 2.So the higher values of androstenone found here compared with previous studies can be partly explained by the weight of the carcasses, but as the breeds of the pigs are unknown we cannot exclude the influence of the genotype of the animals because different studies have shown the importance of the breed for the androstenone levels (Xue et al., 1996; Hortos et al., 2000; Babol et al., 2004). Table 4-2. The correlation of taint compounds with carcass weight. Correlation Coefficient (r ) P-value Skatole Androstenone <0.05 Indole

74 Table 4-3. The correlation of taint compounds with P2. Correlation Coefficient (r ) P-value Skatole Androstenone Indole The results obtained from the analysis of variance (Table 4-4), on both arithmetic and logarithmic values; demonstrate that there was no difference in the average value of each compound in the fat from the three different site of the carcass. This suggest that in addition to backfat other sites of the carcass could be used for the collection of fat samples for boar taint evaluation, and that the mean values of boar taint compounds in backfat are representative of overall adipose tissue concentrations. However, the ANOVA analysis does not compare the individual measurements of the taint compounds for each pig from the different fat samples necessary to confirm a good agreement between them. For that further statistical methods were applied, without using logarithmic values as no differences were found in the ANOVA analysis. Table 4-4. The effect of the site of adipose tissues on taint compounds. Tissue site p-value Variable Backfat Neck Cheek Skatole (μg/g) Log-skatole -1,200-1,185-1, Androstenone (μg/g) Log-androstenone 0,042-0,010-0, Indole (μg/g) Log-indole -1,577-1,517-1,

75 Following the method of comparison by Bland and Altman (1986), the first step to study the agreement between two measurements was to plot the data in a graph and see if the line of equality, normally used when comparing two methods measuring the same variables, and the linear regression fit. This could help the eye in checking the degree of agreement between the measurements; although a good linear relationship does not imply that they are in agreement. If the data of the measurements will be scattered about and close to the line of equality which passes trough zero, we could expect a good agreement between the two measurements. The graphs (Figure 4-3, 4-4 and Appendix 2) obtained in this study suggest that the measurements of skatole in the fat from the neck seem to slightly overestimate its concentration compared with those of backfat (Figure 4-3) and cheek; while the measurements of androstenone in backfat seem to slightly overestimate its concentration compared with those of the fat from neck (Figure 4-4) and cheek. 75

76 Figure 4-3. Linear regression (red line) of skatole (μg/g) in backfat and neck along with the line of equality (black line). Figure 4-4. Linear regression (blue line) of androstenone (μg/g) in backfat and neck along with the line of equality (black line). 76

77 The second step was to calculate the correlation coefficient between the different pairs of values (backfat - neck; backfat - cheek; neck - cheek) of the compounds to observe the extent of the association between the different carcass sites. It can be seen from Table 4-5, 4-6 and 4-7 that for all the taint compounds analyzed the measurements from the three different carcass sites were positively correlated with each other. This confirmed the results of the ANOVA test which showed that there are positive associations of the compounds levels between backfat depot and the other two sites, and also between the fat from the neck region and the cheek. So we can assume that when the levels of taint compounds increase in backfat, they also increase in the whole carcass. The correlation coefficients of skatole (0.979, and 0.993; P < 0.01) and indole (0.960, and 0.958; P < 0.01) are similar to those found by Rius and García-Regueiro (2001), who reported that the levels of skatole and indole in the backfat are highly associated with those in fat covering the L. dorsi muscle (r = 0.99; P < 0.001). Table 4-5. The correlation of taint compounds between backfat and fat from the neck. Correlation Coefficient (r ) P-value Skatole < 0.01 Androstenone < 0.01 Indole < 0.01 Table 4-6. The correlation of taint compounds between backfat and fat from the cheek. Correlation Coefficient (r ) P-value Skatole < 0.01 Androstenone < 0.01 Indole <

78 Table 4-7. The correlation of taint compounds between fat from the neck and the cheek. Correlation Coefficient (r ) P-value Skatole < 0.01 Androstenone < 0.01 Indole < 0.01 The final step was to calculate the average of the differences between all pairs of values (backfat - neck; backfat - cheek; neck - cheek) of the compounds and the crude 95% limit of agreement (Table 4-8, 4-9 and 4-10). If the differences are normally distributed, 95% of the differences will lie between these limits. For both skatole and indole the average of the differences between the three sites of sampling were close to zero, confirming that the levels of these compounds do not vary in the fat from different part of the carcass. For example, the average difference between backfat and fat from the neck for skatole is (µg/g fat) and most of the backfat measurements will be positioned between units below and units above the measurement of fat from the neck. In addition, this and the other two results for skatole confirm what we could notice from the scatter plot graphs, that the measurements from the neck were slightly higher compared with the other two sites. For androstenone the average of the differences between the three site were not close to zero as those for skatole and indole, but this was expected as the concentration levels in the fat of androstenone are usually five times higher than skatole (e.g. androstenone threshold is 1.0 µg/g and skatole threshold is 0.2 µg/g of fat). The crude 95% limits for androstenone were wide in all three pairs of measurements, suggesting that there is not a good agreement between the different sites. However, as the analysis for the taint compounds levels were conducted on fat tissue and not on pure fat it is possible that the variation in 78

79 the composition of the fat (Wood et al., 1986) could have had much more influence on the results for androstenone. Also for androstenone the trend noticed in the scatter plot graphs was confirmed, with higher values found in the backfat compared with those from the other two sites. Table 4-8. Descriptive statistics for differences (backfat - neck) and crude 95% limits of agreement (n=19). Variable Mean St.Dev. lower 95% upper 95% Indole Skatole Androstenone Table 4-9. Descriptive statistics for differences (backfat - cheek) and crude 95% limits of agreement (n=19). Variable Mean St.Dev. lower 95% upper 95% Indole Skatole Androstenone Table Descriptive statistics for differences (neck - cheek) and crude 95% limits of agreement (n=19). Variable Mean StDev lower 95% upper 95% Indole Skatole Androstenone

80 A plot of the difference between two sites of sampling against their mean is a better method to assess between-site differences, compared to the simple plot of the measurements of one site against those of another. Figure 4-5 and 4-6 are examples for skatole and androstenone between backfat and the fat in the neck, the other scatter plot graphs are presented in Appendix 2. The red lines show the average of differences from the two sampling sites, so the closer they get to zero difference the better is the agreement between sites. The black dotted lines indicate the crude 95% limits of agreement. If the crude limits fit the data well, the data will be scattered randomly between these 2 lines. Figure 4-5. The plots of differences between skatole levels from backfat and neck. 80

81 Figure 4-6. The plots of differences between androstenone levels from backfat and neck. 4.5 Conclusions This work has shown a wide variability in the levels of skatole and androstenone in the pigs, as in the previous study. There was no significant difference in the average value of each boar taint compound in the fat from the three different sites of the carcass (backfat, neck and cheek). The similarity in the measurements was also confirmed by the significant (P < 0.01) positive correlation found between the different pairs of values (backfat-neck, backfat-cheek and neck-cheek) for all the compounds. A good agreement between the measurements was found in skatole and indole where the average of the differences between the sites were close to zero (between for skatole, and for indole); while these values for androstenone were bigger, especially those comparing backfat (0.125 vs. neck; vs. cheek) suggesting a possible overestimation from this site. As only 81

82 20 pigs were tested definite conclusions cannot be drawn, but these results suggest that other sites of the carcass of less economical value can be used for the measurement of the boar taint compounds instead of the most common site known as backfat. 82

83 5 Effects of pig growth rate and health status on boar taint and meat eating quality 5.1 Summary The main factors contributing to meat eating satisfaction are considered to be tenderness, juiciness and flavour. Several studies in the literature have shown controversial results in these characteristics when meat from boars and from females were compared (Wood et al., 1986; Von Lengerken et al., 1980; Barton- Gade, 1987; Ramsey et al., 1990; De Smet et al. 1996; Bonneau, 1998; Jeremiah et al., 1999). It is generally accepted that meat from female pigs is more tender and has a better palatability due to the absence of boar taint. Boar taint is an off odour/flavour that can be experienced if the meat from an entire male has high levels in the fat of at least one of the main compounds androstenone and skatole. However, the use of entire males is economically advantageous and widely practised in the UK, while in most other European countries the castration of the pigs at a young age is a common practise used to avoid the problem of boar taint. The use of entire males without the problem of boar taint in UK is usually explained by the lighter weight at which the pigs are commonly slaughtered compare to the rest of Europe. However, in the literature there is not strong evidence of a correlation between the boar taint levels and the weight of the pigs (Walstra et al., 1999). Growth rate is an important performance indicator in pig production. Herds with high growth rates normally achieve higher profitability than those with low growth rates. There is also evidence that faster growing pigs have superior tenderness 83

84 (Meat and Livestock Commission, 1989), possibly linked to faster muscle protein deposition through the activity and expression of proteolytic enzymes. Slow growth could also result in higher boar taint levels in entire male pigs because they will be older and sexually more mature at slaughter (Meat and Livestock Commission, 1989). The current project, consisting of two experiments, was therefore undertaken to investigate the effect of variations in growth rate on boar taint and other meat quality aspects, under experimental conditions and in commercial practice. The first experiment examined 225 pork loins from male and female pigs slaughtered at 90 and 110kg. The Large White x Landrace pigs were fed ad libitum at the University of Leeds and were subsequently divided into 3 growth categories: Fast, Slow and Interrupted. In the second experiment a total of 199 entire male pigs from four farms serving a commercial abattoir in the UK were examined only for boar taint levels. Growth rate had a marked effect on meat quality but not on boar taint. In the first experiment the Fast group had more tender pork when this was measured objectively or subjectively using a taste panel. The effect was present at both slaughter weights and in both sexes although it was more marked at 110kg and in males. There were no differences in meat quality aspects between the two sexes. There was no evidence that slow or interrupted growth adversely affected odours or flavours associated with boar taint. In fact the concentrations of both skatole and androstenone were higher in the faster growing male pigs in both experiments. 84

85 5.2 Introduction Rearing entire male pigs has been shown to be advantageous because of a better commercial performance as the production costs are reduced, mainly due to a better food conversion, and the higher lean meat satisfies the consumers demand (Walstra 1974). However, in most of the European countries the castration of the pigs at a young age is a common practise used to avoid the problem of boar taint. Boar taint is an off odour/flavour, mainly presented at the moment of cooking, due to high levels in the fat of at least one of the main compounds responsible for it, androstenone and skatole. Androstenone is a pheromone produced in the testis, mainly described having a urine-like odour, and skatole is a product of the microbial fermentation in the hind gut, characterised by a faecal-like odour. In the UK most of the males are reared without the use of castration and it is widely accepted that the boar taint problem is limited by the light weight at which the pigs are commonly slaughtered. However, in the literature there are not strong evidences of a correlation between the boar taint levels and the weight of the pigs. In the European study by Walstra et al. (1999) more than 4000 pigs from 6 countries with carcass weight ranging between 49 and 104 kg have been tested, and the correlation between skatole and androstenone concentrations with carcass weight was about 0.1 for both compounds and there were small differences between countries; the highest correlation being 0.20 for androstenone and 0.13 for skatole. In a previous study conducted in the UK by MLC (Meat and Livestock Commission, 1989) the levels of taint compounds were compared at two carcass weight, 65 and 80 kg. There were no differences between the two weights for skatole levels and the percentage of animals over the threshold was 5% in both. 85

86 However, androstenone concentrations exceeding the threshold increased from 3% at the lighter weight to 8% at the heavier weight. From the results of these and other studies is not possible to conclude that the weight of the animal has an important influence on the levels of the taint compounds, and it is even more difficult to give a weight limit under which the boar taint problem could be solved. The use of entire males in the UK limits the exportation of pork and other pig meat products to the European market, especially to those countries where boar taint is considered a main problem in pork quality, although the EU directive 91/497/EEC allows the trade of meat from entire male pigs between member countries. However, this European directive gives a carcass weight limit of 80 kg over which the pig needs to be assessed for boar taint to prevent the entrance of tainted meat into the human food chain. The problems with this directive are that there is no scientific evidence supporting the limit on carcass weight of 80 kg and the absence of a universal test for detecting boar taint on the slaughterline. Flavour is not the only factor contributing to pork eating satisfaction as also tenderness and juiciness are considered fundamental for a positive experience. As well as boar taint other aspects need to be considered before the use of entire male pigs in meat production can be widely accepted. Boars are generally characterised by a leaner carcass compared with castrated and gilts, with the result of carcasses higher graded for thinner backfat, and the smaller development of intramuscular fat results in meat cuts that are more appealing for the consumer (Bonneau, 1998). The different composition of the fat in boars, with higher percentage of water and unsaturated fat (Wood et al., 1986), could be positive from a dietetic point of view 86

87 but the lower levels of intramuscular fat could lead to a tougher meat compare to castrate and gilts (Barton-Gade, 1987). The carcass weight should not be the only aspect to consider as growth rate is an important performance indicator in pig production. Herds with high growth rates normally achieve higher profitability than those with low growth rates. Fast growth rates in the finishing stage may be positive for eating quality through increasing tenderness. However, it is often speculated that boar taint will be higher in slow growing, older pigs, due to them being more sexually mature at slaughter (MLC, 1989). Variability in current growth rates in the British herd is high and anecdotal evidence shows variation in pork tenderness and flavour is also high. Part of the variation in growth rate and potentially in quality is linked to the interrupted growth caused by infection, that could also lead to more mature pigs and to potential problem with tainted carcasses. The current project, consisting in two experiments, was therefore undertaken to investigate the effect of pig growth rate and health status on boar taint and other meat quality aspects, under experimental conditions and in commercial practice. 5.3 Materials and methods Experiment 1 A total of 225 loins were examined from Large White x Landrace pigs involved in a 3 x 2 x 2 factorial experiment based at the University of Leeds. Factors were 3 growth categories (Fast, Slow and Interrupted), 2 sexes and two final live weights, 90 and 110kg. Growth rate was measured between weaning and slaughter with the pigs being fed ad libitum. At the outset, guidelines for Fast growth were > 600 g/d 87

88 to 90kg and > 700 g/d to 110kg. Guidelines for Slow growth were around 500 g/d to 90kg and 590 g/d to 110kg. Interrupted growth pigs were those that had suffered a period of weight loss presumed to be due to infection and hospitalised in separated pen. The pigs have been slaughtered in a commercial abattoir over a period time of one year; every time the pigs were slaughtered the loins were delivered after two days to the University of Bristol for analysis. On arrival, ph in Longissimus dorsi was measured with a Testo 230 ph meter. A steak was removed, placed with the cut surface uppermost on a foam tray, overwrapped with cling film and bloomed for 2 hr in a walk-in chiller maintained at 1ºC. Colour was then recorded using a Minolta colour meter. To measure drip loss, the muscle from a 25mm thick loin steak was suspended in a plastic bag in a cabinet maintained at 1ºC and the drip collected over 48 hr was recorded. The remainder of the loin was then conditioned at 1ºC so that the period from slaughter to the end of conditioning was 10 days. Loins were then blast frozen at -80ºC and stored at -20ºC before the rest of analysis was conducted. Toughness was measured following cooking a 10 cm piece of loin in a water-bath at 80ºC to an internal temperature of 78ºC. After cooling in ice, 10 cores of muscle 20 mm long and 10 mm x 10 mm in cross section were sheared using a Stable Microsystems Texture Analyser fitted with a Volodkevitch jaws. Average peak shear force of these 10 samples was the measure of toughness recorded. Skatole and indole concentrations in subcutaneous fat from the loin were measured in all the pigs using the simultaneous distillation-extraction procedure followed by GC analysis, according to the methodology described by Annor-Frempong, et al. (1997b) (see Appendix 1). Androstenone concentration in subcutaneous fat from 88

89 the loin was measured in the male pigs using a modification of the high resolution gas chromatographic procedure of De Brabander and Verbeke (1986) (see Appendix 1). On the day before sensory analysis, loins were thawed at room temperature then kept in a chiller at 1ºC. Steaks 2.5 cm thick were griddled to an internal temperature of 72ºC, then cubes of cooked muscle were presented to the 10- member taste panel for analysis of the following attributes on 8-point scales (1 = extremely weak, 8 = extremely strong): pork odour of fat, abnormal odour of fat, tenderness, juiciness, pork flavour and abnormal flavour. Finally, the panellists were asked to score flavour liking and overall liking. Data were statistically analysed using general linear models (GLM), with growth category, carcass weight group and sex as factors and including interaction terms, using Minitab Release 14 computer software Experiment 2 A total of 199 entire male pigs from four farms (referred to as: Farm A, Farm B, Farm C and Farm D) serving a commercial abattoir in the UK were examined for boar taint levels. From each farm 25 loins of the first pigs (Pull 1) in the batch to finish (24 for Farm B) and 25 of the last (Pull 2) of the batch to finish were collected to be analysed at University of Bristol. Subcutaneous fat samples from the loins were tested for the boar taint compounds concentrations with the same methods as in the first experiment. 89

90 The data of the compounds levels were subjected to analysis of variance (ANOVA), with batch and farm as factors to determine whether there appeared to be any significant differences in boar taint compounds between pigs of the same group getting to slaughter at different time and to test the variability of the compounds levels between farms. Statistical analysis was performed using SPSS computer software. 5.4 Results and discussion Experiment 1 The final numbers of pigs in the growth categories were: 54 males and 42 females in the Fast group; 36 males and 50 females in the Slow group and 22 males and 21 females in the Interrupted group; for a total of 113 males and 112 females tested. The weaning to finishing average daily gain (ADG) on which the classification into growth categories was made, was for the Fast growth category 0.66 kg/d up to 90 kg and 0.70 kg/d up to 110 kg (average 0.68 kg/d). These values were as originally planned. For the Slow category, the ADG was 0.52 kg/d to 90 kg and 0.55 kg/d to 110 kg (average 0.54 kg/d). These values were also similar to those planned. As we expected, in the Interrupted pigs the average growth rate was even lower than the Slow category with 0.47 kg/d to 90 kg and 0.54 kg/d to 110 kg (average 0.50 kg/d). The growth rates and fat thickness measurements of all pigs are in Table 5-1, which shows that the daily growth of the animal has a significant effect on the fatness, in this case on the P2 fat thickness. 90

91 Table 5-1. Growth rate and P 2 fat thickness in growth category groups. Growth category p-value Fast Slow Interr Finisher ADG (kg) 0.74a 0.57b 0.56b <0.05 Wean to finish ADG (kg) 0.68a 0.54b 0.49c <0.05 P 2 (mm) 10.13a 9.06b 8.43b <0.001 Numerous studies have established that the production of entire males is economically advantageous as they need less feed, grow faster and have leaner carcasses (Walstra and Kroeske, 1964; Walstra, 1974; Fowler et al, 1981; Andersson et al 1997); in this study there was a significant difference in the daily growth of the males compared to the female, especially in the fast growing pigs (Table 5-2 and 5-3). However, there was no difference in the fatness of the carcasses as the fat thickness P2 values in the two sexes were similar. Table 5-2. Differences between the 2 sexes (all 225 pigs) in growth rate and P 2 fat thickness. Male Female p-value Finisher ADG (kg) <0.05 Wean to finish ADG (kg) <0.005 P 2 (mm) Table 5-3. Differences between the 2 sexes in growth rate and P 2 fat thickness in the Fast group. Male Female p-value Finisher ADG (kg) <0.001 Wean to finish ADG (kg) <0.001 P 2 (mm) The meat quality measurements for all 225 pigs recorded on the Longissimus dorsi (loin) muscle are in Table 1-4. For some characteristics, there were carcass weight group x growth category interactions indicated by. There were small differences 91

92 in colour and drip loss between the growth categories, the samples from Fast growing pigs tending to be paler. There was a strong tendency for toughness to be lower in the Fast growing group but the interaction prevented this being a statistically significant effect. The reasons for the interactions between carcass weight group and growth category are explained in Table 5-5 (as for Table 5-4, these are mean values for males and females). Differences were much more marked in the 110kg pigs. In these pigs, the Fast growth group had less drip, less intensely coloured muscle. Toughness was much higher in the Slow and Interrupted growth groups than in the Fast group. In the 90kg pigs, toughness was also higher in the Slow and Interrupted groups, particularly in the latter although growth category differences were smaller than at 110kg. Drip and colour parameters were not different although muscle ph was lower in the Fast growth category group. Table 5-4. Effects of growth category on meat quality in the loin muscle. Growth category vr p-value sig. sed Variable Fast Slow Interr ph 5.40b 5.43a 5.45a *** Drip (%) ns L* 54.1b 54.5b 55.6a ** a* ns b* ns c ns H 33.1b 34.3ab 35.7a * 1.01 Toughness (kg) ns Significant carcass weight group x growth category interaction. Therefore vr, sed and significance recalculated using interaction error term. 92

93 Table 5-5. Effects of growth category on meat quality in the loin muscle in 90 and 110 kg carcass weight groups. 90 kg Pigs Growth category Variable Fast Slow Interr p-value ph 5.38b 5.43a 5.47a <0.001 Drip (%) L* a* b* c* H Toughness (kg) 4.49b 4.81ab 5.35a kg Pigs Growth category Variable Fast Slow Interr p-value ph Drip (%) 4.09b 4.59ab 5.24a <0.001 L* a* 5.78b 6.30ab 6.67a b* 3.73b 4.32ab 4.97a c* 6.90b 7.66ab 8.33a h 32.3b 34.4ab 36.3a Toughness (kg) 4.32b 5.63a 5.47a <0.001 The importance of growth rate on tenderness was confirmed after the 113 males were analysed as there was a significant difference (P < 0.001) between the values of toughness in the Fast growing pigs (4.42 kg) and the other two groups (5.38 kg Slow; 5.56 kg Interrupted). Between all the male and female pigs the differences in the meat quality measurements were small and not significant; differences that were confirmed to be not significant between the sexes also in the three different growth rate categories. There are controversial results in the literature regarding 93

94 the effect of gender on meat quality of pork. Most of the studies reported that gender has no influences on carcass composition and on meat quality aspects such as tenderness, juiciness and palatability (Von Lengerken et al., 1980; Barton-Gade, 1987; Ramsey et al., 1990; De Smet et al. 1996; Jeremiah et al., 1999). However, other reports have found significant differences between sexes in ultimate ph (Eikelenboom and Hoving-Bolink, 1993), with gilts having higher values, and carcass composition (Angelov and Apostolov, 1995), gilts carcasses being fatter than those of boars. The differences between the results from this study and those from the literature could be due to the different breeds used. Even if skatole is a product of the fermentation in the hind gut of the pig and then produced in both sexes in the same amount (Agergaard and Jensen, 1993), its level in the fat is influenced by the presence of androstenone, a pheromone produced in the testis, that interacts with skatole metabolism in the liver (Babol et al. 1999; Doran et al., 2002) with the results of higher levels of this compounds in entire males compared to castrated and female pigs. The concentration of skatole in loin backfat was significantly (P < 0.005) higher in the males (0.07 μg/g of fat) than in the females (0.05 μg/g of fat); while there was no difference in indole levels (0.04 μg/g in both). The concentrations of skatole and androstenone in backfat of males only are shown in Table 5-6 and 5-7. Table 5-6 shows that the concentrations of both compounds were highest in the Fast growth category group nullifying the hypothesis that a slow growth could lead to the problem of boar taint. While there were no significant differences in taint compound concentrations between 90 and 110kg live weight groups in males, confirming the results found by Walstra et al. (1999) that the weight of the pig has no influence on boar taint levels. However, the 94

95 average values for both skatole (0.07 μg/g) and androstenone (0.44 μg/g) for all the male pigs are lower to those found in the commercial pigs reported in the previous chapters and with those that can be found in the literature (e.g. Walstra et al.,1999; skatole 0.13 µg/g, androstenone 0.81 µg/g). Furthermore only one pig on 113 males (0.8%) had skatole value over the threshold of 0.20 µg/g and there were 3/113 (2.6%) pigs with an average concentration of androstenone above the threshold value of 1.0 µg/g. The breeds used in this experiment, Large White and Landrace, are most probably the cause of these low levels in the fat as different studies have reported that these two breeds are characterized by low levels of taint compounds (Squires, 1992; Xue et al., 1996; Pedersen, 1998; Hortos et al., 2000; Babol et al., 2004). Table 5-6. Effects of growth category on skatole and androstenone concentrations in males. Growth category Variable Fast Slow Interr p-value Androstenone µg/g 0.500a 0.383b 0.397b Skatole µg/g 0.080a 0.056b 0.060ab Table 5-7. Effects of carcass weight group on skatole and androstenone concentrations in males. Weight group (kg) p-value Variable Androstenone µg/g Skatole µg/g The results for eating quality obtained with the taste panel are in Table 5-8 which shows the pooled results for sexes and carcass weight groups. The score for tenderness was higher in the Fast growing pigs than in the Slow and Interrupted groups; however the carcass weight group x growth category interaction prevented 95

96 this being expressed as a statistically significant effect. Juiciness was significantly higher in the Fast growth group and overall liking tended to be higher. The differences between growth categories were greater in the 110 kg group which explains the interaction between carcass weight group and growth category group. At 110kg, juiciness and overall liking were significantly greater (P < 0.005) in the Fast group than in the Slow and Interrupted groups. Numerical differences were even greater for tenderness but the sex x carcass weight group interaction prevented a statistically significant result being expressed. The means for tenderness and juiciness were in the same direction in the 90kg pigs (higher values in the Fast group) but differences were not significant. Table 5-8. Effects of growth category on eating quality of griddled loin steaks (1 to 8 scales) in all pigs. Growth category Fast Slow Interr vr p-value sig. sed Pork odour of fat ns Abnormal odour of fat ns Tenderness ns Juiciness 4.59 a 4.34 b 4.37 b * Pork Flavour ns Abnormal Flavour ns Hedonic Flavour liking ns Overall liking ns Significant carcass weight group x growth category interaction, so vr, sed and significance recalculated using interaction error term. As for the quantitative measurements of meat quality (i.e. shear force, ultimate ph) also the results in the literature for qualitative aspects of meat quality from taste panels and consumer surveys on differences between sexes are controversial. It is not entirely correct to compare the results from different taste panels and 96

97 consumer surveys as they are conducted using different methods and conditions; however it helps to give an idea of the real situation. There is speculation that meat from boars is less tender than meat from gilts or barrows and this is confirmed in studies which reported that consumers found meat from boars to be the lowest in tenderness (Martin et al. 1968; Cliplef, 1970; Gullett et al., 1993; Ellis et al., 1995). However, Jeremiah et al. (1999) reported that their sensory panel rated cuts from boars higher in both initial and overall tenderness than cuts from gilts and barrows. In this study there were no significant differences in tenderness between sexes when all the pigs were compared and when the sexes were compared between the different growth and weigh groups. Among the eight attributes evaluated by the taste panel here, only juiciness and abnormal odour presented significant differences between boars and gilts (Table 5-9). The meat from boars had higher values for juiciness than the meat from gilts (P < 0.05), this results agree with those reported by Ellis et al. (1995) and Jeremiah et al. (1999) but not with the consumer results reported by Gullett (1993). Jeremiah et al. (1999), using a 7-point hedonic scale, suggested that a difference of one full panel unit is required for consumers to detect a difference in meat quality. In this study an 8-point scale was used and the differences between the sexes in the juiciness values were of the magnitude of 0.2 panel units, therefore we can consider this difference of little practical importance. Regarding the results more correlated with the boar taint problem, only abnormal odour was significantly different (P < 0.05), with higher values in boars compared with the female pigs. As we could expect as consequence of the low levels in the fat of the taint compounds, also the mean values for abnormal odour were low, in the region of 2.4/8 (boars = 2.5; females = 2.3), with a small difference of 0.2 panel unit between the sexes that makes this result not 97

98 practically important. These results agree with those of Ellis et al. (1995), who also observed a more bland flavour in meat from gilts, while Jeremiah et al. (1999) found no differences in flavour intensity or desirability and in overall palatability. Table 5-9. Eating quality of griddled loin steaks (1 to 8 scales). Comparison of male and female pigs. Sex Male Female p-value Pork odour of fat Abnormal odour of fat Tenderness Juiciness Pork Flavour Abnormal Flavour Hedonic Flavour liking Overall liking Experiment 2 Analysis of the fat samples from the 199 male pigs obtained from the four commercial farms revealed an overall average of 0.08 µg/g of fat for skatole, 0.46 µg/g of fat for androstenone and 0.06 µg/g of fat for indole. There were 17/199 (8%) of pigs with skatole concentration above the 0.2 μg/g threshold for this compound and 16/199 (8%) of pigs with androstenone concentration above the threshold of 1.0 μg/g, while only 5/119 (2%) of pigs had both the compounds levels over these thresholds. These average values for skatole and androstenone are similar to the average values found in the Experiment 1 and the percentages of pigs over the thresholds are similar to those found in previous UK studies (Meat and Livestock Commission, 1989; Babol and Squires, 1995) but lower than those 98

99 μg/g of fat reported by Walstra et al. (1999) and those found in the farms analysed in chapter 2. The level of skatole in the 100 pigs of Pull 1 (0.11 μg/g) was significantly higher (P < 0.005) than the mean value of skatole in the 99 pigs of Pull 2 (0.05 μg/g), as found in the first experiment (Figure 5-1). Also the level of indole was higher (P< 0.001) in the pigs of Pull 1 (0.08 μg/g) compared to the level in Pull 2 pigs (0.05 μg/g). While in this experiment the levels of androstenone in the two pulls were similar with an average value of 0.46 μg/g in Pull 1 and 0.47 μg/g in Pull Pull 1 Pull 2 Indole Skatole Androstenone Taint Compound Figure 5-1. Levels of taint compounds in the 2 pulls. 99

100 Table Concentration of taint compounds (μg/g) in the 4 farms for the 2 pulls. Farm/Pull Indole Skatole Androstenone A A B B C C D D Table 5-10 shows that the differences in the taint compounds between the two pulls were not similar if we look at the individual farms. The skatole levels were higher in Pull 1 in Farm A, Farm C (P < 0.005) and Farm D, while in Farm B the levels were higher in Pull 2 compared to those in Pull 1 but not significantly so (Figure 1-2). In Farm A and C the 20% of the pigs from Pull 1 were over the threshold level of 0.2 μg/g, while in Farm D only the 4% were over and in Farm D none of the pigs was over the threshold; while the percentages of pigs were lower in Pull 2 in Farm A and C (4%), was the same in Farm D and as expected higher in Farm B (13%). As for skatole the trends for the concentrations of androstenone between the two pulls were different in the farms (Figure 5-3), but there were no significant differences in these levels. Androstenone was higher in the pigs of Pull 1 in Farm B, Farm C and D, while in Farm A the levels of androstenone were higher in Pull 2. Also the number of pigs above the threshold of 1.0 μg/g of fat for androstenone varied between farms and pulls. In Farm A 4% of the pigs of the first pull were over the threshold against 20% of pigs for the second pull; in Farm B there was no pig in the first pull over the threshold and only 4% in Pull 2; in Farm C in both pulls 8% of the pigs were tainted for androstenone and finally in Farm D

101 Androstenone μg/g fat Skatole (μg/g) and 8% of the animals, respectively for Pull 1 and Pull 2, had concentration above the threshold Pull 1 Pull Farm A Farm B Farm C Farm E Figure 5-2. Levels of skatole in the 4 farms Pull 1 Pull Farm A Farm B Farm C Farm E Figure 5-3. Levels of androstenone in the 4 farms. These results confirm the variability in the levels of the taint compounds that exist in the pig British population found in the previous chapters and also reported in the literature. The results from this experiment also confirm those of the first one 101

102 that showed as the growth rate of the pig, in this case the first and the last ones of the batch to be slaughtered, has no influence on boar taint levels and contradicting the popular idea that a slow growth leads to higher probability of boar taint. 5.5 Conclusions This study with the first experiment has shown important differences in eating quality between pigs allowed to feed ad libitum and differing naturally in growth rate. Pigs growing fast between weaning and slaughter, at around 0.68 kg/d, had significantly more tender pork than those growing at 0.54 kg/d. The difference was more marked in pigs slaughtered at the heavier weight of 110 kg but was also present at 90kg, and when only the 113 male pigs were analysed. The taste panel results also showed that juiciness of pork was increased by faster growth and as a result the overall liking score for eating quality was higher; however, in this case there were no significant differences for tenderness. It is often speculated that meat from female pigs is superior in quality, being more tender, with better taste and without the boar taint problem. This study showed no significant differences in both quantitative and qualitative aspects of meat quality when the meat from boars and females were compared. The only two aspects with small significance (P < 0.05) were juiciness and abnormal odour, but the differences between the sexes were on the range of 0.2 panel unit which makes these results not practically important. The meat from boars had higher values for juiciness than the meat from gilts, as shown in previous studies (Ellis et al., 1995; Jeremiah et al., 1999). Also the values for abnormal odour were higher in boars, but in this case as there were no differences for the values of flavour liking and overall liking the problem of boar taint can be excluded. 102

103 Even if the concentration of skatole in loin backfat was significantly (P < 0.005) higher in the males (0.07 μg/g of fat) than in the females (0.05 μg/g of fat), these values and those of androstenone in the boars (mean = 0.44 μg/g) are considered too low to manifest a problem of boar taint in these meat samples. There was no evidence that slower growth increased boar taint in both experiments. In experiment 1 concentrations of both androstenone and skatole were higher in the fast growth groups and scores for flavour attributes were not different between the groups. In experiment 2 the levels of skatole were higher in the first pull, while there were no differences in the levels of androstenone between the pulls. Both the experiments showed wide variability in the levels of skatole and androstenone in the pigs, as in the previous chapters, but with average values lower than the previous, confirming that genetics and farm management factors have a big influence on boar taint. These conclusions concerning tenderness and boar taint are similar to those reached in the MLC study of genotypes and growth rates (Meat and Livestock Commission, 1989) although the absolute values for growth rate in Experiment 1 were lower and more relevant to modern commercial pigs and Experiment 2 was conducted on commercial farms. The results should encourage producers to aim for growth rates above 0.7kg/d from weaning to slaughter for reasons of both lower production costs and better eating quality. 103

104 6 Effect of dietary chicory on boar taint 6.1 Summary Boar taint is a problem that can occur during the cooking or eating of meat or meat products from entire male pigs. The main compounds, which accumulate in the fat, responsible for it are androstenone and skatole. Androstenone is a pheromone produced in the testis that increases after sexual maturity. Skatole is a product of fermentation in the hind gut and its level in the fat can be reduced by acting on managerial aspects on the farm. The most effective way to reduce skatole levels appears to be the use of diets rich in fibre or in fermentable carbohydrates. Different studies showed that chicory (Cichorium intybus L.), a source of a fermentable carbohydrate know as inulin, is significantly effective in reducing the skatole levels in faeces, blood and fat of pigs (Claus et al., 1994; Jensen and Jensen, 1998; Rideout et al., 2004; Byrne et al., 2006; Hansen et al., 2006). However, it is still not clear which is the minimum percentage of inulin necessary in a diet to be effective and the current project was therefore undertaken to see if a short feeding period with inclusion of chicory at different percentage levels before slaughter will be sufficient to significantly reduce the level of skatole and improve the sensory aspects of pork from entire males. After a preliminary study and a first feeding trial, in the main feeding trial a total of 360 entire male pigs were used to evaluate the effects of 1 and 2 weeks feeding period with chicory before slaughter on skatole and androstenone levels in backfat. The pigs had been divided into 4 groups fed different levels of chicory: 0% (control), 3%, 6% and 9% supplemented diet. For each group 30 entire pigs were 104

105 sampled at 3 different times: a first time (called week 0) to measure the base level of skatole and androstenone in all the pigs, then the supplement of chicory was introduced and the pigs were sampled after 1 and 2 weeks on the test diet. All 360 backfat samples from the neck region (backfat) were assessed for skatole concentration; androstenone was measured in 110 pigs (all 9% pigs). Samples of backfat were presented to a 10 member taste panel after cooking for sniff tests to determine if reducing skatole had also reduced boar taint. Chicory fed at the level of 9% for 2 weeks was successful in reducing skatole to a level well below the threshold for this compound (0.2µg/g) with only 1 pig with a skatole value over the threshold. In the 9% group there was a downward trend in skatole by 1 week and 55% of pigs had levels between 0 and 0.05µg/g, typical of levels in castrated males. The other levels of chicory (3% and 6%) were not effective. The concentration of androstenone increased slightly in the pigs fed 9% chicory after 2 weeks. There were no differences between the groups in the abnormal odour scores after 2 weeks. The 9% chicory group had values as high as in the other treatments. The results show that the inclusion of dried chicory in the diet for 2 weeks was effective in reducing skatole concentrations. However no improvement in odour scores occurred, probably because androstenone remained high. 6.2 Introduction Entire male pigs have a better commercial performance then castrates, mainly due to more efficient food conversion and higher lean meat percentage of the carcass. In most European countries but not in UK, however, male pigs are still castrated at a young age to avoid the potential problem of boar taint, an offensive odour/flavour that mainly manifests when the meat is cooked. It is due to high concentrations of 105

106 skatole and/or androstenone in the meat which are driven off during cooking. These compounds are derived from different metabolic processes. Skatole, exhibiting a faecal-like odour, is produced by fermentation of the amino acid tryptophan in the hind gut and androstenone, exhibiting a urine-like odour, is a pheromone produced along with testosterone as part of male sex hormone metabolism. Both compounds are metabolised in the liver, showing an interaction that influences their concentrations (Doran et al., 2002), and in most pigs levels in fat are below the thresholds where taint problems can arise. These are 0.2 µg/g of fat for skatole and 1.0 µg/g of fat for androstenone (Bonneau et al., 1992). However in some pigs, with percentages that are variable (Walstra et al., 1999), these compounds reach very high concentrations for reasons which are still not clearly understood. In the literature we can find several possible methods that could reduce the incidence of boar taint in slaughter pigs, but the right solution to the problem yet has to be found. Some methods have significant effects and others seem to have only marginal effects, but the main problem is that androstenone and skatole do not always respond in the same way to the measures. Fat androstenone levels are mostly affected by genetic factors controlling its production and excretion as well as by the degree of sexual maturity. The idea is commonly accepted that increasing slaughter weight results in animals being more sexually mature and consequently presenting high levels of androstenone, but the study by Walstra et al. (1999). In previous chapters in this thesis, conflicting results were seen, both positive relationships with body weight (Chapters 2 and 3) and no relationship (Chapter 4). This was possibly explained by the different genotypes used. Genetic selection 106

107 seems more efficient at lowering androstenone content (Bonneau, 2006; Jensen, 2006); however, even if the patterns of androstenone production and metabolism are know (Doran et al., 2004; Nicolau-Solano et al., 2006) more work is still necessary to identify the right genetic markers for low androstenone level. On the contrary as skatole production is due to the bacterial degradation of tryptophan in the large intestine, its level in the fat can be reduced by modulating management aspects of pig farming such as diet, feeding and rearing conditions. Several studies in the last twenty years (Claus et al., 1994; Jensen and Jensen, 1998; Rideout et al., 2004; Whittington et al., 2004; Zamaratskaia et al., 2005; Hansen et al., 2006; Kjos et al., 2010) showed how a change in the dietary composition, especially in the last weeks before slaughter, possibly through changes in bacterial activity affecting tryptophan (Hansen et al., 1997; Jensen et al., 1997) has a significant effect on skatole level. Different feed components have been tested; most of the studies have investigated the effects of diets rich in fibre, such as sugar beet pulp, or rich in fermentable carbohydrates that escape digestion in the small intestine, such as raw potato starch and chicory. In different studies the use of chicory (Cichorium intybus L.) has reduced skatole levels in faeces, blood and fat (Claus et al., 1994; Jensen and Jensen, 1998; Rideout et al., 2004; Byrne et al., 2006; Hansen et al., 2006). Chicory roots are particularly rich in a fructo-oligosaccharide, inulin, which is not digested in the small intestine and is able to alter the patterns of microbial fermentation in the large intestine with a consequent reduction in the production of skatole. In the past studies both chicory roots and pure inulin, extracted from chicory, have been tested with positive results in reducing the level of skatole; however, it is still not clear which is the minimum percentage of inulin necessary in a diet to be effective or how long the feeding period needs to be. 107

108 It has been demonstrated that androstenone and skatole are the main contributors to the problem of boar taint but the respective importance of each of them to it is still not clear. Some publications have underlined the importance of the level of androstenone for tainted pork (Squires et al., 1992; Babol et al., 1996), others have suggested skatole as the main contributor (Lundström et al., 1988; Bejerholm and Barton Gade, 1993). However, in a consumer survey carried out in seven countries Matthews et al. (2000) found that a high level of skatole has a bigger impact on consumer acceptability compared to androstenone and from another study by Annor-Frempong et al. (1997b) it also seems that there are interactions between skatole and androstenone that affect the sensory response, as skatole has been demonstrated to enhance the sensory perception of androstenone. It is possible that reducing skatole alone may be an effective way to control boar taint. The current project was therefore undertaken to see if a short feeding period with inclusion of chicory before slaughter will be sufficient to significantly reduce the level of skatole and improve the sensory aspects of pork from entire males from British commercial farms. 6.3 Materials and methods The work was conducted on farms supplying pigs to a major pig processing company in the East of England. The study consisted of three trials Preliminary study A total of 30 farms supplying a commercial abattoir were screened for levels of androstenone and skatole in the fat. The farms were all operated in a similar way including the use of common diets and similar genetics. Breeds represented were 108

109 Large White, Duroc and Pietrain and pigs on each farm were crosses between these, although the exact combination was unknown and likely to differ between farms. For each farm, samples of subcutaneous fat from the dorsal neck region (referred as backfat) were collected from a minimum of 22 to a maximum of 50 entire male pigs, of unknown breeds, slaughtered on the same day. The sample collected from each pig was an area of skin plus attached fat all the way down to the intersection with the meat of max 15x5 cm, taken from the cervical region close to where the head is removed. After an initial training visit to the abattoir, all samples were taken over several weeks by the staff of the abattoir. Samples per farm were individually labelled and bagged and stored at -20ºC before being sent in batches to University of Bristol (UoB). For each farm a single sample was obtained from all the pigs. After removing the skin and the subcutaneous glands, 2 g of fat, including all the layers, was obtained from each pig. The 100 g of fat obtained for each farm (when 50 pigs were sampled) was then blended together using a food processor Moulinette S (Moulinex ) and sub-sampled for analysis. In the obtained subsamples of fat, skatole and indole concentrations were measured using the simultaneous distillation-extraction procedure followed by GC analysis, according to the methodology described by Annor-Frempong, et al. (1997b) (see Appendix 1). Androstenone concentrations were measured using a modification of the high resolution gas chromatographic procedure of De Brabander and Verbeke (1986) (see Appendix 1). This study provided baseline information on the levels of taint compounds in this farming group. 109

110 6.3.2 First feeding trial In the first feeding trial the effects of feeding 5% dried chicory (Fibrofos 60, SOCODE, Cosucra Groupe Warcoing S.A., Belgium) in the finishing diet for two weeks before slaughter on 7 farms was compared with feeding a nonsupplemented feed on 6 control farms. Each farm was represented by 50 pigs and they were sampled and tested with the same methods as in the preliminary study. The level of 5% was based on results of previous research on chicory in the literature and views of the cost effectiveness of using Fibrofos 60, but this is not the amount recommended by the producing company (10%). Fibrofos 60 is a powder product obtained from chicory roots dried at low temperature and containing 60% of inulin on dry matter (D.M) Main feeding trial In the main feeding trial only one of the 30 farms from the preliminary study was tested. A total of 360 cross-breed growing-finishing pigs {[(Large White x Landrace) x White Duroc] x Pietrain}, all entire males, were fed 4 different levels of dried chicory: 0% (basal diet), 3%, 6% and 9% of diet. The pigs were raised in 4 different buildings (90 pigs each), with 18 pigs/pen, on straw based solid floors. For each group, 30 entire pigs were sampled at slaughter at three different times: a first time (called week 0) to determine the baseline level of skatole and androstenone in all the pigs that have received only the basal diet. Then the supplement of chicory was introduced and the pigs were sampled after 1 and 2 weeks on the test diet. From each pig a sample of backfat was collected, individually labelled and bagged, stored at -20ºC until all 360 pigs were sampled and then sent to UoB for analysis. Hot carcass weight and P2 fat thickness were 110

111 recorded for each pig. All 360 boars were individually tested for skatole concentration; androstenone was measured individually in 110 pigs (all 9% pigs and 20 pigs of 0% chicory, week 2), using the same procedures as before. All the samples were presented to a 10-members taste panel (all female) for sniff tests to determine pork odour intensity and abnormal odour intensity using 8 point category scales (1= extremely weak; 2 = very weak; 3 = moderately weak; 4 = slightly weak, 5= slightly strong, 6 = moderately strong, 7 = very strong, 8 = extremely strong). In addition, certain descriptive terms for specific odours, used by Annor-Frempong et al. (1997a), were assessed on scales. The descriptors for skatole were: mothballs and musty. Those for androstenone were: acrid, stale, sweaty, nose feel and piggy; then parsnip, stewed vegetables and sweetness that are typical for values below the recognition threshold. For cooking, each fat sample was cut into 10 approximately equal cubes, placed in a foil container covered with foil, and cooked in pre-heated ovens set at 200ºC for 15 minutes. Each cube was then removed and placed in a bottle maintained on a hotplate at 60ºC and presented to each member of the panel. Data were statistically analysed using general linear models (GLM), comparing the different levels of chicory in the diet and the duration of feeding, using Minitab Release 14 computer software. 6.4 Results and discussion Preliminary study The number of entire males collected from each farm varied from a minimum of 22 to a maximum of 50, with an average of 47 pigs sampled for each farm. The results 111

112 for skatole and androstenone levels in the 30 farms involved in the first trial are shown in Figure 6-1. Each red bar represents the skatole value and each blue bar the androstenone value for the sub-sample obtained from all the pigs of one farm, and the black horizontal lines represent the threshold levels for the compounds. There was wide variability between farms, as shown in the previous chapters. Only 4/30 (13%) farms had androstenone concentrations above the threshold value of 1.0 µg/g but 12 (40%) farms had skatole concentrations above the threshold of 0.2 µg/g fat. Overall 13/30 (43%) farms had at least one of the compounds over the threshold level and 3 (10%) farms had both values over the thresholds. The overall means were: 0.71 µg/g for androstenone and 0.19 µg/g for skatole (Table 6-1), and 0.09 µg/g for indole. The mean value for androstenone is similar to that found by Walstra et al. (1999) of 0.8 µg/g, lower to those of the farms in chapter 2 (1.65 µg/g) and 3 (1.16 µg/g) but higher than the values for the pigs in chapter 4 (0.44 and 0.46 µg/g). While the mean value for skatole is higher than that found by Walstra et al. (1999) and from those in the previous chapters; showing a hypothetical problem for these farms with skatole levels but ideal for testing the effects of chicory. There were different genetics and production systems on the different farms which could explain these variations and also the differences from the other studies. Table Concentrations of taint compounds in the 30 farms (µg/g fat). Skatole Androstenone Mean ± Standard Dev ± ± Geometric mean Median Minimum Maximum

113 1.4 Preliminary study Androstenone (µg/g) Skatole (µg/g) Figure 6-1. Skatole and androstenone concentrations in the 30 farms First feeding trial Analysis of the sub-samples of fat from the 13 farms tested in the first feeding trial revealed an overall average of 0.13 µg/g of fat for skatole and 0.43 µg/g of fat for androstenone. The results for skatole and androstenone levels involving all the 13 farms are shown in Figure 6-2, in which each bar represents the mean value for one farm whose backfat samples have been minced together. The bars from 1 to 6 represent the control farms and those from 7 to 13 are the farms in which 5% chicory was tested. The threshold level for skatole is represented by the black horizontal line. 113

114 Feeding trial Androstenone (µg/g) Skatole (µg/g) Figure 6-2. Skatole and androstenone concentrations in the 13 farms. The overall mean value for skatole in the farm tested for chicory was 0.11 µg/g; this was lower than the mean value for the control farms of 0.15 µg/g, however there was no significant difference between the two groups of farms. While the mean value for androstenone was higher in the chicory group (0.46 µg/g) than in the control group (0.39 µg/g) but not significantly. There was a tendency for the farm tested with chicory to have lower values for skatole; with 5/7 farms having values lower than 0.1 µg/g of fat typical of castrated pigs. This trend was considered a positive result and so we proceeded with the final trial, but to demonstrate a significant effect of chicory on skatole a bigger number of animals was necessary. Also the use of individual measurements instead of the sub-sample and the use of different percentages of chicory were introduced, to give a more detailed study of the effect of chicory on boar taint. 114

115 6.4.3 Main feeding trial The mean value for hot carcass weight of all 360 pigs was 78.1 ± 2.3 kg similar to the UK national mean of 78 kg (BPEX Annual Technical Report 09-10), and the mean value for P2 fat thickness was 13 ± 0.5 mm. For both the carcass weight and P2 there were no statistically significant difference between the 4 groups of pigs. There was therefore no effect of chicory on feed intake and daily gain, showing good palatability and utilization of chicory. As Fibrofos 60 contains 60% of inulin on D.M., the effective levels of it in the three diets (3, 6 and 9%) was: 1.8, 3.6 and 5.4% on D.M. Concentrations of skatole in the backfat of the 12 groups of 30 pigs (3 weeks x 4 levels of chicory, 360 pigs in total) are shown in Figure 6-3 and Table 6-2. The levels of skatole in the different groups were variable even though this trial took place on one farm with the same genetics and feeding systems in the different houses where the pigs were reared. This variability in the skatole levels is evident when we look at the concentrations in the 4 groups at week 0, when the animals were on the same diet, with a significant difference (P < 0.05) between the mean value of the 3% pigs and the 0 and 6% groups. If we look at the individual values of skatole also the percentage of pigs with values above the threshold of 0.2 µg/g of fat was different. In the 0% group 7/30 (23%) pigs were above the threshold, 16 (53%) in the 3% group, 9 (30%) in the 6% group and in the 9% group 6 (20%) pigs were above the threshold value. The mean skatole value for all the 120 pigs from week 0 was 0.23 µg/g and there were 38/120 (32%) pigs with skatole concentrations above the threshold. This mean value was similar to the one found in the 30 farms of the preliminary trial (0.19 µg/g) but much higher than the farms 115

116 Skatole (µg/g backfat) tested in chapter 4 (0.08 µg/g) and the mean value of 0.13 µg/g found by Walstra et al. (1999), confirming a potential problem with high skatole levels for these producers Week 0 Week 1 Week % chicory in diet Figure 6-3. Skatole levels in the 4 groups during the 3 weeks of sampling (horizontal line representing the threshold level). Table 6-2. Effect of feeding chicory on skatole levels (µg/g). Week Chicory levels 0% 3% 6% 9% p-value b a b ab a a a b p-value Superscript letters within the same row and numbers within the same column are significantly different (P < 0.05). After 2 weeks of feeding the 9% chicory diet, the level of skatole was significantly lower (P < 0.05) than in all the other groups. This effect was also significant (P < 116

117 0.05) within the same group between the pigs fed with chicory for 2 weeks and those fed for only 1 week and the boars of week 0. These results are similar to those found in a recent study by Kjos et al. (2010) in which a supplementation of 6 and 9% of chicory (70% of inulin) for a period of 4 weeks was significantly able to reduce the skatole concentration in both fat and faeces. In a previous study Hansen et al. (2006) have found a significant effect of chicory (crude and dry) or pure inulin on skatole levels in the fat but using higher percentages in the diet, from 10 to 25%, and for longer period of time, from a minimum of 4 weeks to a maximum of 9. However, in a subsequent study the use of lower percentages, 10 and 13.3, of dry chicory for 1 and 2 weeks had no effect in reducing the level of skatole in the fat (Hansen et al., 2008). In this study an increase in skatole concentration was present in the pigs fed only with the base diet (0%) during the 2 weeks of the trial. There were no differences in carcass weight between these pigs, which were reared in the same conditions and the pens were cleaned out 6 times a week. Differences in growth rate could have been responsible but this result would be in disagreement with those found in the two experiments presented in chapter 4, where a slow growth did not influence skatole levels in the fat. In the 30 pigs fed for 2 weeks with 9% chicory, the levels of skatole were well below the threshold for this compound, with only 1 pig (3%) with a skatole value over the threshold (Figure 6-4). The numbers of pigs exceeding the threshold value of 0.2 µg/g skatole in the 0, 3, 6% chicory groups were respectively: 19 (63%), 11 (37%) and 13 (43%). In the 9% group after 2 weeks 17/30 (56%) pigs had levels between 0 and 0.05 µg/g of fat, typical of levels in castrated males; while between 117

118 Percentage of pigs Percent chicory all the control pigs of week 0 only 8 (7%) had concentrations between these values (Figure 6-5). Dotplot of Skatole vs Percent chicory Skatole Figure 6-4. Distribution of skatole concentrations (µg/g of fat) in the 120 pigs fed different levels of chicory at 2 weeks control pigs (120) 9% pigs (30) Figure 6-5. Comparison of skatole levels (µg/g of fat) of all control pigs from week 0 with those fed 9% chicory for 2 weeks. 118

119 The concentration of androstenone in the pigs fed 9% chicory showed an opposite trend to skatole during the 2 weeks of the trial. The mean value after two weeks (1.39 µg/g) was significantly higher (P < 0.005) than those of the other two groups of pigs fed 9% (0.75 µg/g and 0.85 µg/g, respectively for week 0 and week 1) (Figure 6-6). The result of increasing values of androstenone during the 2 weeks of the trial is in disagreement with the results found in the two experiments in chapter 4, where the slow growth of the pigs did not correspond in higher values of androstenone. Between the 30 pigs on the 9% diet for two weeks, 15 had values of androstenone over the threshold of 1 µg/g of fat. The percentage of pigs over the threshold (50%) and the mean value for androstenone after two weeks of 1.39 µg/g of fat are similar to those found in the farms studied in chapter 2, but are much higher than the results obtained from the farm tested in the preliminary trial and from the pigs studied in chapters 3 and 4. As the pigs in this study had the same genetics, the high values for androstenone and the significant increase in the second week of the trial could be explained by the formation of a new hierarchy inside the pens after the 30 pigs have been slaughtered on week 0 and the other 30 in week 1. In some studies, the formation of a new hierarchy in a group of pigs and the food competition are responsible of increasing the androstenone levels in plasma and fat (Claus et al., 1994; Giersing et al., 2000). 119

120 Androstenone (ug/g fat) Week 0 Week 1 Week 2 Figure 6-6. Concentrations of androstenone in the groups fed 9% chicory for 0, 1 and 2 weeks. There was no influence of housing the animals in 4 different buildings on the sensory aspects of pork fat. At week 0 there were no significant differences between all 4 groups of animals in pork odour and abnormal odour evaluation and in all the descriptors used (Appendix 2). However, the values for abnormal odour in the 4 groups were in the region of 4 out of 8 while in the pigs of chapter 4 the values for this attribute was in the region of 2. The reason for this significant difference could be the higher values for skatole in this study but also the differences in the genetics of the pigs and the rearing conditions. Feeding chicory for 1 week did not have any significant effect on the sensory aspects of the fat. Table 6-3 shows the sensory results after 2 weeks of feeding chicory. There was no trend in the pork odour scores. The abnormal odour score was significantly (P < 0.001) lower in the 9% group than in the 0% controls, but there were inconsistent results for the 3 and 6% chicory groups. A clue to the reason of this small change in 120

121 abnormal odour score is shown by the tendency to increase present in the 9% group of some scores for the terms used to describe the odour of androstenone, like parsnip, related to the high concentrations of androstenone found in this group. However, the inclusion of 9% chicory for 2 weeks did have a clear effect in reducing odour descriptors as musty and mothballs (P < 0.05) which are typical of skatole. This result was confirmed by the cluster analysis which shows that the chemical determination of skatole is closely associated with the sensory descriptors for skatole (Figure 6-7). Also the 3 and 6% groups showed a shift from skatole but to a lesser degree. The results for the sensory panel in this study are in disagreement with those reported in the study of Byrne et al. (2008), in which higher levels of chicory (25%) and pure inulin (14%) had a significant effect in reducing off flavour related to boar taint and in increasing the overall liking score of the meat. Table 6-3. Sensory results after 2 weeks. Attribute/ Chicory levels p-value Descriptor 0% 3% 6% 9% Pork odour x ns Abnormal odour x 4.30 a 3.90 c 4.22 ab 4.04 bc <0.001 Stewed vegetables y ns Parsnip y ns Acrid y 22.5 a 18.7 b 19.6 ab 17.0 b <0.005 Musty y ns Mothballs y 11.2 a 8.2 b 9.2 ab 7.4 b <0.05 Stale Sweaty y ns Burnt Oil y ns Nose feel y ns Piggy y ns Sweetness y ns x 1-8 scales, y scales 121

122 Figure 6-7. Dendrogram with word linkage and correlation coefficient distance. 6.5 Conclusions The large variation in the boar taint compounds found between the 30 farms tested in the preliminary study has been seen in all the pigs in the previous chapters and is thought to be due to genotype and husbandry differences between farms. Twelve of the farms exceeded the threshold value for skatole taken to indicate tainted pork (0.2 µg/g) and 4 exceeded the threshold for androstenone (1.0 µg/g). Of the 4 farms high in androstenone, 3 also exceeded the threshold for skatole and the fourth was at the limit (0.19 µg/g). These high levels of skatole concentration, that have been confirmed in the main feeding trial too, could lead to problems for these producers with boar taint. This problem could be reduced by adding chicory to the diet for 2 weeks before slaughter. A value of 5% dried chicory of the commercial product Fibrofos 60, that correspond to 3% of inulin, had an evident effect on the skatole levels in 5/7 farms tested in the first trial, with values lower than 0.10 µg/g 122

123 typical of castrated male pigs. The results of the main feeding trial showed the same wide variation between groups observed in the other 2 trials. However, there was a clear effect of 9% chicory in reducing skatole to extremely low levels. After 2 weeks, the skatole concentration was on average 0.065µg/g in the 9% group. Only 1 pig out of 30 had a value greater than the threshold of 0.2µg/g compared with 19 given the 0% diet, 11 given the 3% diet and 13 given the 6% diet. This was very good evidence that the inclusion of 9% dried chicory in the diet for 2 weeks before finishing reduced skatole concentrations in backfat to a level typical of castrates. However no improvement in odour scores was detected by the panellists, probably because the levels in the fat of androstenone in the 9% increased after 2 weeks. Overall the results show that under in the commercial conditions of this trial, chicory was effective in reducing skatole but not boar taint. It is possible that as skatole declined, the perception of androstenone increased, causing no change in overall abnormal odours. But it is surprising that the sensory perception of boar taint was not reduced because several studies show that skatole is the taint compound most closely associated with boar taint and it has also been demonstrated that skatole enhances the sensory perception of androstenone (Lundström et al., 1988; Bejerholm and Barton Gade, 1993; Annor-Frempong et al., 1997b; Matthews et al., 2000). In other pigs, with lower concentrations of androstenone, the result may have been more positive for pork odour and flavour. However, it seems difficult to justify the extra cost to pig production of using chicory based on the current results, as it was estimated that the cost of a supplementation of 9% Fibrofos 60 for 2 weeks is around 3 per pig. 123

124 7 Comparison of heating methods for sensory assessment of boar taint in pig meat 7.1 Summary The trade of meat and meat products derived from entire male pigs is allowed in EU by the directive 91/497/EEC under the carcass weight of 80 kg, above which the carcass needs to be tested free from boar taint. Boar taint is due to high concentrations in the fat of at least one of the main compounds involved, skatole and androstenone. In the literature we can find several possible methods that could be used to detect boat taint on the slaughter-line but a universal test for detecting boar taint in the slaughterhouse has yet to be found. Four different cooking methods which could possibly be used in abattoirs were examined. They were: microwave, hotwire, boiling at 25ºC and 75ºC and melting. The methods were applied to 4 groups of backfat samples having different concentrations of skatole and androstenone. The groups of 10 samples were high skatole/high androstenone (HS/HA), HS/LA, LS/HA and LS/LA. The cut off between high and low was 0.2 µg/g for skatole and 1.0 µg/g for androstenone. From the pigs of the same groups, a composite sample comprising 80% head fat, 10% cheek muscle and 10% salivary glands was also cooked using the microwave and boiling methods. The odours produced during cooking were evaluated by a panel of 3 experienced assessors who evaluated abnormal odour and pork odour on 1 to 8 scales and skatole odour and androstenone odour on 0 to 8 scales. Backfat odours generated by all 4 methods were scored significantly differently between the extreme samples HS/HA and LS/LA. Samples high in only 1 compound (HS/LA 124

125 and LS/HA) were distinguished using the microwave, hotwire and boiling (75 C) methods. Samples high in skatole produced higher abnormal odour scores than those high in androstenone and correlations with abnormal odour were higher for skatole than androstenone. There was evidence that skatole enhanced androstenone odour. Scores for abnormal, skatole and androstenone odour were higher in composite samples containing 10% salivary glands than in backfat. The concentration of androstenone but not skatole was higher in this sample and use of it in an on-line abattoir test would overestimate the incidence of boar taint in comparison with backfat. As well as using the scoring scales (1 to 8 and 0 to 8), the assessors rated each sample as negative, moderately undesirable or highly undesirable in terms of odours. HS/HA samples were the most undesirable and, as with the scoring scales, HS/LA received more highly undesirable scores than LS/HA for both backfat and composite samples. Composite samples had a higher proportion of samples marked highly undesirable than backfat. Two of the assessors were more sensitive to the abnormal odours of boar taint than the third, a factor which would limit the usefulness of these cooking methods in abattoir tests. 7.2 Introduction In the European Union the trade of meat from entire male pigs is allowed with no restrictions for carcass weights lower than 80 kg by the Directive 91/497/EEC, while over this weight the carcass needs to be assessed for boar taint with an objective method. Boar taint is an off odour/flavour, mainly presented at the moment of cooking, due to high levels in the fat of at least one of the main compounds responsible for it, androstenone and skatole. Androstenone is a 125

126 pheromone produced in the testis, mainly described having a urine-like odour; skatole is a product of the microbial fermentation in the hind gut, characterised by a faecal-like odour. The problem of boar taint is considered important inside the EU as has been underlined also in the Regulation 852/2004/EC which states that meat manifesting a pronounced sexual odour is to be declared unfit for human consumption. The main problems are that both the directive and the regulation do not state the best method to detect boar taint, and a universal test for detecting boar taint on the slaughterline has yet to be found. There is also no scientific evidence supporting the limit on carcass weight of 80 kg. The absence of a specific regulation on the detection of boar taint explains the present situation in the EU where there is no harmonised method. So far only in Denmark a device for skatole quantification (Mortensen and Sorensen, 1984) has been commercially applied in most of the slaughter plants. Here a backfat sample is removed from the carcass, analysed with a spectrophotometric method for skatole and the results used on the production line to remove tainted carcasses. The limit of this automated system is that can only analyse a maximum of 180 samples per hour. Other problems are that androstenone is not measured and that not only skatole but also other indole-related compounds are measured, resulting in a possible overestimation of boar taint incidence for skatole-like compounds and underestimation for androstenone. Bonneau and Russeil (1984) suggested that the measurement of the length and size of the bulbourethral glands could be use to predict the incidence of boar taint in the carcasses. Later Babol et al. (1995) tried to improve this strategy correlating these measurements with skatole quantification in the fat. However, even if this strategy 126

127 is fast and of easy application on the slaughter-line, it results not completely efficient and reliable. The off-odours from the meat related to boar taint can be evaluated by sensory tests which may involve boiling in water, microwave heating, and melting of fat, but it is still not clear which is the best one to be used. A soldering iron (hot wire) method has also been described (Jarmoluk et al., 1970). While in Germany meat from the head with the inclusion of salivary glands is used for the heating test (Bundesanzeiger, 2007). The use of salivary glands is likely to increase abnormal odours associated with androstenone since the latter is a pheromone, that accumulates in the salivary glands and is released in the saliva of sexually active boars (Booth, 1975). There is no documentation that shows which of these heating methods is more reliable to detect boar taint and also practical to use on the slaughter-line. This study was therefore conducted to determine if these heating tests, on backfat and the composite sample, are effective, to compare them and to evaluate the contributions of androstenone and skatole to any abnormal odours generated. 7.3 Materials and methods A total of 120 entire male pigs of mixed breeding (Large White, Duroc, and Hampshire) with carcass weights of 79.4 ± 7.6kg were obtained in a commercial abattoir in the UK. They came from 8 farms and were slaughtered on the same day. From the same pigs different type of samples were collected: subcutaneous fat from the dorsal neck region (referred to as backfat), samples of fat and muscle from the cheek, and one sub-maxillary salivary glands (total sample weighed approximately 127

128 150g). The backfat sample collected from each pig was an area of skin plus attached fat all the way down to the intersection with the meat of max 15x5 cm, taken near the P2 area. The cheek sample was a square of 10x10 cm, comprehensive of skin, subcutaneous fat and muscle. From each backfat samples skatole and indole concentrations were measured using the simultaneous distillation-extraction procedure followed by GC analysis, according to the methodology described by Annor-Frempong, et al. (1997b) (see Appendix 1). Androstenone concentrations were measured using a modification of the high resolution gas chromatographic procedure of De Brabander and Verbeke (1986) (see Appendix 1). Four groups of 10 samples with extreme values of androstenone and skatole were then selected for subsequent sensory testing. The groups were: high skatole/high androstenone (HS/HA); high skatole/low androstenone (HS/LA); low skatole/high androstenone (LS/HA); and low skatole/low androstenone (LS/LA) using threshold values of 0.2 µg/g of fat for skatole and 1.0 µg/g of fat for androstenone. This approach was similar to that used by Annor-Frempong et al. (1997b). The composite samples of muscle and fat from the cheek and submaxillary glands were prepared from the 40 pigs selected for the sensory testing. These samples weighed about 20g in total and comprised 80% cheek fat, 10% muscle and 10% salivary glands based on the protocol described by Bundesanzeiger (2007), however this author does not specify the ratios of these tissues. The different tissues were finely chopped and blended using a food processor Moulinette S (Moulinex ) and then concentrations of skatole and androstenone were measured using the same methods as for backfat. In addition to androstenone, androstene-α- 128

129 ol and androstene-β-ol were also measured according to the method of De Brabander et al. (1985) since these are known to be present in the salivary glands (Booth, 1975) Heating methods used for sensory assessment of fat samples Preliminary work (not described here) was carried out to achieve the optimal heating times and to standardise the methods. All 4 methods were tested separately and all 40 samples were assessed for each method Microwave Test Twenty grams of clean back fat was chopped into 1 cm length cubes and evenly distributed on the bottom of a 400ml Pyrex glass beaker which was capped with food grade plastic film. Four samples (one from each of the four groups) were heated in a 750W microwave for 90 seconds before being presented on a hotplate at 60ºC to the panellists for assessment within 2 minutes Melting test Twenty grams of clean backfat was chopped into 5mm length cubes and evenly distributed on the bottom of a 400ml Pyrex glass beaker and capped with food grade plastic film before placing on the hotplate of a Lincat griddle setting 4 (surface temperature of 185ºC), until the fat had melted but not browned. Four samples at a time (one from each of the four groups) were presented to the panellists for assessment within 2 minutes. 129

130 Hotwire test Cubes of clean backfat (4cm x 4cm) were presented in glass Petri dishes to the panellists. The panellists applied the tip of a specially adapted soldering iron (held at 180ºC) to the fat before sniffing the volatiles given off. The tip of the soldering iron was cleaned using vegetable fat between each application. Again, the samples were tested in groups of four Boiling test Thirty grams of clean backfat was chopped into 1cm length cubes and placed in a 250ml Erlenmeyer flask and covered with 90ml cold water before capping with aluminium foil and a glass lid. The flasks were placed on the hotplate of a Lincat griddle at setting 4 (surface temperature of 185ºC), until boiling (approximately 5 minutes). When the contents of the flasks had cooled to 75ºC, the samples (one from each of the four groups) were immediately presented to the panellists for assessment of the vapour. The samples were presented again as soon as the contents had cooled to 25ºC (samples were renumbered for the second presentation to the panel) Composite samples Preliminary work showed that when blended in advance of heating, the resulting odour was significantly less intense than when the samples were freshly prepared, so all composite samples were finely chopped just before heating for sensory assessment and only briefly blended with fat and muscle before sensory testing and sampling for chemical analysis of the mixture. 130

131 Only two methods were used for the sensory assessment of composite samples, microwave and boiling, as described for backfat. Both methods were tested separately and all 40 samples were tested on the same day. There was only sufficient material remaining from 25/40 samples for subsequent chemical analysis Sensory assessment The panel consisted of 3 experts, two women and one man (BSEN ISO :2008 Part two: Expert Sensory Assessors) and panels were performed as 4 x 10 replicates for each heating method and sample type (4 methods for backfat and 2 methods for composite samples). Samples were presented in 4 x 4 incomplete Latin squares and labelled with 3-digit codes. Oil of wintergreen and oils with floral notes were used by the panellists to reset noses when necessary. All sensory assessment was carried out over five days. Samples were scored for abnormal odour and pork odour intensity (defined as the amount of abnormal odour in the fat and the amount of cooked pork odour in the fat respectively) using 8 point category scales (1= extremely weak; 2 = very weak; 3 = moderately weak; 4 = slightly weak, 5= slightly strong, 6 = moderately strong, 7 = very strong, 8 = extremely strong). Samples were also scored for skatole odour intensity (odours associated with skatole such as mothballs, musty) and androstenone odour intensity (for odours associated with androstenone such as acrid, ammonia, silage, parsnips, sweaty, dirty) using 9 point category scales (0= none; 1 = extremely weak; 2 = very weak; 3 = moderately weak, 4= slightly weak, 5 = slightly strong, 6 = moderately strong, 7 = very strong, 8 = extremely strong). 131

132 Odours were also classified as: negative, moderately undesirable or extremely undesirable, as described by Bundesanzeiger (2007). The panellists entered their scores directly into a computerised sensory assessment programme (Fizz, version 2.10c, Biosytemes, Couternon, France). Sensory data were analysed using analysis of variance, with heating method and assessor as factors using Minitab Release 14 computer software. Simple linear regression and calculation of pie charts were performed using Microsoft Excel Results and discussion The chemical analysis of the 120 samples of backfat obtained from the commercial abattoir revealed an overall average of 0.17 µg/g for skatole and 0.87 µg/g for androstenone (Table 7-1). Table 7-1. Concentrations of taint compounds in backfat of 120 entire male pigs (µg/g fat). Skatole Androstenone Mean ± Standard Dev ± ± Geometric mean Median Minimum Maximum These average values were higher than those found in pigs analysed in chapter 4, lower than those of the farms tested in the first chapters but similar to the values found in the 30 farms from the preliminary trial in chapter 5. These differences confirm the high variability for the boar taint compounds present in British herds of pigs, which is mostly due to different genetic and rearing conditions. There were 132

133 40/120 (33%) pigs with values for skatole above the threshold of 0.2 µg/g and 31 (26%) pigs with values for androstenone above the threshold of 1.0 µg/g. There were 17 (14%) pigs with both compounds concentrations above the threshold levels. The 4 sample groups of 10 animals per group selected for sensory assessment are described in Table 7-2 and Figure 7-1. Table 7-2. Concentrations of taint compounds present in backfat (mean µg/g fat ± standard deviation). Sample Group Skatole Androstenone (µg/g) (µg/g) High Skatole/High Androstenone ± ± (HS/HA) ( ) ( ) High Skatole/Low Androstenone ± ± (HS/LA) ( ) ( ) Low Skatole/High Androstenone ± ± (LS/HA) ( ) ( ) Low Skatole/Low Androstenone ± ± (LS/LA) ( ) ( ) = Ten pigs in each group; = Range 133

134 Skatole (µg/g fat) HS/HA HS/LA LS/HA LS/LH Androstenone (µg/g fat) Figure 7-1. Classification of samples into 4 groups differing in levels of taint compounds (µg/g fat). The low values for both compounds in the pigs of the LS/LH group were similar to those for Large White x Landrace pigs in chapter 4. While the high values were more close to those of the pigs in the main chicory trial in chapter 5. Concentrations of skatole, androstenone, androstene-α-ol and androstene-β-ol in the composite sample are shown in Table 7-3. Table 7-3. Concentrations of skatole, androstenone, androstene-α-ol and androstene-β-ol in composite samples containing fat, muscle and submaxillary salivary gland in proportions of 8:1:1. Group Skatole Androstene-α-ol Androstene-β-ol Androst. HS/HA (n=4) ( ) ( ) ( ) ( ) HS/LA (n=6) ( ) ( ) ( ) ( ) LS/HA (n=7) ( ) ( ) ( ) ( ) LS/LA (n=8) ( ) ( ) ( ) ( ) = Range 134

135 The concentrations of skatole were lower than in backfat and the concentrations of androstenone were higher. By including 10% muscle and 10% salivary gland it was expected that levels of skatole in the composite blend would be about 20% lower than that found in backfat, as in chapter 3 there were no differences in compounds concentration in the fat from different part of the carcass. Tuomola et al. (1996) found that skatole was present in submaxillary glands of boars, castrates and gilts although at much lower concentrations than found in adipose tissue. Surprisingly, the concentration of skatole in composite samples expressed as a proportion of their value in backfat was found to be as low as 40% in animals with high levels of skatole, suggesting a possible difference in fat tissue levels between the head and back regions, in disagreement with the results in chapter 3 (Figure 7-2) Skatole in composite as % of back fat y = ln(x) R² = Skatole in backfat (µg/g) Figure 7-2. Effect of adding 10% each of muscle and salivary gland tissue on skatole concentration in composite sample expressed as a proportion of the backfat value. In contrast, androstenone was expected to be higher in composite samples since the sub-maxillary salivary gland has been reported to contain high levels of this and the other androstenes compounds, androstene-α-ol and androstene-β-ol 135

136 (Booth, 1975). In this study, animals with low levels of backfat androstenone contained highly varying proportions (1-7 times higher) of androstenone in composite samples whereas those at high levels of backfat androstenone had similar levels in the composite sample as in backfat (Figure 7-3). This suggests a possible risk of some carcasses below the threshold value for androstenone based on backfat measurements being miss-classified as tainted when salivary gland is added to the test sample Androst. 400 in composite as % 300 of backfat y = ln(x) R² = Androstenone in backfat (µg/g) Figure 7-3. Effect of adding 10% each of muscle and salivary gland tissue on androstenone concentration in composite sample expressed as a proportion of the backfat value. Comparisons between HS/HA and LS/LA, the most extreme groups, are in Table 7-4. Each heating method was assessed separately. Overall, lower scores were recorded for the Boiling method than for the other three methods, particularly when assessing the samples at 25ºC. However, the Boiling method at 75ºC distinguished between the extreme samples as well as the other methods. The 136

137 microwave and hotwire methods produced the highest scores for the intensity of odours. In general, pork odour increased as abnormal odour decreased. For all four methods of heating, the intensity of pork odour tended to be higher in the groups with low concentration of taint compounds although the differences were not significant. As expected, the reverse was true for abnormal, skatole and androstenone odour scores which were significantly higher in HS/HA compared with LS/LA groups for all methods except the melting method where androstenone odours were not rated differently between these two extreme samples. Significant differences in abnormal and skatole odours were recorded between the two groups which were high in only one taint compound (HS/LA and LS/HA) when using the microwave, hotwire and boiling (75ºC) methods (Table 7-5). Table 7-4. Scores for backfat odours in HS/HA and LS/LA using different heating methods (values with different superscripts on the same row within method are significantly different, P < 0.05) Odour Score Microwave Hotwire Melting Boiling 75 o C Boiling 25 o C HS/ HA LS/ LA HS/ HA LS/ LA (Scale 1-8) Pork Abnormal 4.63 b 2.57 a 4.23 b 2.30 a 4.03 b 2.70 a 3.70 b 1.63 a 2.43 b 1.53 a (Scale 0-8) Skatole 4.27 b 1.77 a 3.17 b 1.73 a 3.67 b 1.90 a 2.97 b 1.27 a 2.47 b 1.40 a Androstenone 3.57 b 2.57 a 4.27 b 2.17 a b 1.87 a 2.17 b 1.47 a HS/ HA LS/ LA HS/ HA LS/ LA HS/ HA LS/ LA 137

138 Table 7-5. Scores for backfat odours in HS/LA and LS/HA using different heating methods (values with different superscripts on the same row within method are significantly different, P < 0.05). Odour Score Microwave Hotwire Melting Boiling 75 o C Boiling 25 o C HS/ LA LS/ HA HS/ LA LS/ HA (Scale 1-8) Pork Abnormal 4.13 b 3.33 a 4.63 b 3.20 a b 2.97 a (Scale 1-8) Skatole 3.93 b 2.30 a 4.17 b 2.23 a b 1.90 a Androstenone HS/ LA LS/ HA HS/ LA LS/ HA HS/ LA LS/ HA In all cases, HS/LA were given higher odour scores than LS/HA. The same trend was apparent for androstenone odour although this was not significantly different between the groups. These results in this study showed a dominance of skatole in odour perceptions. Panellists were unable to distinguish significantly between HS/HA and HS/LA groups for abnormal odour or androstenone odour using any of the heating methods (the results for the hotwire method are in Table 7-6) although there were differences in skatole odour. This result showed again the dominance of skatole in odour responses. Using the hotwire method, lower scores for androstenone odour were recorded in LS/HA than HS/HA showing that skatole enhanced androstenone odour. The same trends occurred with the other methods. 138

139 Table 7-6. Scores for backfat odours in the 4 taint groups using the hotwire method (means derived from analysis of variance with method and assessor as factors, with 10 replicates). Attribute HS/HA HS/LA LS/HA LS/LA p-value (Scale 1-8) Pork odour Abnormal odour 4.23 a 4.63 a 3.20 b 2.30 c <0.001 (Scale 0-8) Skatole odour 3.17 b 4.17 a 2.23 bc 1.73 c <0.001 Androstenone odour 4.27 a 4.00 ab 3.27 b 2.17 c Correlations between the scores for abnormal odour assessment and the concentrations of individual taint compounds measured in backfat (µg/g fat) for all 40 samples are summarised in Table 7-7. Table 7-7. Correlations between odour scores and boar taint compounds. Regression Microwave Hotwire Melting Boiling Boiling analysis 75ºC 25ºC Skatole odour v *** *** *** *** *** Skatole content Abnormal odour v *** *** ** *** *** Skatole content Androst. odour v *** * Androst. content Abnormal odour V Androst. content R 2 x 100, n= 40, 38 deg of freedom; P<0.05*, P<0.01**, P<0.001*** All relationships between the skatole concentration of backfat and abnormal and skatole odour were highly significant, even for the boiling method when assessed at 139

140 Abnormal odour score 25ºC. In contrast, only the microwave method showed a highly significant relationship between androstenone concentration and odour (P < 0.001). The melting test revealed a less significant relationship (P < 0.05) between androstenone content and odour score. These results again showed the dominance of skatole in odour scores. The linear regression relationships between abnormal odour scores and the skatole content of backfat are shown in Figure 7-4. The slopes of the regression lines are similar for the different methods apart from the melting method. Abnormal odour scores were higher for the microwave and hotwire methods of assessment than for the boiling methods y = x y = x y = x y = x y = x Microwave Hotwire Melting Boiling 75 Boiling Skatole (µg/g fat) Figure 7-4. Relationships between abnormal odour score and skatole concentration in backfat in the different heating methods. Since the hotwire method is unsuitable for use with blended samples and the melting test proved to be less discriminatory in identifying tainted backfat samples, 140

141 pork odour only two methods were compared for composite samples, microwave and the boiling methods (at 75ºC and 25ºC). In contrast to backfat samples, sensory assessment of composite samples after microwave heating revealed a more significant relationship between pork odour and abnormal odour over all samples (Figure 7-5). This was also true for the boiling method when assessed at 75ºC but not at 25ºC. The results for sensory assessment of composite samples showed that abnormal odour as well as skatole and androstenone odour scores were higher than those recorded for backfat samples (Tables 7-8 and 7-9). Although significant differences between the two extreme groups of composite samples (HS/HA and LS/LA) were obtained for both microwave and boiling test (assessed at 75 and 25ºC) in general, the microwave method was once again the most effective y = x R² = y = x R² = abnormal odour Back fat samples Composite samples Figure 7-5. Effect of sample type on the inverse relationship between pork odour and abnormal odour using the microwave test (40 samples). 141

142 Table 7-8. Effect of adding 10% muscle and 10% salivary gland to 80% head fat (Comp) compared with back fat (BF) on odour scores for the microwave test. Odour Score Microwave BF Comp BF Comp BF Comp p- value p-value HS/HA LS/LA (Scale 1-8) Pork < Abnormal < < (Scale 0-8) Skatole < < Androstenone < Table 7-9. Effect of adding 10% muscle and 10% salivary gland to 80% head fat (Comp) compared with backfat (BF) on odour scores for the Boiling test 75 0 C. Odour Score Boiling 75 0 C BF Comp BF Comp BF Comp P value P value HS/HA LS/LA (Scale 1-8) Pork Abnormal < < (Scale 0-8) Skatole < Androstenone < In addition to category scoring, the panellists were asked to record odours as negative, moderately undesirable or highly undesirable as suggested by the German guidelines (Bundesanzeiger, 2007). It must be noted that all three panellists remarked how there were differences in underlying/background odours between heating methods that were difficult to categorise. 142

143 The percentage classification of samples as either negative, moderately or highly undesirable, within each of the four groups, was combined for the three judges. Results of the microwave method for both backfat and composite samples are represented as pie charts (Figure 7-6 and 7-7). Results for the other methods are in Appendix % HS/HA 33.3% HS/LA 30.0% Category highly undesirable mod undesirable negative 40.0% 40.0% 30.0% LS/HA 3.3% LS/LA 30.0% 50.0% 46.7% 70.0% Figure 7-6. Charts representing the percentage classification of samples into degrees of undesirability for backfat using the microwave method of cooking (HA = 1.0 µg, HS = 0.20 µg, LA = 0.99 µg, LS = µg). 143

144 6.7% HS/HA 20.0% HS/LA Category highly undesirable mod undesirable negative 40.0% 43.3% 50.0% 40.0% LS/HA LS/LA 26.7% 16.7% 30.0% 70.0% 56.7% Figure 7-7. Charts representing the percentage classification of samples into degrees of undesirability for composite samples using the microwave method of cooking (HA = 1.0 µg, HS = 0.20 µg, LA = 0.99 µg, LS = µg). In the untainted group LS/LH no samples were classified as highly undesirable and only 30% were recorded as moderately undesirable for both backfat and composite samples. For the three remaining groups, the percentage of samples classified as highly undesirable increased markedly and was higher for composite samples than backfat (33.3 and 50% for HS/HA, 30 and 40% for HS/LA and 3.3 and 16.7% for LS/HA in backfat and composite samples respectively). A similar result for the percentage of highly undesirable samples was also found when the boiling method (assessed at 75ºC) was used (16.7 and 46.7% for HS/HA, 16.7 and 23.3% for HS/LA and 6.7 and 20% for LS/HA groups in backfat and composite samples respectively). However, 10% of samples in the LS/LA group of 144

145 composite samples were also classified as highly undesirable compared with 0% for the backfat samples in the same group. The microwave method appeared to discriminate between the groups of samples more effectively than the boiling method. In backfat, HS/LA and HS/HA samples received similar scores for highly undesirable, showing the dominance of skatole. In composite samples, however, the proportion of highly undesirable scores was higher for HS/HA than for HS/LA, showing the importance of androstenone in this tissue category. Pooling the scores for the 3 assessors has disguised some differences between them. Results in Table 7-10 for backfat show that assessor 1 ranked few HS/HA samples as highly undesirable and ranked most LS/LA samples as negative. Assessor 2 gave more highly undesirable ratings to HS/HA and fewer negative ratings to LS/LA. Assessor 2 is therefore more sensitive to boar taint than assessor 1. This variability between people in their sensitivity to boar taint odours has been observed in previous research (Annor-Frempong et al, 1997b). Table Scores for highly undesirable in HS/HA and negative in LS/LA for the 3 assessors when evaluating backfat. Assessor Heating method Microwave Hot wire Boiling 75 Boiling 25 Melting % of highly undesirable scores for HS/HA % of negative scores for LS/LA

146 7.5 Conclusions The results of this study show that the 4 cooking methods that can be quickly applied to backfat in the abattoir (boiling in water, microwave, melting and hotwire) can clearly distinguish between samples with extreme ratios of skatole and androstenone, i.e. HS/HA and LS/LA. Separation was achieved for abnormal odour, skatole odour and androstenone odour when these were assessed by 3 experienced assessors. In general, the microwave and hot wire methods produced higher scores for abnormal, skatole and androstenone odours while the boiling method resulted in lower scores although it did allow discrimination between the two extreme groups, especially at 75ºC. The melting method could not distinguish a difference in androstenone odour between these extreme samples. The intermediate groups of backfat samples (HS/LA and LS/HA) were also separated by the 4 methods in terms of abnormal odour and skatole odour. However, androstenone odour was not scored differently between these samples in any method. Again, the microwave and hotwire methods produced higher scores. The panellists remarked that each cooking method produced a different background odour. Correlations between odour scores and concentrations of skatole and androstenone for all methods were higher for skatole showing that it was more closely associated with boar taint. The correlations were higher for the microwave and hotwire methods than for the other methods. A composite sample containing 80% fat, 10% muscle and 10% sub-maxillary salivary glands from the head was formed from the same 4 groups of 10 pigs that constituted the backfat groups. The boiling and microwave methods easily 146

147 separated the 4 groups. Including salivary glands greatly increased the concentration of androstenone compared with backfat and produced different concentrations of androstene-ol compounds in proportion to androstenone (Table 1-3). Scores for skatole odour, androstenone odour and abnormal odour were higher in the composite sample than in backfat although skatole concentration was lower. The three experienced assessors also classified the backfat and composite samples into negative, moderately undesirable and highly undesirable categories. For backfat, the 4 methods showed similar trends, with the proportion of highly undesirable samples increasing in the order LS/LA < LS/HA < HS/LA < HS/HA. However, the proportion of highly undesirable samples was similar between HS/LA and HS/HA showing the dominance of skatole in the perception of abnormal odour in backfat. In the composite samples, only the boiling and microwave methods were assessed. There was the same trend in the proportion of samples classified as highly undesirable between the groups although in this case there was a clear distinction between HS/LA and HS/HA with more of the latter classified as highly undesirable. This result suggests that the inclusion of only a small quantity of salivary gland in the composite sample has resulted in a significant enough increase in androstenone content (Table 7-3) to produce a different sensory response to that of backfat samples between these two groups. The differences in sensory scores between the 4 backfat groups show that high skatole concentrations enhance the perception of androstenone odour (higher androstenone odour in HS/HA than LS/HA for all methods). This is evidence of the 147

148 synergism between the boar taint compounds observed by other workers (e.g. Annor-Frempong et al, 1997). Even if some of the heating methods were found to be efficient for the detection of boar taint the three assessors differed in their responses to the different groups of samples. Assessors 2 and 3 gave higher scores to the HS/HA samples than assessor 1 showing they were more sensitive to boar taint. This variability between assessors is a feature of boar taint evaluation that can cause results to differ within and between abattoirs, making these methods not universally usable as standard tests for boar taint. 148

149 8 General conclusions During the progress of the present work, it was clear that the British pig population is characterized by a wide variability in the levels of the main compounds responsible for boar taint, androstenone and skatole. Two different surveys on the taint levels have been conducted on farms serving commercial abattoirs in the UK and a total of 93 farms have been tested, using a combined sample obtained from all the pigs of one farm. In these farms the levels for androstenone varied from a minimum of µg/g to a maximum of µg/g of fat, and 60% of the farms had levels over the threshold of 1.0 µg/g of fat. The levels of skatole varied from a minimum of 0.37 µg/g to a maximum of µg/g of fat, and 20% of the farms had levels over the threshold of 0.2 µg/g of fat. This variability was also confirmed when the pigs have been tested individually and can be explained by the different breeds involved and by the differences of their rearing condition. There was no significant difference in the average value of each boar taint compound in the fat from the three different sites of the carcass (backfat, neck and cheek). The similarity in the measurements was also confirmed by the significant (P < 0.01) positive correlation found between the different pairs of values (backfatneck, backfat-cheek and neck-cheek) for all the compounds. As only 20 pigs were tested definite conclusions cannot be drawn, but these results suggest that other sites of the carcass of less economical value can be used for the measurement of the boar taint compounds instead of the most common site known as backfat. British pigs producers should aim for growth rates above 0.7kg/d from weaning to slaughter for reasons of both lower production costs and better eating quality. Pigs 149

150 growing fast between weaning and slaughter, at around 0.68 kg/d, had significantly more tender pork than those growing at 0.54 kg/d. The difference was more marked in pigs slaughtered at the heavier weight of 110 kg but was also present at 90kg. The taste panel results also showed that juiciness of pork was increased by faster growth and as a result the overall liking score for eating quality was higher. However, there was no evidence that slower growth and high carcass weight increases boar taint. In fact the concentrations of androstenone and skatole were highest in the fastest growing pigs. The use of entire males should not be limited as there were no significant differences in both quantitative and qualitative aspects of meat quality when the meat from boars and females were compared. A supplementation of the diet with 9% chicory for 2 weeks before slaughter had a clear effect in reducing skatole to extremely low levels. After the 2 weeks on the diet the skatole concentration was on average 0.065µg/g and 55% of pigs had levels between 0 and 0.05µg/g, typical of levels in castrated males. However no improvement in odour scores was detected by the panellists, probably because the levels in the fat of androstenone in these animals increased after the 2 weeks of the trial. It was surprising that the sensory perception of boar taint was not reduced because several studies and the results from the last chapter show that skatole is the taint compound most closely associated with boar taint and it has also been demonstrated that skatole enhances the sensory perception of androstenone. These results suggest that the use of chicory to control boar taint might not be cost effective in pig populations where levels of androstenone are high. Where these are more normal, the use of chicory might be cost effective, depending on the extra price received for taint-free pigs. 150

151 Between 4 different cooking methods which could possibly be used in abattoirs on backfat (boiling in water, microwave, melting and hotwire) for boar taint detection, microwave and hot wire were the best methods to detect successfully abnormal, skatole and androstenone odours. All methods showed that correlations between odour scores and concentrations of skatole and androstenone were higher for skatole showing that it was more closely associated with boar taint. Skatole appears to give a biggest contribution to the detection of boar and this was also demonstrated by the differences in sensory scores between the 4 backfat groups (LS/LA, LS/HA, HS/LA and HS/HA) that showed how high skatole concentrations enhance the perception of androstenone odour (higher androstenone odour in HS/HA than LS/HA for all methods). Even if some of the heating methods resulted efficient for the detection of boar taint the three assessors differed in their responses to the different groups of samples and this variability between assessors is a feature of boar taint evaluation that can cause results to differ within and between abattoirs, making these methods not universally usable as standard tests for boar taint. 151

152 9 Appendix Determination of skatole and indole concentration in fat Skatole and indole concentration in fat was determined following the methodology of Annor-Frempong, et al. (1997b). After removing the skin and the subcutaneous glands, approximately 10 g of cleaned adipose tissue were finely chopped and blended in 80 ml distilled water, brought to ph 10.5 with sodium hydroxide, using an MSE overhead blender and glass vortex flask. The homogenate was poured into a 150 ml bound bottom flask containing few anti-bumping granules and an antifoaming tablet. Once three samples in duplicate had been blended (6 flask in all), they were placed on a 6-place heating mantle. Internal standard (20 µl 5-methylindole in hexane) was added to all samples before gently refluxing for 2 hours using Likens-Nikerson condensers to trap volatiles in 25 ml of pentane/ether (9:1, v/v), held in 50 ml round bottomed flasks in a water bath at 43ºC. After cooling, the solvent was decanted into glass vials and reduce to approximately 200 µl under nitrogen gas before transfer to amber gas chromatography (GC) vials. Samples were analysed using a Chrompak CP57 Wax CB capillary column 25 m x 0.32 mm, in a Fisons 8000 Series GC equipped with FIN-NDP detector. Quantification was achieved by use of the internal standard added prior to extraction of the sample, and any discrimination by the detector was corrected for by use of a calibration cocktail containing known amounts of all 3 components. 152

153 9.2 Determination of androstenone in fat Androstenone concentration in the fat was determined using a modification of the procedure of De Brabander and Verbeke (1986). After removing the skin and the subcutaneous glands, approximately 0.4 g (± g) of clean sliced fat were putted into a 15 ml Pyrex tube with a polytetrafluoroethylene cup and 2ml of 10% KOH in methanol were added. Once 12 samples in duplicate were prepared, internal standard (α-androstenol in hexane) was added to all the samples and washed down into the tube with 1.4 ml of toluene and capped. The tubes were than manually shaken every ten minutes while they were kept for 1 hour in a water bath at 60º. After cooling, 2ml of water, 2.5 ml of methanol and 3ml of petroleum ether/diethyl ether (1:1, v/v) were added. The tubes were then manually shaken for a minute and the phases of the liquid were separated by centrifugation at 5000 rpm for 2 minutes. The clear upper phase was transferred into a clean tube. This extraction was then repeated twice adding only 3 ml of petroleum ether/diethyl ether (1:1, v/v). The obtained liquid was reduced under nitrogen gas while kept at 60ºC in a water bath. When a cloudy residual was obtained 2 ml of petroleum ether/diethyl ether (1:1, v/v) were added to the tube. These tubes were than centrifuge at 5000 rpm for a minute to separate the added liquid from a white precipitate present in the tubes. The liquid was passed into a clean tube and reduced under nitrogen gas to approximately 1 ml. Under a fume cupboard 4-8 drops of diazomethane in diethylether were added to each tube and left to evaporate for 10 minutes at 60ºC in a water bath. The residual liquid was then evaporated with nitrogen gas and immediately 200 µl of hexane were added to the tube. The obtained sample was 153

154 then transferred into a GC vial. The samples have been analysed using a Fisons 8000 series GC in hot split mode (20:1), equipped with a Chrompak Sil8 WCOT capillary column (25m x 0.25mm), helium as the carrier gas and a flame ionisation detector. Quantification was achieved by the use of response factor of androstenone relative to the internal standard. 154

155 10 Appendix 2. Tables and figures 10.1 Androstenone and skatole levels in pigs on British farms Table Concentrations of taint compounds in the 30 farms from A1. Farm Indole Skatole Androstenone µg/g fat µg/g fat µg/g fat

156 Table Concentrations of taint compounds in the 33 farms from A2. Farm Indole Skatole Androstenone µg/g fat µg/g fat µg/g fat

157 10.2 Androstenone and skatole levels in different fat tissues Table Skatole concentrations (µg/g fat) in the 20 pigs. Site Pig Backfat Neck Cheek

158 Table Androstenone concentrations (µg/g fat) in the 20 pigs. Site Pig Backfat Neck Cheek

159 Figure Linear regression (red line) of skatole (μg/g) in backfat and cheek along with the line of equality (black line). Figure Linear regression (red line) of skatole (μg/g) in neck and cheek along with the line of equality (black line). 159

160 Figure Linear regression (blue line) of androstenone (μg/g) in backfat and cheek along with the line of equality (black line). Figure Linear regression (blue line) of androstenone (μg/g) in neck and cheek along with the line of equality (black line). 160

161 Figure The plots of differences between skatole levels from backfat and cheek. Figure The plots of differences between skatole levels from neck and cheek. 161

162 Figure The plots of differences between androstenone levels from backfat and cheek. Figure The plots of differences between androstenone levels from neck and cheek. 162