Universidad Politécnica De Madrid Effect of Short-Term Fasting Prior to Slaughter on Welfare and Meat Quality of Rainbow Trout

ESCUELA TÉCNICA SUPERIOR DE INGENIEROS AGRÓNOMOS

UNIVERSIDAD POLITÉCNICA DE MADRID

EFFECT OF SHORT-TERM FASTING PRIOR TO SLAUGHTER ON WELFARE AND MEAT QUALITY OF RAINBOW TROUT (Oncorhynchus mykiss)

TESIS DOCTORAL

JAVIER LÓPEZ LUNA Ingeniero Agrónomo

2013

ESCUELA TÉCNICA SUPERIOR DE INGENIEROS AGRÓNOMOS

UNIVERSIDAD POLITÉCNICA DE MADRID

EFFECT OF SHORT-TERM FASTING PRIOR TO SLAUGHTER ON WELFARE AND MEAT QUALITY OF RAINBOW TROUT (Oncorhynchus mykiss)

JAVIER LÓPEZ LUNA Ingeniero Agrónomo

Director de la Tesis

Morris Villarroel Robinson Dr. en Biología

ACKNOWLEDGEMENTS

First, I would like to express my sincere gratitude to my supervisor, Dr. Morris Villarroel Robinson for his support and guidance during the last four years, for generously sharing his knowledge and time and for being patient with my questions.

I am really grateful to all those who helped me during my PhD study. Many people have made invaluable contributions, both directly and indirectly to my research. These include Francisco Berdejo and the colleagues who participated in the trials of this Thesis. Teachers and members of the Departamento Producción Animal (ETSI Agrónomos, UPM), Fernando Torrent and members of the Piscifactoría (ETSI Montes, UPM), Miguel Ángel Ibáñez (Departamento de Estadística y Métodos de Gestión en la Agricultura (ETSI Agrónomos, UPM), Jesús de la Fuente and members of the Departamento de Producción Animal (Facultad de Veterinaria, UCM) and Pilar Barreiro and Adolfo Moya (Departamento de Ingeniería Rural, ETSI Agrónomos, UPM).

This Thesis was made possible with a grant from the Ministerio de Ciencia e Innovación, project AGL2010-19479.

III

AGRADECIMIENTOS

En primer lugar quisiera agradecer a mi tutor Morris Villarroel Robinson todo el apoyo y la confianza prestados durante el periodo de doctorado en la UPM, por su ayuda personal y técnica. Gracias por ser tan buen consejero.

Gracias también a Francisco Berdejo y a todos los compañeros que tomaron parte en la realización de las pruebas de esta Tesis, a los profesores y personal del Departamento de Producción Animal de la ETSI Agrónomos, a Fernando Torrent y personal de la Piscifactoría de la ETSI Montes, a Miguel Ángel Ibáñez del Departamento de Estadística y Métodos de Gestión en Agricultura de la ETSI Agrónomos, a Jesús de la Fuente y personal del Departamento de Producción Animal de la Facultad de Veterinaria (UCM) y a Pilar Barreiro y Adolfo Moya del Departamento de Ingeniería Rural de la ETSI Agrónomos.

Las pruebas experimentales que componen esta Tesis Doctoral han sido realizadas gracias al proyecto AGL2010-19479 del Ministerio de Ciencia e Innovación.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... III AGRADECIMIENTOS ...... V LIST OF FIGURES ...... XI LIST OF TABLES ...... XVII LIST OF ABBREVIATIONS ...... XIX SUMMARY ...... XXI RESUMEN ...... XXIII INTRODUCTION 1. The importance of aquaculture ...... 3 1.1. Aquaculture in Spain ...... 4 1.2. Trout ...... 5 2. Stress and welfare ...... 6 2.1. Animal welfare ...... 6 2.2. Fish welfare ...... 8 3. Pre-slaughter handling ...... 10 3.1. Fasting ...... 10 3.2. Hour of slaughter ...... 13 3.3. Water temperature ...... 14 4. How to measure welfare ...... 15 4.1. Biometric indicators ...... 15 4.2. Stress indicators...... 16 4.3. Haematological indicators ...... 17 4.3.1. Plasma cortisol ...... 19 4.3.2. Plasma glucose ...... 20 4.3.3. Plasma lactate ...... 21 4.3.4. Haematocrit ...... 21 4.3.5. Leucocytes ...... 22 4.4. Metabolic indicators ...... 22 4.4.1. Liver and muscle glycogen ...... 22 4.4.2. R-value (IMP/ATP ratio) ...... 23 5. Product quality ...... 24 5.1. Muscle pH ...... 26 5.2. Rigor mortis ...... 27

VII

5.3. Water holding capacity ...... 30 5.4. Meat colour ...... 32 6. Objectives ...... 34 MATERIAL AND METHODS 1. Installations and fish material ...... 37 2. Experimental design ...... 38 3. Slaughtering and sampling procedures...... 39 3.1. Biometric indicators ...... 40 3.2. Haematological indicators ...... 40 3.2.1. Plasma cortisol ...... 40 3.2.2. Plasma glucose ...... 41 3.2.3. Plasma lactate ...... 41 3.2.4. Red and white cells ...... 42 3.3. Metabolic indicators ...... 43 3.3.1. Liver and muscle glycogen ...... 43 3.3.2. R-value (IMP/ATP ratio) ...... 43 3.4. Product quality ...... 44 3.4.1. Muscle pH ...... 44 3.4.2. Rigor mortis ...... 44 3.4.3. Water holding capacity ...... 45 3.4.4. Meat colour ...... 46 4. Statistics ...... 46 4.1. Biometric and haematological indicators ...... 46 4.2. Metabolic indicators ...... 48 4.3. Product quality ...... 48 RESULTS 1. TRIAL 1 ...... 53 1.1. Biometric indicators ...... 53 1.2. Haematological indicators ...... 53 1.3. Product quality ...... 55 1.3.1. Muscle pH ...... 55 1.3.2. Rigor mortis ...... 57 1.3.3. Percentage of water released ...... 58 2. TRIAL 2 ...... 60 2.1. Biometric indicators ...... 60 2.2. Haematological indicators ...... 61

VIII

2.3. Product quality ...... 64 2.3.1. Muscle pH ...... 64 2.3.2. Rigor mortis ...... 66 2.3.3. Percentage of water released ...... 68 2.3.4. Meat colour ...... 69 3. TRIAL 3 ...... 70 3.1. Biometric indicators ...... 70 3.2. Haematological indicators ...... 72 3.3. Metabolic indicators ...... 73 3.4. Product quality ...... 74 3.4.1. Muscle pH ...... 74 3.4.2. Rigor mortis ...... 75 3.4.3. Percentage of water released ...... 77 3.4.4. Meat colour ...... 79 4. TRIAL 4 ...... 79 4.1. Biometric indicators ...... 79 4.2. Haematological indicators ...... 82 4.3. Product quality ...... 84 4.3.1. Muscle pH ...... 84 4.3.2. Rigor mortis ...... 85 4.3.3. Percentage of water released ...... 87 5. Overall analysis of haematological indicators including water temperature ...... 88 DISCUSSION 1. Biometric indicators ...... 95 2. Haematological indicators ...... 97 2.1. Plasma cortisol ...... 97 2.2. Plasma glucose ...... 98 2.3. Plasma lactate ...... 100 2.4. Haematocrit ...... 101 2.5. Leucocytes ...... 102 3. Metabolic indicators ...... 103 3.1. Liver and muscle glycogen ...... 103 3.2. R-value ...... 105 4. Product quality ...... 105 4.1. Muscle pH ...... 105 4.2. Rigor mortis ...... 106

4.3. Water holding capacity-percentage of water released ...... 108 4.4. Meat colour ...... 109 5. Overall analysis of haematological indicators including water temperature ...... 110 CONCLUSIONS 1. Biometric indicators ...... 117 2. Haematological indicators ...... 117 3. Metabolic indicators ...... 117 4. Product quality ...... 118 REFERENCES ...... 121

LIST OF FIGURES

Figure 1. World capture fisheries and aquaculture production (FAO, 2012a) ...... 3

Figure 2. Reported aquaculture production in Spain from 1950 (FAO, 2012a) ...... 4

Figure 3. Global aquaculture production of Onchorhynchus mykiss (FAO, 2012b) ...... 6

Figure 4. Overview of the HPI axis in fish (adapted from Alsop & Vijayan, 2009) ...... 17

Figure 5. Overview of dynamics of cortisol and catecholamines in the production of glucose

(adapted from Martínez-Porchas et al., 2009) ...... 18

Figure 10. Post mortem ATP degradation in fish muscle. (from Gill, 1992) ...... 23

Figure 6. Aerobic and anaerobic breakdown of glycogen in fish muscle (Huss, 1995) ...... 25

Figure 7. Tissue stress and quality indicators in fish (Poli et al., 2005) ...... 26

Figure 8. Rigor mortis as a function of meat chemistry and time (www.qpc.adm.slu.se) ...... 29

Figure 9. CIE L*a*b* colour space (from www.linocolor.com)...... 32

Figure 11. Photo of the raceway holding treatment fish (fasted) with three sections. Water

entered through the main channel at the top of the photo and exited through a drain at the

bottom. Note that the sections were covered with a net (not present in photo) to avoid fish

jumping from one section to another...... 38

Figure 12. Setup of the raceways and sections at the School of Forestry Engineering. Each day

(1, 2 and 3) represents a batch of 10 fish ...... 39

Figure 13. Muscle pH sampling...... 44

Figure 14. Rigor mortis sampling. The length of the horizontal (X) and vertical (Y) legs of the

right-angled triangle and the rigor angle (α) are represented...... 45

Figure 15. Changes in plasma glucose concentrations of rainbow trout during three days of

fasting in TRIAL 1. Data are means ± SEM (n=30). Different letters indicate significant

differences among groups (P<0.001)...... 54

Figure 16. Changes in plasma lactate concentrations of rainbow trout during three days of

fasting in TRIAL 1. Data are means ± SEM (P=0.004; n=10). Different letters indicate

significant differences between fasted and fed fish at sampling points (P<0.05)...... 54

Figure 17. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00

n=15 in controls; SEM=0.07). Significant differences among days of fasting at the same

sampling points are denoted with different letters (P<0.05)...... 56

Figure 18. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and

20h00 (c) during 72 h of storage in TRIAL 1. Data represent means (n=5 in fasted groups

and n=15 in controls; SEM=4.4 in 08h00 and 3.86 in 14h00 and 20h00). Significant

differences among days of fasting at the same sampling points are denoted with different

letters (P<0.05)...... 58

Figure 19. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a),

14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 1. Data represent means (n=3 in

fasted groups and n=9 in controls; SEM=2.55). Significant differences among days of

fasting at the same sampling points are denoted with different letters (P<0.05)...... 59

Figure 20. Changes in the coefficient of condition (CC) of rainbow trout during three days of

fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters indicate significant

differences among groups (P<0.001)...... 60

Figure 21. Changes in the relative weight of the whole gut (RWWG) of rainbow trout during

three days of fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters

indicate significant differences among groups (P<0.001)...... 61

Figure 22. Changes in plasma glucose concentrations of rainbow trout during three days of

fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters indicate significant

differences among groups (P=0.010)...... 62

Figure 23. Changes in haematocrit levels of rainbow trout during three days of fasting in

TRIAL 2. Data are means ± SEM (n=20). Different letters indicate significant differences

among groups (P<0.001)...... 63

XII

Figure 24. Changes in leucocytes levels of rainbow trout during three days of fasting in TRIAL

2. Data are means ± SEM (n=30). Different letters indicate significant differences among

groups (P=0.023)...... 63

Figure 25. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00

n=15 in controls; SEM=0.09). Significant differences among days of fasting at the same

sampling points are denoted with different letters (P<0.05)...... 65

Figure 26. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and

20h00 (c) during 72 h of storage in TRIAL 2. Data represent means (n=5 in fasted groups

and n=15 in controls; SEM=4.0). Significant differences among days of fasting at the

same sampling points are denoted with different letters (P<0.05)...... 67

Figure 27. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a),

14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 2. Data represent means (n=3 in

fasted groups and n=9 in controls; SEM=2.04). Significant differences among days of

fasting at the same sampling points are denoted with different letters (P<0.05)...... 68

Figure 28. Colour evolution (L* and a* coordinates, CIELab) of rainbow trout flesh during 72 h

of storage in TRIAL 2. Data represent means±SEM (n=90). Significant differences

among sampling points within a coordinate are denoted with different letters (PL* <0.001

and Pa* =0.030)...... 69

Figure 29. Changes in the coefficient of condition (CC) of rainbow trout during three days of

fasting in TRIAL 3. Data are means ± SEM (n=30). Different letters indicate significant

differences among groups (P=0.009)...... 70

Figure 30. Changes in the relative weight of the gut content (RWGC) of rainbow trout during

three days of fasting in TRIAL 3. Data are means ± SEM (n=30). Different letters

indicate significant differences among groups (P<0.001)...... 71

Figure 31. Changes in the leucocytes levels of rainbow trout during three days of fasting in

TRIAL 3. Data are means ± SEM (n=20). Different letters indicate significant differences

among groups (P=0.003)...... 73

XIII

Figure 32. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00

n=15 in controls; SEM=0.07). Significant differences among days of fasting at the same

sampling points are denoted with different letters (P<0.05)...... 75

Figure 33. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and

20h00 (c) during 72 h of storage in TRIAL 3. Data represent means (n=5 in fasted groups

and n=15 in controls; SEM=3.5). Significant differences among days of fasting at the

same sampling points are denoted with different letters (P<0.05)...... 76

Figure 34. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a),

14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 3. Data represent means (n=3 in

fasted groups and n=9 in controls; SEM=1.27). Significant differences among days of

fasting at the same sampling points are denoted with different letters (P<0.05)...... 78

Figure 35. Colour evolution (L*a*b* coordinates, CIELab) of rainbow trout flesh during 48 h

of storage in TRIAL 3. Data represent means±SEM (n=90). Significant differences

among sampling points within a coordinate are denoted with different letters (PL* and Pb*

<0.001; Pa*=0.0013)...... 79

Figure 36. Changes in the body weight of rainbow trout during three days of fasting in TRIAL

4. Data are means ± SEM (n=30). Different letters indicate significant differences among

groups (P<0.001)...... 80

Figure 37. Changes in the coefficient of condition (CC) of rainbow trout during three days of

fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant

differences among groups (P=0.013)...... 80

Figure 38. Changes in the relative weight of the gut content (RWGC) of rainbow trout during

three days of fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters

indicate significant differences among groups (P<0.001)...... 81

Figure 39. Changes in the hepatosomatic index (HSI) of rainbow trout during three days of

fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant

differences among groups (P<0.001)...... 81

XIV

Figure 40. Changes in plasma glucose levels of rainbow trout during three days of fasting in

TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant differences

among groups (P=0.003)...... 82

Figure 41. Changes in plasma lactate levels of rainbow trout during three days of fasting in

TRIAL 4. Data are means ± SEM (P=0.047; n=10). Different letters indicate significant

differences between fasted and fed fish at sampling points (P<0.05)...... 83

Figure 42. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00

n=15 in controls; SEM=0.09). Significant differences among days of fasting at the same

sampling points are denoted with different letters (P<0.05)...... 85

Figure 43. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and

20h00 (c) during 48 h of storage in TRIAL 4. Data represent means (n=5 in fasted groups

and n=15 in controls; SEM=3.3). Significant differences among days of fasting at the

same sampling points are denoted with different letters (P<0.05)...... 86

Figure 44. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a),

14h00 (b) and 20h00 (c) during 48 h of storage in TRIAL 4. Data represent means (n=3 in

fasted groups and n=9 in controls; SEM=1.61). Significant differences among days of

fasting at the same sampling points are denoted with different letters (P<0.05)...... 88

Figure 45. Overall changes in plasma glucose levels of rainbow trout during three days of

fasting according to water temperature. Data are means ± SEM (n=60). Different letters

indicate significant differences among groups (P=0.006)...... 89

Figure 46. Overall changes in plasma lactate levels of rainbow trout during three days of fasting

according to water temperature. Data are means ± SEM (n=120). Different letters indicate

significant differences among groups (P<0.001)...... 90

Figure 47. Overall changes in haematocrit levels of rainbow trout at three different hours of

slaughter according to water temperature. Data are means ± SEM (n=120). Different

letters indicate significant differences among groups (P<0.001)...... 90

LIST OF TABLES

Table 1. Average initial parameters and analyses in the four trials ...... 38

Table 2. Summary of measurements and analyses carried out in the four trials ...... 46

Table 3. Summary of P-values of biometric indicators in TRIAL 1 for the statistical model used

( )...... 53

Table 4. Summary of P-values of haematological variables in TRIAL 1 for the statistical model

used ( )...... 55

Table 5. Summary of P-values of biometric indicators in TRIAL 2 (three days of fasting) for the

statistical model used (

)...... 61

Table 6. Summary of P-values of haematological variables in TRIAL 2 (three days of fasting)

for the statistical model used (

)...... 64

Table 7. Mean values of the b* coordinate of rainbow trout flesh during 72 h of storage in

TRIAL 2 for the statistical model used (

) ...... 70

Table 8. Summary of P-values of biometric indicators in TRIAL 3 for the statistical model used

( )...... 72

Table 9. Summary of P-values of haematological variables in TRIAL 3 for the statistical model

used ( )...... 73

Table 10. Changes in liver and muscle glycogen and R-values (IMP/ATP ratio) of rainbow trout

during three days of fasting in TRIAL 3. P refers to the factor type of fasting (day of

slaughter)...... 73

Table 11. Summary of P-values of biometric indicators in TRIAL 4 for the statistical model

used ( )...... 82

XVII

Table 12. Summary of P-values of haematological variables in TRIAL 4 for the statistical

model used ( )...... 83

Table 13. Summary of P-values of haematological variables in the overall analysis including

water temperature for the statistical model used (

)...... 91

XVIII

LIST OF ABBREVIATIONS

ANOVA: analysis of variance CC: coefficient of condition CIWF: Compassion in World Farming cm: centimetres CP: creatine phosphate CPA: cold perchloric acid DFD: dark, firm and dry meat EDTA: ethylenediaminetetraacetic acid EFSA: European Food Safety Authority FAO: Food and Agriculture Organization of the United Nations FAWC: Farm Animal Welfare Council Fig.: figure FK: potassium fluoride FSBI: Fisheries Society of the British Isles GOD: glucose oxidase h: hours HSA: Humane Slaughter Association HSI: hepatosomatic index l: litres µl: microlitres m: meters min: minutes ml: millilitres OIE: World Organisation for Animal Health POD: peroxidase PSE: pale, soft and exudative meat PWR: percentage of water released RWGC: relative whole gut content RWWG: relative weight of the whole gut SEM: standard error of the mean SR: sarcoplasmic reticulum t: tonne TMB: tetramethylbenzidine WHC: water holding capacity

XIX

SUMMARY

Pre-slaughter handling in fish is important because it affects both physiological reactions and post mortem biochemical processes, and thus welfare and product quality. Pre-slaughter fasting is regularly carried out in aquaculture, as it empties the viscera of food and faeces, thus reducing the intestinal bacteria load and the spread of gut enzymes and potential pathogens to the flesh. However, it is unclear how long rainbow trout can be fasted before suffering unnecessary stress. In addition, very little is known about the best time of the day to slaughter fish, which may in turn be dictated by diurnal rhythms in physiological stress parameters. Water temperature is also known to play a very important role in stress physiology in fish but the combined effect with fasting is unclear. Current recommendations regarding the optimal duration of pre-slaughter fasting do not normally consider water temperature and are only based on days, not degree days (ºC d).

The effects of short-term fasting prior to slaughter (1, 2 and 3 days, between 11.1 and 68.0 ºC days) and hour of slaughter (08h00, 14h00 and 20h00) were determined in commercial-sized rainbow trout (Oncorhynchus mykiss) over four trials at different water temperatures (TRIAL 1, 11.8 ºC; TRIAL 2, 19.2 ºC; TRIAL 3, 11.1 ºC; and TRIAL 4, 22.7 ºC). We measured biometric, haematological, metabolic and product quality indicators. In each trial, the values of fasted fish (n=90) were compared with 90 control fish kept under similar conditions but not fasted.

Results show that fasting affected biometric indicators. The coefficient of condition in fasted trout was lower than controls 2 days after food deprivation. Gut emptying occurred within the first 24 h after the cessation of feeding, with small traces of digesta after 48 h. Gut emptying was faster at higher water temperatures. Liver weight decreased in food deprived fish and differences between fasted and fed trout were more evident when gut clearance was faster.

The overall effect of fasting for up to three days on haematological indicators was small. Plasma cortisol levels were high in both fasted and fed fish in all trials. Plasma glucose response to fasting varied among trials, but it tended to be lower in fasted fish as the days of fasting increased. In any case, it seems that water temperature played a more important role, with higher concentrations at lower temperatures on days 2 and 3 after the cessation of feeding. Plasma lactate levels indicate moments of high muscular activity and were also high, but no variation related to fasting could be found. Haematocrit did not show any significant effect of fasting, but leucocytes tended to be higher when trout were less stressed and when their body condition was higher. Finally, the loss of liver weight was not accompanied by a decrease in

XXI liver glycogen (only measured in TRIAL 3), suggesting that a different strategy to maintain plasma glucose levels was used.

Regarding the hour of slaughter, lower cortisol levels were found at 20h00, suggesting that trout were less stressed later in the day and that pre-slaughter handling may be less stressful at night. Haematocrit levels were also lower at 20h00 but only at lower temperatures, indicating that higher temperatures increase metabolism.

Neither fasting nor the hour of slaughter had a significant effect on the evolution of meat quality during 3 days of storage. In contrast, storage time seemed to have a more important effect on meat quality parameters. The lowest pH was reached 24-48 h post mortem, with a higher variability among fasting durations at 20h00, although no clear pattern could be discerned. Maximum stiffening from rigor mortis occurred after 24 h. The water holding capacity was very stable throughout storage and seemed to be independent of pH changes. Meat lightness (L*) slightly increased during storage and a* and b*-values were relatively stable.

In conclusion, based on the haematological results, slaughtering at night may have less of a negative effect on welfare than at other times of the day. Overall, our results suggest that rainbow trout can cope well with fasting up to three days or 68 ºC d prior to slaughter and that their welfare is therefore not seriously compromised. At low water temperatures, trout could probably be fasted for longer periods without negative effects on welfare but more research is needed to determine the relationship between water temperature and days of fasting in terms of loss of live weight and the decrease in plasma glucose and other metabolic indicators.

XXII

RESUMEN

El manejo pre-sacrificio es de vital importancia en acuicultura, ya que afecta tanto a las reacciones fisiológicas como a los procesos bioquímicos post mortem, y por tanto al bienestar y a la calidad del producto. El ayuno pre-sacrificio se lleva a cabo de forma habitual en acuicultura, ya que permite el vaciado del aparato digestivo de restos de alimento y heces, reduciendo de esta manera la carga bacteriana en el intestino y la dispersión de enzimas digestivos y potenciales patógenos a la carne. Sin embargo, la duración óptima de este ayuno sin que el pez sufra un estrés innecesario no está clara. Además, se sabe muy poco sobre la mejor hora del día para realizar el sacrificio, lo que a su vez está regido por los ritmos diarios de los parámetros fisiológicos de estrés. Finalmente, se sabe que la temperatura del agua juega un papel muy importante en la fisiología del estrés pero no se ha determinado su efecto en combinación con el ayuno. Además, las actuales recomendaciones en relación a la duración óptima del ayuno previo al sacrificio en peces no suelen considerar la temperatura del agua y se basan únicamente en días y no en grados día (ºC d).

Se determinó el efecto del ayuno previo al sacrificio (1, 2 y 3 días, equivalente a 11,1-68,0 grados día) y la hora de sacrificio (08h00, 14h00 y 20h00) en trucha arco iris (Oncorhynchus mykiss) de tamaño comercial en cuatro pruebas usando diferentes temperaturas de agua (Prueba 1: 11,8 ºC; Prueba 2: 19,2 ºC; Prueba 3: 11,1 ºC; y Prueba 4: 22,7 ºC). Se midieron indicadores biométricos, hematológicos, metabólicos y de calidad de la carne. En cada prueba, los valores de los animales ayunados (n=90) se compararon con 90 animales control mantenidos bajo condiciones similares pero nos ayunados.

Los resultados sugieren que el ayuno tuvo un efecto significativo sobre los indicadores biométricos. El coeficiente de condición en los animales ayunados fue menor que en los controles después de 2 días de ayuno. El vaciado del aparato digestivo se produjo durante las primeras 24 h de ayuno, encontrándose pequeñas cantidades de alimento después de 48 h. Por otra parte, este vaciado fue más rápido cuando las temperaturas fueron más altas. El peso del hígado de los animales ayunados fue menor y las diferencias entre truchas ayunadas y controles fueron más evidentes a medida que el vaciado del aparato digestivo fue más rápido.

El efecto del ayuno hasta 3 días en los indicadores hematológicos no fue significativo. Los niveles de cortisol en plasma resultaron ser altos tanto en truchas ayunadas como en las alimentadas en todas las pruebas realizadas. La concentración media de glucosa varió entre pruebas pero mostró una tendencia a disminuir en animales ayunados a medida que el ayuno

XXIII progresaba. En cualquier caso, parece que la temperatura del agua jugó un papel muy importante, ya que se encontraron concentraciones más altas durante los días 2 y 3 de ayuno en animales mantenidos a temperaturas más bajas previamente al sacrificio. Los altos niveles de lactato obtenidos en sangre parecen sugerir episodios de intensa actividad muscular pero no se pudo encontrar relación con el ayuno. De la misma manera, el nivel de hematocrito no mostró efecto alguno del ayuno y los leucocitos tendieron a ser más altos cuando los animales estaban menos estresados y cuando su condición corporal fue mayor. Finalmente, la disminución del peso del hígado (índice hepatosomático) en la Prueba 3 no se vio acompañada de una reducción del glucógeno hepático, lo que sugiere que las truchas emplearon una estrategia diferente para mantener constantes los niveles de glucosa durante el periodo de ayuno en esa prueba.

En relación a la hora de sacrificio, se obtuvieron niveles más bajos de cortisol a las 20h00, lo que indica que las truchas estaban menos estresadas y que el manejo pre-sacrificio podría resultar menos estresante por la noche. Los niveles de hematocrito fueron también más bajos a las 20h00 pero solo con temperaturas más bajas, sugiriendo que las altas temperaturas incrementan el metabolismo.

Ni el ayuno ni la hora de sacrificio tuvieron un efecto significativo sobre la evolución de la calidad de la carne durante los 3 días de almacenamiento. Por el contrario, el tiempo de almacenamiento sí que parece tener un efecto claro sobre los parámetros de calidad del producto final. Los niveles más bajos de pH se alcanzaron a las 24-48 h post mortem, con una lata variabilidad entre duraciones del ayuno (1, 2 y 3 días) en animales sacrificados a las 20h00, aunque no se pudo distinguir ningún patrón común. Por otra parte, la mayor rigidez asociada al rigor mortis se produjo a las 24 h del sacrificio. La capacidad de retención de agua se mostró muy estable durante el período de almacenamiento y parece ser independiente de los cambios en el pH. El parámetro L* de color se incrementó a medida que avanzaba el período de almacenamiento de la carne, mientras que los valores a* y b* no variaron en gran medida.

En conclusión, basándose en los resultados hematológicos, el sacrificio a última hora del día parece tener un efecto menos negativo en el bienestar. De manera general, nuestros resultados sugieren que la trucha arco iris puede soportar un período de ayuno previo al sacrificio de hasta 3 días o 68 ºC d sin que su bienestar se vea seriamente comprometido. Es probable que con temperaturas más bajas las truchas pudieran ser ayunadas durante más tiempo sin ningún efecto negativo sobre su bienestar. En cualquier caso, se necesitan más estudios para determinar la relación entre la temperatura del agua y la duración óptima del ayuno en términos de pérdida de peso vivo y la disminución de los niveles de glucosa en sangre y otros indicadores metabólicos.

XXIV

CHAPTER I

INTRODUCTION

1. The importance of aquaculture

The Food and Agriculture Organization (FAO) of the United Nations defines aquaculture as “the farming of aquatic organisms in inland and coastal areas, involving intervention in the rearing process to enhance production and the individual or corporate ownership of the stock being cultivated” (FAO, 2008). Capture fisheries and aquaculture supplied the world with about 159.2 million tonnes of fish in 2010 (68.4 million tonnes from aquaculture), of which about 128 million tonnes was to feed people. World fish food supply has grown dramatically in the last five decades, with an average growth rate of 3.2% per year in the period 1961-2009, while fish consumption reached a new high of 18.4 kg per capita in 2009 (FAO, 2012a). Fish production represents 49.5% of the total aquaculture production, with carps and tilapias as the most produced species. The aquaculture production of the European Union (UE-27) represents 2.1% of world production, with 1.2 million tonnes. Spain is the main producer with 252,352 tonnes (21.0% of European aquaculture production), followed by France (224,520 t) and United Kingdom (201,091 t), but Norway is the largest producer in Europe (including the UE-27) with 1 million tonnes.

If we consider that most fish produced for aquaculture weigh between 250-500 g, 1 tonne of fish represents approximately 2600 fish (using an average weight of 375 g). Thus, in terms of fish welfare, a large number of individual fish are involved in aquaculture production.

Figure 1. World capture fisheries and aquaculture production (FAO, 2012a)

1.1. Aquaculture in Spain

As mentioned above, Spain is the largest producer within UE-27 in terms of total aquaculture production (252,352 t), including aquatic plants. Production increased until the late 90s (up to 318,000 t; see Fig. 2).

Most aquaculture in Spain is carried out in sea water (229,322 tonnes) but fish represent only 41,460 tonnes of this production. At present, marine fish species cultivated at a commercial level include: gilthead seabream (Sparus aurata) with 23,690 t, European seabass (Dicentrarchus labrax) with 13,840 t, turbot (Psetta maxima) with 8,320 t and meagre (Argyrosomus regius) with 1,660 t. Other species with minor production include European eel (Anguilla anguilla), blackspot seabream (Pagellus bogaraveo), common sole (Solea vulgaris), red tuna (Thunnus thynnus), white seabream (Diplodus sargo), pollack (Pollachius pollachius) and Atlantic salmon (Salmo salar).

The most important species cultivated in Spain in 2010, in terms of their economic value, are: mussels, gilthead seabream, European seabass, turbot, tunids, meagre and Japanese clam (Ruditapes philippinarum). As mentioned above for general aquaculture, mariculture involves a great number of fish: cages can house 60 tonnes, which represents about 120,000 individuals, so welfare in these types of systems becomes complex if we consider it from an individual point of view.

Figure 2. Reported aquaculture production in Spain from 1950 (FAO, 2012a)

As for inland and brackish water fish, the most important species is rainbow trout (Oncorhynchus mykiss), constituting 6.8% of the total aquaculture production (17,384 t in 2010) and 99% of inland aquaculture production. There is also a small production (< 500 tonnes) of eel (Anguilla anguilla), tench (Tinca tinca) and sturgeon (Acipenser nacarii).

1.2. Trout

Trout is the name for a number of species of freshwater fish belonging to the genera Oncorhynchus, Salmo and Salvelinus, all from the subfamily Salmoninae of the family Salmonidae.

Rainbow trout is native to the Pacific drainages of North America, ranging from Alaska to Mexico. Since 1874 it has been introduced to waters on all continents except Antarctica, for recreational angling and aquaculture purposes (FAO, 2012b).

The species was originally named “trout” by Johann Julius Walbaum in 1792 based on type specimens from Kamchatka. Richardson named a specimen of this species Salmo gairdneri in 1836, and in 1855, W. P. Gibbons found a population and named it Salmo iridia, later corrected to Salmo irideus. These names fell out of use, however, once it was determined that Walbaum's type description was conspecific and therefore had precedence (see e.g. Behnke, 1966). More recently, DNA studies showed rainbow trout are genetically closer to Pacific salmon (Oncorhynchus species) than to brown trout (Salmo trutta) or Atlantic salmon (Salmo salar).

The production of rainbow trout has grown exponentially since the 1950s, especially in Europe during the 60s and 70s, reaching a peak in 2001 with more than 323,000 t (primarily due to increased inland production in countries such as France, Italy, Denmark, Germany and Spain to supply the domestic markets and mariculture in cages in Norway). Now it represents 257,200 tonnes, of which about 6.7% are produced in Spain (17,384 t), where production reached a maximum in 2001 (35,000 t) but there has been a marked reduction since then, as in the rest of species, mainly due to an increase in production costs, decrease in market value, competition from other species and rural abandonment (IS-AC, 2010). World production increased to 728,500 t in 2010 (Fig. 3), as with much of the rest of the aquaculture sector and is the third most produced species of fish around the world. Chile is currently the largest producer (220,000 t in 2010). Other major producing countries include Norway, France, Italy, Spain, Denmark, USA, Germany, Iran and the UK. The market worth of rainbow trout is 1,037 million USD in Europe and 61 million USD in Spain (5.9%).

Figure 3. Global aquaculture production of Onchorhynchus mykiss (FAO, 2012b)

2. Stress and welfare

2.1. Animal welfare

The study of animals and of human society are often considered as belonging to two different realms, with little or no contact or relevance for each other. However, ever since domestication, farm animals have been an integrated and important part of human society. Although the proportion of the population engaged in animal husbandry is decreasing in the Western world, the lives of farm animals and of humans are still entangled in many ways (Lund et al., 2006). Animal agriculture is not only of great economic interest, employing a considerable number of people. It is also on the basis of a significant part of human culture and tradition. As for the farm animals themselves, their quality of life depends on human care. This has become a matter of increasing concern in society, and is discussed in terms of animal welfare. In much of the Western world, animal welfare legislation is continually being tightened. The issue is also being addressed by intergovernmental bodies. In 2004, the World Organisation for Animal Health (OIE, Office International des Epizooties) declared that it had taken on the task of developing animal welfare standards for international trade agreements, thus preparing the ground for applying animal welfare requirements when trading in animal products or live animals. This will affect countries where animal welfare issues are less in focus today. In the European Union, animals were granted the status of ‘‘sentient beings’’ through the Treaty of Amsterdam in 1997

(European Union, 1997). As a result of all this, there is a demand for more knowledge regarding animal welfare – for example, what affects it, how to measure it, and how to implement it.

‘‘Animal welfare science’’ has become established as a research field of its own (Mench, 1998). The complex character of animal welfare problems seems to make animal welfare science exceptionally well suited for studies across disciplinary boundaries. Further, it can be argued that such studies are actually necessary for resolution of problems: this is an arena where consumers really ought to be involved. However, as is seen by surveying the literature, this seems to happen surprisingly seldom.

Farm animal welfare concern is based upon the belief that animals can suffer, and it is now clearly an important issue for ordinary people across Europe, who demand that animals be reared, transported and slaughtered in a humane way (Velarde & Dalmau, 2012).

Animal welfare is not a straightforward concept since it is difficult to measure directly and involves moral and ethical issues which may vary from person to person. To assure an objective approach, different governmental bodies and non-governmental organisations have developed guidelines for several species. For example, the farm Animal Welfare Council (FAWC) base their guidelines on the ‘‘Five Freedoms’’ framework, which defines ideal states rather than specific levels of acceptable welfare (FAWC, 2009):

Free from hunger and thirst Free from discomfort Free from pain, injury or disease Freedom to express normal behaviour Free from fear and distress

These freedoms have been highly influential and underpin much legislation and regulation about the welfare of farmed birds and mammals and, increasingly, of fish. They provide a logical framework with which to assess welfare issues (Ashley, 2007).

Animal welfare as an added value should be also taken into account. A European survey carried out in 2007 (European Commission, 2007) revealed that more than 63% of the 29,152 interviewed persons in the 25 member states of the European Union and four candidate countries show some willingness to change their usual place of shopping in order to be able to purchase more animal welfare friendly products. Furthermore, producers, retailers and other food chain participants increasingly recognize that consumer concerns for good animal welfare

7 represent a business opportunity that could be profitably incorporated in their commercial strategies (Roe & Buller, 2008). In general, animal welfare is increasingly perceived and used as an important attribute of an overall concept of “food quality” (Buller & Cesar, 2007).

2.2. Fish welfare

As mentioned above, animal welfare has become a hot topic during the last years. It is important since many human activities potentially affect the welfare of animals, and specifically of fish and people have strong views on whether and how they should be protected.

Fish welfare has arrived rather late on the scene since there is a heated debate on whether they are able to feel pain. Reasonable arguments have been made both to support and refute the claim. In 2010 Victoria Braithwaite published a book entitled “Do feel fish pain?”, which has had a great impact in the Anglosaxon world. Until recently, fish were not included in our concept of sensitive creatures but this is now changing. We now know that fish are not only capable of feeling pain but of suffering. Sneddon et al. (2003), at the University of Edinburgh, performing research on rainbow trout, demonstrated the presence of nociceptors on the face and snout of the trout. These are sensory receptors that respond to potentially damaging stimuli by sending nerve signals to the spinal cord and brain. Later research has demonstrated that fish feel both reflexive and cognitive pain (Sneddon, 2003). Those results have been confirmed during the following years in other laboratory species such as goldfish. For example, Nordgreen et al. (2009) reported that goldfish can feel pain and that when they were given morphine, anxiety, wariness and fear decreased.

These and many other studies indicate that there should be concern about fish welfare. This however is not a universally held belief and there are some authors that question the evidence that fish feel pain (Rose 2002). Rose argues that, since the fish brain is different from ours, fish are probably not conscious in the manner humans are, and while fish may react in a way similar to the way humans react to pain, the reactions in the case of fish have other causes. Rose published his own opinion a year earlier arguing that fish cannot feel pain because they lack the appropriate neocortex in the brain. However, most farmers have included stress and welfare indicators in their production program and many UE governments have developed legislation that promotes animal welfare in fish. These are discussed below.

The question of why it matters that human activities may compromise fish welfare is complex but there are both practical and ethical reasons for taking fish welfare seriously. On the practical

8 front, industry should be engaged with improving the welfare of farmed fish because good production and good flesh quality often follow from good welfare. Additionally, the public is increasingly concerned about welfare of fish and it has become a marketing issue, so there is a price premium for producing fish in a way that maximizes welfare. On the ethical front, the industry should be concerned with fish welfare because there are moral obligations to the animals which are farmed (Huntingford & Kadri, 2008). In this sense, it is essential to develop frameworks to guide clear thinking on this complex moral issue.

Partly due to the rapid expansion of aquaculture all over the world, the welfare of farmed fish has received increasing attention. As said above, fish welfare is an important issue for the industry, not just for public perception, marketing and product acceptance, but also often in terms of production efficiency, quality and quantity (FSBI, 2002). However, in a number of areas, conflicts between welfare and production exist, where procedures are seemingly associated with diminished welfare at the level of the individual fish. Such practices have been the subject of much commentary and discussion in recent years (Lymbery, 2002; Hastein, 2004). In the EU, Council Directive 98/58/EC lays down minimum standards for the protection of animals bred or kept for farming purposes, although no specific rules concerning fish are mentioned. International organisations have also issued recommendations and guidelines concerning fish welfare. In 2005 the Council of Europe adopted a recommendation on the welfare of farmed fish (Council of Europe, 2005), where starvation before certain management practices or slaughter was recommended in order to reduce metabolism and excretion of waste products. In 2008 the World Organisation for Animal Health (OIE) adopted guiding principles for fish welfare in the Aquatic Animal Code, which has been updated in 2012 (OIE, 2012). During the last few years, the Panel on Animal Health and Welfare (AHAW) adopted an opinion from the European Food Safety Authority (EFSA) on husbandry systems, on the general approach to welfare and on the welfare aspects of stunning and killing methods for different species of farmed fish (EFSA, 2009). Finally, a number of codes of practice have also been adopted by industry that includes measures to safeguard fish welfare.

Some organisations such as the FAWC and the Humane Slaughter Association (HSA) have published recommendations regarding the welfare of farmed fish (FAWC, 1996; HSA, 2005). This is a growing area where research is lacking and these reports identify the need for an improvement in scientific knowledge on which to base future guidelines and potential legislation where necessary.

3. Pre-slaughter handling

Harvest and slaughter come at the end of the production cycle of all animals farmed for food. The focus is to avoid financial loss due to premature death or visible injury. During the process, a great number of actions are implemented which can affect welfare. Stress during the ante mortem period may interact and influence both the animal’s physiological reactions – in this way worsening the animal performance – and the post mortem biochemical processes that turn muscle into meat, in this way affecting the quality and durability of the final product (Azam et al. 1989; Lowe et al. 1993; Marx et al. 1997; Thomas et al. 1999). Often there are unavoidable situations where multiple stressors interact and affect the final condition. Among these situations, crowding, medication withdrawal, removal from the cage or raceway, transport, stunning and the killing method are the most important. Parisi et al. (2001) reported the effects of pre-slaughter procedures such as crowding prior to slaughter and repeated catching on cortisol and haematocrit levels. They both increased in stressed fish. As in poultry production, aquaculture involves the handling of relatively small animals (less than 1 kg) with many thousands of individuals in small spaces (between 37 and 120 billion farmed fish were slaughtered for food in 2010; Mood & Brooke, 2012). Thus, from an individual point of view, animal welfare is an important issue in aquaculture and it may be difficult to ensure all five freedoms under current conditions.

3.1. Fasting

The restriction of feed for ﬁsh is a natural occurrence in wild populations and also occurs regularly in cultured ﬁsh. Wild populations undergo feed restrictions due to limitation in the forage base, weather events and during certain phases of their reproductive cycle. Aquaculture species undergo periods of feed deprivation in response to some of the same cues as their wild counterparts but also imposed by producers. Fasting in aquaculture species is often carried out for several reasons, some of which may be to ameliorate water quality problems, reduce handling stress, reduce negative effects of disease outbreaks, or merely due to an inability to feed due to inclement weather (Davis & Gaylord, 2011). Another reason may be to induce compensatory growth after starvation and re-feeding (Gaylord & Gatlin, 2000). This starvation period is usually several weeks long (Pottinger et al., 2003) and has dramatic consequences for the fish metabolism, varying the use of carbohydrates, lipids and proteins from different body compartments and mobilizing energy reserves (Barcellos et al., 2010). However, it is essential to differentiate between physiological states of fasting (i.e., the early phase of food deprivation during which animals mobilize readily available metabolite reserves, usually less than 7-10 days (Soengas et al., 1996)) and starvation (i.e., the chronic fasting period associated with weight

10 loss and pronounced protein and lipid catabolism). Although the two stages of food deprivation represent a physiological continuum, they do represent significantly different metabolic states and thus endocrine control of the two states might be expected to differ markedly (Farbridge & Leatherland, 1992b).

Fish are subject to periods of fasting before certain management procedures are carried out, such as transport, treatment of disease, transfer of smolts from fresh water to seawater and slaughter. This is designed to reduce physiological stress (Ashley, 2007), as it evacuates the gut and reduces metabolism, oxygen demand and waste production (Lines & Spence, 2012). By removing the feed, the bacterial populations are reduced, reaching their lowest after about three days of feed withdrawal and then they start to increase again. In addition, the empty gut is necessary from a food safety point of view, as a full gastrointestinal tract is very likely to become punctured during evisceration, which will result in gut enzymes and potential pathogens being spread on the flesh (Robb, 2008). However, according to FAWC (1996), depriving a farmed fish which has been fed regularly will normally cause some adverse effect on welfare. Though controlled food deprivation can be beneficial, it should not be carried out as a matter of course. In order to assure that the overall effect of food deprivation on welfare is low, any beneficial consequence must be sufficient to counterbalance direct adverse effects.

As fish are ectothermic, periods of food deprivation may be of less detriment than in endotherms (FSBI, 2002), since they can adjust their metabolic rate in response to the availability of nutrition. However, the motivation to eat is clearly strong and as necessary to fish as to any other animal. The critical question is whether the duration has been minimised to that simply needed to empty the gut.

Very little is known about the effect of short term fasting on fish that are normally fed regularly. If prolonged, fasting will decrease live weight (Sumpter et al., 1991) but it is not clear how much pre-slaughter fasting affects trout, since it normally only lasts a few days. In commercial practice, a range of food withdrawal periods are to be found, often far longer than is necessary to simply empty the gut. Some associations and institutions provide general rules regarding starving before slaughter in fish. According to FAWC (1996), trout should not be totally deprived of food except during a period of up to 48 hours before slaughter for food hygiene reasons, or where the overall effect of food deprivation is an improvement in fish welfare. However, the HSA (2005) points out that current research suggests that 72 hours is adequate for the complete emptying of the fish gut whilst minimising adverse welfare effects and fish should therefore not be fasted for longer than 72 hours. The Compassion in World Farming Association (CIWF, 2009) indicates that prolonged periods of starvation, exceeding 72 h, should be avoided

11 in Atlantic salmon and rainbow trout. Lymbery (2002) conclude that starvation periods longer than 48-72 h for farmed fish prior to slaughter are excessive and highly likely to be detrimental to the welfare of fish. Wall (2001) recommends that pre-slaughter fasting should last long enough to empty the stomach, but they do not provide information on the optimum period or rates of stomach emptying. EFSA (2008) indicates that it is not possible to specify a maximum acceptable duration of fasting, since its impact on welfare is related to size, lipid reserves, life stage and water temperature. However, EFSA (2009) points out that prolonged feed deprivation of more than 50 degree days (ºC d) can result in the utilization of body fat reserves and then functional tissue (which is associated with poor welfare). Robb (2008) considers that no evidence exists for additional benefits from fasting salmon for longer than 72 h. However, Lines & Spence (2012) conclude that a fasting period of 1-5 days prior to slaughter is unlikely to cause significant welfare problems and Nikki et al. (2004) reported no weight significant effects of short term fasting in rainbow trout (14 days).

There have been relatively few studies on the effects of starvation on stress physiology or behaviour (Barton et al., 1988; Pottinger at al., 2003; Jentoft et al., 2005) and the majority of work in this area concerns the effect of prolonged starvation on growth, muscle protein and fat composition (Sumpter et al., 1991; Einen et al., 1998; Pirhonen et al., 2003; Tripathi & Verma, 2003), not the effect on stress indicators.

Evidence for quality benefits from prolonged fasting is weak and mixed. In the past, long periods of fasting were used in the belief that this could improve the flesh quality by reducing oil levels (Wall, 2001), but other research has not demonstrated that there is any significant loss of fat even over extended periods of food withdrawal (Einen et al., 1998). Trials with brown trout have indicated that prolonged pre-harvest fasting can slightly improve the flesh colour, but this is at the expense of weight loss and fillet yield (Regost et al., 2001). Two commercial trials (unpublished) have been carried out, concluding that flesh flavour from non-fasted fish was preferred (Robb, 2008). Work with sea bream has found a measurable but possibly insignificant increase in muscle firmness but a decrease in the shelf life (Ginés et al., 2002; Álvarez et al., 2008). Mørkøre et al. (2008) concluded that prolonged fasting improves the ability of salmon to withstand severe stress during harvesting; however, this problem is more properly handled by directly addressing the causes of stress. Periodic feeding and starving throughout the growing cycle have been shown to have some beneficial quality effects in halibut culture (Foss et al, 2009). However, there is no reason to expect that prolonged fasting just at the end of the production cycle will have the same effect (Lines & Spence, 2012).

It is also important to consider other effects of starvation and reduced nutrition, such as changes in metabolic activity (Martínez et al., 2003), changes in behaviour related to competition, and the potential for increased aggression (Brannas et al., 2003). For example, reduced food abundance can cause changes in territorial behaviour strategies and activity patterns in brown trout, Salmo trutta (Alanara et al., 2001).

3.2. Hour of slaughter

Very little is known about the best time of the day to slaughter fish, which may in turn be dictated by diurnal rhythms in physiological stress parameters. These rhythms of activity in fish play an important role in the temporal organization and their adaptive value is to permit the adjustment of the life processes to the chronological arrangement of the external world (De Pedro et al., 1998). In mammals, plasma concentration of metabolic and hormonal products related to energy balance varies with a well-established dynamic, depending on whether the organism is well fed, fasted or starved (Escobar et al., 1998).

In fish, the effects of fasting on daily changes of several plasma metabolites and hormones like glucose, lactate, cortisol, GH or thyroxin have been characterized in a number of species demonstrating in several cases that some of those changes are dependent on feeding (Figueroa et al., 2000; De Pedro et al., 2005).

Plasma cortisol levels have been extensively studied in fish and are perhaps the best characterized physiological variable with respect to temporal rhythms (Boujard & Leatherland, 1992). Cortisol show a post-prandial peak (catfish: Small, 2005; rainbow trout: Polakof et al., 2007), with minimum values at night (Reddy & Leatherland, 1995) and these changes seem to be independent from feeding (Polakof et al., 2007).

Plasma glucose follows the same trend in fed rainbow trout, with higher values after feeding time (Holloway et al., 1994; Figueroa et al., 2000) and at night (Polakof et al., 2007), as also reported in other species (Pavlidis et al., 1999a). Fasting is the most frequently used approach to study whether a rhythm in glucose metabolism is independent of the disturbing/masking effect of feeding since those daily rhythms dependent on feeding should disappear in food deprived animals (De Pedro et al., 2005; Polakof et al., 2007).

Lactate diurnal rhythms have also been studied in rainbow trout, with no pattern discerned (Polakof et al., 2007) in fasted or fed fish, in contrast with the circadian-like pattern of plasma

13 lactate observed in fed red porgy (Pavlidis et al., 1999b), with higher values at 06h00 and 14h00.

No haematocrit rhythm has been found in rainbow trout, although Hannah & Pickford (1981) found a rise in haematocrit levels 8 h after the onset of light in killfish (Fundulus heteroclitus) and De Pedro et al. (2005) report higher values at night in tench, similarly to cortisol and glucose in several fish species.

Finally, leucocytes have been reported to vary seasonally and greatly influenced by exogenous factors such as photoperiod, temperature changes, feeding and stress in tench (De Pedro et al., 2005) and these factors in turn are dependent on the season.

The choice of the optimal time of day to slaughter trout is of great importance because it can determine when the effects of transport, pre-slaughter handling and slaughter may be less stressful. A lower stress response during those procedures should also result in fewer problems in relation to flesh quality. According to the literature reviewed above, no clear conclusions can be made so far about the best time to slaughter trout. In some studies, stress indicators are higher in the morning and in others at night while in others there is not clear rhythm at all. In addition, few studies have considered a wide variety of indicators and their diurnal rhythms, comparing fasted fish with a significant number of control (i.e., fed) fish. That is important to be able to suggest a time frame in which commercial slaughter is most feasible.

3.3. Water temperature

Temperature is a vitally important physical property of the water in aquaculture systems. The temperature of the water regulates the amount of dissolved oxygen, the ionisation of ammonia, growth and infectiousness of many fish pathogens and the toxicity of many dissolved contaminants (MacIntyre et al., 2008). But it also affects feeding, reproduction, immunity and the metabolism of aquatic animals (as fish are exothermic, increasing the water temperature increases the metabolic rate and hence oxygen consumption). All of these factors have the capacity to compromise the health and welfare of farmed fish.

Optimal temperatures for growth have been widely studied for many species important to aquaculture. The optimum temperature range for rainbow trout is suggested to be 13-19 ºC when oxygen is not limited (Schurmann et al., 1991) and can be fed at a temperature up to 20- 22 °C (FAO, 2011) without loss of appetite but it greatly depends on the length of acclimation

14 time, the dissolved oxygen concentration and the ions present in the water (Wedemeyer, 1996). Upper lethal temperature has been estimated to be within the range of 25-30 ºC (Cherry et al., 1977; Currie et al., 1998), whereas EFSA (2008) point out that this higher limit is above 24 ºC. Long-term exposure to elevated temperatures during farming of rainbow trout will impair growth performance by increasing the fish metabolism making them more sensitive to hypoxia because the oxygen carrying capacity is low (Skov et al., 2011) and increasing the excretory products (Azevedo et al., 1998). Another important factor is that warm water can hold less oxygen than cold water. It is well known that low dissolved oxygen levels trigger stress responses in fish, particularly those involving adrenergic stimulation of respiratory adjustments, such as erythrocyte swelling or cardiovascular changes (Caldwell & Hinshaw, 1994; Pickering, 1998). Appetite and thus the growth rate are also reduced (FAO, 2012). According to these data, high water temperatures may have a major role on stress. However, very little is known about how temperature is related with fasting stress prior to slaughter, as there is very little information on short-term exposure within the high-end optimal range of temperatures.

Water temperature is a determinant factor affecting fish productive period. During fish farming, degree days are used to estimate the length of time required to develop to the next growing stage, such as the egg incubation, hatching or fattening (From & Rasmussen, 1991). However, very little research has been carried out in fish. Caggiano (1999) pointed out that the length of the starving period prior to slaughter is influenced by feeding rate and water temperature, so recommendations should be given in degree days instead of only days. In addition, EFSA (2009) has also recommended to avoid fasting beyond 50 degree days.

4. How to measure welfare

4.1. Biometric indicators

The effect of food deprivation on body weight reduction is obvious but the time that it takes to start decreasing is unclear, although several days or even weeks are necessary in rainbow trout (Sumpter et al., 1991; Pottinger et al., 2003) and other species (Peterson & Small, 2004; Caruso et al., 2011). There are some exceptions, as Figueiredo-Garutti et al., 2002, who found significant reductions after only 24 h in Brycon cephalus. However, other studies report no differences between fasted and fed fish (rainbow trout, 6 days: McMillan & Houlihan, 1991; Atlantic salmon, 7 days: Soengas et al., 1996; rainbow trout, 14 days: Nikki et al., 2004).

The body condition or coefficient of condition (CC) is a widely used indicator of the nutritional status of fish (Ratz et al., 2003; Bavcevic et al., 2010; Chatzifotis et al., 2011), as it is more

15 accurate than body weight because it also involves body length. It has been found to vary between fasted and fed fish after 7 days of fasting in rainbow trout (Sumpter et al., 1991; Pottinger et al., 2003) and also in other species (Peterson & Small, 2004; Caruso et al., 2011).

Gut clearance is important because it determines the period of time that fish need to be fasted before certain handling procedures such as transport or slaughter (Wall, 2001; CIWF, 2009). It normally takes around 2-3 days in rainbow trout (McMillan & Houlihan, 1992) and other species (Robb, 2008). However, current recommendations should be probably reviewed, as a great number of factors, including water temperature and body, size play a very important role in gastric evacuation (Usher et al., 1991, Koed et al., 2001).

The hepatosomatic index (HSI) is defined as the ratio of liver weight to body weight. It provides an indication on status of energy reserve in an animal, as the liver is an important regulatory organ in nutrient utilisation in both teleosts and mammals (Christiansen & Klungsoyr, 1987). Liver weight normally decreases with food deprivation in fish (Blasco et al., 1992; McMillan & Houlihan, 1992; Barcellos et al., 2010; Costas et al., 2011) but the time that it takes to start decreasing may vary with species and experimental conditions. In any case, the minimum period seem to be 2 days (Farbridge & Leatherland, 1992b). This reduction in liver weight is a consequence of the utilization of nutrients stored (Davis & Gaylord, 2011), mainly glycogen (Soengas et al., 1996) but other liver metabolites, such as lipids or protein, have been found to decrease (Blasco et al., 1992; McMillan & Houlihan, 1992).

4.2. Stress indicators

As discussed above, welfare is difficult to define. However, we know that human activities can compromise fish welfare. Thus, there is a need to measure welfare objectively (Huntingford & Kadri, 2008). Although there is no simple relationship between stress and welfare, stress responses can be used to give an idea about the degree of metabolic challenge, especially where several components of the stress response are all influenced in a similar way by the same condition. The key issue in welfare evaluation at the time of slaughter is animal suffering, which is also difficult to measure. According to several authors, although stress physiology does not necessarily equate to suffering and diminished welfare and will not tell us all we need to know about the welfare of fish, concurrent deleterious effects in several of the indicators that can be measured (reductions in growth, suppressed reproductive function, diminished immune function, disease resistance and other physiological indicators) may provide a strong indication that welfare is poor (Ashley, 2007). Thus, a reliable assessment of animal welfare-suffering and

16 of the impact on product quality, requires a multidisciplinary approach that takes into account the different biochemical and physiological processes involved (Poli et al., 2005).

4.3. Haematological indicators

As in all the other vertebrates, fish react with a well-characterised neuroendocrine ‘‘stress response’’ of adaptive value (Wendelaar Bonga, 1997) to whatever undesirable challenge is threatening the animal homeostasis, i.e. the complex physiological processes that contribute to maintain the specific ranges of temperature, pH and solute concentration, necessary for the maintenance of the normal physiological functions. The study of stress in fish in an aquaculture setting has a relatively short history compared to stress research in domestic and laboratory animals but we already have a sophisticated, though incomplete, appreciation of the complex nature of the stress response in fish. This response can be divided into the primary neuroendocrine response and the secondary response.

Figure 4. Overview of the HPI axis in fish (adapted from Alsop & Vijayan, 2009)

Acute primary response to stress is the activation of the sympathetic nervous system whose endpoint is the immediate massive release into the blood of catecholamines (adrenaline and noradrenaline) from the chromaffin cells (the teleostean homologue of the adrenal medulla). This is followed by a gradual ACTH release that triggers off the release of corticosteroids, such as cortisol, from the interrenal tissue into the blood stream. This arm of the response is often referred to as the hypothalamus-pituitary-interrenal (HPI) axis (Barton, 2002). The elevated

17 levels of catecholamines and cortisol in the blood represent the primary response to a stressor. Catecholamines (adrenaline and noradrenaline) are not commonly used as stress indicators because they are not easy to determine and quickly removed from blood (Wendelaar Bonga 1997).

The consequences of these endocrine responses lead to a suite of cardio-respiratory changes that include an increase in heart-beat, higher oxygen uptake, energy source mobilisation and increase in plasma glucose (glycogenolysis). The last is easy to determine so it is frequently used as a stress indicator, although some authors have found a delay in its release (Barry et al. 1993).

Figure 5. Overview of dynamics of cortisol and catecholamines in the production of glucose (adapted from Martínez-Porchas et al., 2009)

Higher energy mobilization and utilisation, following the increase of muscular activity, initiates anaerobic glycolysis and a related increase in plasma lactate. Therefore, the increase of plasma lactate is used as stress index (Lowe et al. 1993; Erikson et al. 1999), even if most fish store lactate in muscle tissue.

The increase in heart-beat and the need of higher oxygen uptake increase the number of moving erythrocytes and of the haematocrit value. This is also used as a stress index because it is very simple to determine (Reddy & Leatherland, 1988). The primary response also implies mobilization of carbohydrates and lipid reserves (Pickering & Pottinger, 1995).

4.3.1. Plasma cortisol

Cortisol is the principal glucocorticoid secreted by the interrenal tissue (steroidogenic cells) located in the head-kidney of teleost fish (Iwama et al. 1999). This hormone is released by the activation of the hypothalamus-pituitary-interrenal axis (HPI axis) (Mommsen et al. 1999). When an organism undergoes stress conditions, the hypothalamus releases corticotropin- releasing factor (CRF) toward blood circulation. This polypeptide further stimulates secretion of adrenocorticotrophic hormone (ACTH) from the anterior pituitary gland (Fryer & Lederis 1986) which finally activates the release of cortisol by the interrenal tissue (Mommsen et al. 1999). The secretion of cortisol is slower than catecholamines, but its effects are more prolonged (Gamperl et al. 1994; Waring et al. 1996), combining mineral and glucocorticoid actions to restore homeostasis (Wendelaar-Bonga 1997; Colombe et al. 2000). Cortisol activates glycogenolysis and gluconeogenesis processes in fish; but also causes that chromaffin cells increase the release of catecholamines which further increase glycogenolysis and modulate cardiovascular and respiratory function (Reid et al. 1992, Reid et al. 1998). This whole process increases the substrate levels (glucose) to produce enough energy according with the demand.

It is by far the most used indicator of both long and short term stress (Pickering et al. 1982; Pickering & Pottinger 1985). This is due to the ease with which it can be measured and the ubiquity of its response to virtually all the forms of stress. In experiments of acute stress, the cortisol response is rapid but regularly becomes weak or disappears some hours after the exposure to stress (Davis Jr. & McEntire 2006). In most fishes, cortisol reaches highest concentration 1 hour after being stressed, and returns to basal levels after 6 hours (Iwama et al. 2006). Cortisol levels of red drum during some handling procedures grew rapidly, but decreased to the basal state within 48 hours (Robertson et al. 1987) to avoid tissue damage (Wendelaar- Bonga 1997). In chronic-stress experiments some fish showed a weak increase of cortisol (Barton et al. 2005, Fast et al. 2008) probably caused by exhaustion of the endocrine system as a result of prolonged hyperactivity (Hontela et al. 1992) or an habituation of the organism to that condition.

Nowadays, cortisol is detectable using very precise and sensitive assay systems, among which RIA (radioimmunoassay) and ELISA (enzyme-linked immunosorbent assay or enzyme immunoassay) are the most used. Although cortisol is a reliable indicator of stress, there are several factors that can modify the response when measured (developmental stage, sexual maturity, water temperature, acclimation to repeated stress, inter-individual variation within a population and between-species variation).

Plasma levels of cortisol are consistently reported to be elevated in response to fasting in mammals (Ortiz et al., 2001; Chang et al., 2002). In contrast, the evidence in fish is contradictory. In the current literature, data about the effects of fasting on cortisol levels were inconsistent. There are results reporting no effect of starvation on cortisol levels (Holloway et al., 1994; Reddy et al., 1995). Others report decreased cortisol levels in fasted fish (Barton et al., 1988; Farbridge & Leatherland, 1992a; Small, 2005) and still others reported increased cortisol levels (Sumpter et al., 1991; Varnavsky et al., 1995; Blom et al., 2000; Peterson & Small, 2004). Due to these highly variable results, the effects of food deprivation on cortisol are not fully understood.

4.3.2. Plasma glucose

Glucose is a carbohydrate that has a major role in the bioenergetics of animals, being transformed to chemical energy (ATP), which in turn can be expressed as mechanical energy (Lucas, 1996). In suboptimum or stressful conditions (internal or external) the chromaffin cells release catecholamine hormones, adrenaline and noradrenaline toward blood circulation (Reid et al. 1998). Those stress hormones in conjunction with cortisol mobilize and elevate glucose production in fish through glycogenesis and glycogenolysis pathways (Iwama et al. 1999) to cope with the energy demand produced by the stressor. This glucose production is mostly mediated by the action of cortisol which stimulates liver gluconeogenesis and also halts peripheral sugar uptake (Wedemeyer et al. 1990). Glucose is then released (from liver and muscle) toward blood circulation and enters into cells through the insulin action (Nelson & Cox 2005). In the short-term, hyperglycaemia is mediated primarily as a consequence of the rapid increase in blood catecholamines which occurs during the first phase of the stress response (Wendelaar Bonga, 1997). Although in the longer-term, cortisol-dependent gluconeogenesis may be responsible (van Raaij et al., 1996). Regardless of the wide use of glucose as an indicator of stress, some authors (Mommsen et al. 1999, Flodmark et al. 2001) emphasize that care has to be taken when using plasma glucose as a single parameter to measure stress. It has been reported that glucose content is a less precise indicator of stress than cortisol (Wedemeyer et al. 1990; Pottinger, 1998) and Simontacchi et al. (2008) stated that glucose cannot be considered by itself as a reliable stress indicator. Therefore, the utilization of several parameters may be considered as a more solid indicator of stress.

Nutritional status and thus fasting, is a factor that can have an effect in the glucose response. In most fish species, food deprivation leads to hypoglycaemia (Pérez-Jiménez et al., 2012) and it appears as soon as 1-2 days after the cessation of feeding (Soengas et al., 1996; Pérez-Jiménez

20 et al., 2007). However, at least 7 days of fasting is necessary to produce a significant reduction in most species (Navarro et al., 1992; Holloway et al., 1994), although other authors report no changes after prolonged starvation (Barcellos et al., 2010), probably because the elevated cortisol levels played a role in the mobilization of energy substrates, stimulating glycogenolysis and gluconeogenesis. In rainbow trout, Pottinger et al. (2003) found a reduction in the glucose concentrations after three days of fasting and Furné et al. (2012) demonstrated that 5 days was sufficient to induce hypoglycaemia.

4.3.3. Plasma lactate

Lactate is a chemical compound that plays a role in anaerobic metabolism of animals, produced from pyruvate via the enzyme lactate dehydrogenase during exercise in fast glycolytic white muscle fibres during “burst” swimming. It is considered as an acute stress indicator in fishes because its levels are enhanced under adverse situations (Thomas et al. 1999; Grutter & Pankhurst, 2000) and some of the acid and lactate produced in the muscle enters the blood space, resulting in an increase in blood lactate. Although lactate is more associated with acute stress, it can be affected by fasting, since the lack of food may lead adjustments in social hierarchy or increased aggressiveness (Blanco Cachafeiro, 1995).

4.3.4. Haematocrit

Haematocrit is the volume percentage of red blood cells in blood. As mentioned above, higher heart beat frequency and oxygen demand will increase the haematocrit value. Along with lactate, haematocrit is linked to high muscular activity during acute stress (e.g. before transport, the capture process or during pre-slaughter procedures). Parisi et al. (2001) reported the effects of pre-slaughter procedures such as crowding prior to slaughter and repeated catching on cortisol and haematocrit levels in several fish species. They both increased in stressed fish. However, very little information is available regarding food deprivation. Davis & Gaylord (2011) found a high variability in haematocrit values in fasted and fed sunshine bass, while it was not consistently affected by fasting. In contrast, De Pedro et al. (2005) and Ríos et al. (2002) state that haematocrit is clearly affected by long-term fasting. However, there are no studies in rainbow trout.

4.3.5. Leucocytes

Leucocytes, or white blood cells, are cells from the immune system involved in defending the body against both infectious disease and foreign materials. There are different types of white cells such as neutrophils, eosinophils, basophils, lymphocytes and monocytes. Differential effects of stressors can be found on different immune cell types. During the last few years they have been used as a complementary method of physiological stress assessment, as an increase in glucocorticoid hormones cause characteristic changes in the leucocyte component of the vertebrate immune system that can be quantiﬁed and related to hormone levels (Davis et al., 2008). Actually, cortisol has been reported to have a major effect on lymphocytes, both decreasing the number and their proliferation (Harris & Bird, 2000). The leucocyte response to stress has been studied in ﬁsh, inducing lymphopenia (decrease of the blood lymphocytes) following stress (Maule & Schreck, 1990; Ainsworth et al., 1991; Pulsford et al., 1994; Engelsma et al., 2003) and neutrophilia (high number of neutrophil granulocytes: Morgan et al., 1993; Narnaware & Baker, 1996). It also may cause leucopenia or leucocytosis (Wedemeyer et al., 1990), suggesting possible immune function alterations (Huffman et al., 1997). Although leucocytes have been extensively studied, especially as a stress indicator after the exposure to heavy metals (Dethloff et al., 1999; Witeska, 2005), no research has been carried out to evaluate the effect of fasting on this parameter.

4.4. Metabolic indicators

4.4.1. Liver and muscle glycogen

Carbohydrates are the primary and earliest source of energy when fish are food deprived (Navarro & Gutiérrez, 1995; Davis & Gaylord, 2011). Although food deprivation leads to hypoglycaemia in most ﬁsh species (Pérez-Jiménez et al., 2012), glycogen (the form of carbohydrate stored in the liver) is enzymatically broken down and transported to the extrahepatic tissues as glucose to maintain circulating glucose levels for normal metabolic fuel during periods of fasting. The extent to which reserves are sequestered depends on many variables, including size of the fish, reproductive history, availability, and type of food and thermal environment (Metcalfe & Thorpe, 1992). Usually, glycogen depletion is a continuous process from the beginning of fasting (Navarro and Gutiérrez, 1995). However, the time that it takes to start decreasing glycogen depots is unclear, although it is generally of 5-20 days after the cessation of feeding (De Pedro et al., 2003; Sangiao- Alvarellos et al., 2005; Barcellos et al., 2010). However, Gaylord & Gatlin (2000) demonstrated that feed deprivation of channel catfish for as little as 2 days will dramatically influence liver size and liver glycogen stores.

Regarding glycogen in muscle, it can be decreased or otherwise maintained at the expense of hepatic glucose release (Navarro & Gutiérrez, 1995). Navarro et al. (1992) and Furné et al (2012) found significant reductions in rainbow trout and brown trout after fasting, respectively, whereas Kieffer & Tufts (1998) and Barcellos et al. (2010) report no changes in rainbow trout and jundia.

4.4.2. R-value (IMP/ATP ratio)

It has been accepted that ATP depletion is the actual cause of the onset of rigor mortis (Greaser, 1986). After death, ATP is rapidly split into ADP (adenosine diphosphate) and subsequently into AMP (adenosine monophosphate) and IMP (inosine monophosphate). This compound is then degraded to inosine (Ino) and hypoxanthine (Hx). Thus, the IMP/ATP ratio may be considered as a reliable indicator of the degree of ATP depletion and thus of the state of the animal before the slaughter. It has been described that the breakdown of adenosine nucleotides to IMP and inosine occurs much faster in stressed versus non stressed animals. These types of muscles contain low levels of ATP and high levels of IMP at early stages post mortem while high concentrations of ATP and low amounts of IMP are observed in normal muscles (Essen- Gustavsson, 1991).

In fish, this ratio has been also used as a very useful indicator of stress just prior to slaughter (Korhonen et al., 1990). According to these authors, stressed animals have shown significantly higher values both in the pre rigor state and at rigor onset compared to unstressed fish. However, it has been mainly used to evaluate the effect of different slaughter methods in fish (Giuffrida et al., 2007) but no studies have been carried out to determine the effect of fasting prior to slaughter.

Figure 10. Post mortem ATP degradation in fish muscle. (from Gill, 1992)

5. Product quality

As clearly proved in many other terrestrial meat animals, good welfare often means good production. Pre-slaughter handling should be carried out with care, since severe stress can have a negative impact on meat quality and shelf life. This is an ethical aspect that also has clear implications for fish quality. Animal welfare and product quality are tied together and influence the total quality of the fish, therefore any conflict between the requirements of fish welfare and efficient aquaculture should be avoided. To maintain the best original quality, fish should be stunned and bled with minimal stress using appropriate handling procedures.

An animal which has not been stressed will have normal levels of glycogen in its body. When the animal is slaughtered and exsanguinated, the metabolic process continues, however there is no longer circulating oxygen. Without the presents of oxygen, the breakdown of glycogen/glucose results in a buildup of lactic acid which then causes a drop in pH of the meat.

In rested animals, creatine phosphate (CP) helps immediately to regenerate muscle ATP when there is a need to maintain high levels. However, many animals have little or no CP in muscles at the time of slaughter because it has been used up during animal transport, lairage or during any stressful procedure prior to slaughter. Therefore, ATP must be generated from another source of energy production. Glycogen is the major storage form of energy in muscle and is metabolized to help supply ATP, thus keeping muscles in the pre-rigor state. If there was a great lactic acid buildup before slaughter, mainly due to the stress, the pH of the meat declines too quickly after slaughter and a pale, soft and exudative (PSE) condition may develop. If carcasses have an accelerated pH decline (as with severe PSE) due to the rapid production and consumption of ATP, the accumulation of lactic acid is more rapid and rigor onset is much faster (less than 1 h to 4 h). Unfortunately, the muscle temperature is often still high, which facilitates higher than normal denaturation of protein and increasing the WHC (water holding capacity). Conversely, if the animal is glycogen depleted before slaughter the pH may not drop quickly enough after slaughter because there is not enough lactic acid produced. In this case the meat will be very dry and dark in colour. This condition is known as dark, firm and dry (DFD) meat. An additional problem with this type of meat is that it is more susceptible to spoiling since it lacks the lactic acid which normally helps retard growth of microorganisms after slaughter. Note that glycogen deficiency may also be the result of too much physical activity or inadequate diet before slaughter (HSI, 2001).

Autolysis means "self-digestion". It has been known for many years that there are at least two types of fish spoilage: bacterial and enzymatic (Huss, 1995). Regarding enzymatic spoilage,

24 increased muscular activity, stress at slaughter and relative endocrine response can greatly influence fish post mortem biochemical processes, mostly the anaerobic glycolysis and ATP degradation rate. This in turn can markedly influence the onset and release of rigor mortis, which in turn largely determines the involution rate of fish freshness. In this way it leads to undesirable changes mostly in the marketable, physical and freshness quality parameters. The close interrelationships observed among endocrine acute stress responses and fish post mortem biochemical processes suggest to use not only haematic stress indicators like plasma cortisol, lactate and glucose, but also tissue stress indicators such as muscular pH, glycogen, water holding capacity or rigor mortis.

Figure 6. Aerobic and anaerobic breakdown of glycogen in fish muscle (Huss, 1995)

At the point of death, the supply of oxygen to the muscle tissue is interrupted because the blood is no longer pumped by the heart and is not circulated through the gills where, in the living fish, it becomes enriched with oxygen. Since no oxygen is available for normal respiration, the production of energy from ingested nutrients is greatly restricted. Figure 6 illustrates the normal pathway for the production of muscle energy in most living teleost fish (bony finfish). Glycogen (stored carbohydrate) or fat is oxidized or "burned" by the tissue enzymes in a series of reactions which ultimately produce carbon dioxide (CO2), water and the energy-rich organic compound adenosine triphosphate (ATP). This type of respiration takes place in two stages: an anaerobic and an aerobic stage. The latter depends on the continued presence of oxygen (O2) which is only available from the circulatory system. This figure also illustrates that, under anaerobic conditions, ATP may be synthesized by two other important pathways from creatine phosphate or from arginine phosphate. The former source of energy is restricted to vertebrate muscle (teleost fish) while the latter is characteristic of some invertebrates such as the cephalopods (squid and octopus). In either case, ATP production ceases when the creatine or arginine phosphates are depleted.

Figure 7 presents an outline of the main post mortem biochemical processes with the most used stress indices used to measure the metabolic response to the pre-slaughter and slaughter management.

Figure 7. Tissue stress and quality indicators in fish (Poli et al., 2005)

5.1. Muscle pH

Immediately after the cessation of the circulation of blood and, consequently, of oxygen supply, the stored carbohydrate glycogen is anaerobically degraded and lactic acid accumulates in muscle resulting in a pH drop from a value close to 7.4 to 6.5 in fish, sometimes below. In cod, the pH drops from 6.8 to an ultimate pH of 6.1-6.5. In some species of fish, the final pH may be lower: in large mackerel, the ultimate rigor pH may be as low as 5.8-6.0 and as low as 5.4-5.6 in tuna and halibut, however such low pH levels are unusual in marine teleosts. These pHs are seldom as low as those observed for post mortem mammalian muscle. For example, beef muscle often drops to pH levels of 5.5 in rigor mortis. Since glycogen is converted to lactate via glycolysis under the anaerobic conditions existing after slaughter, the in vivo level of muscle glycogen in fish is the principal determinant of pH in the post-rigor state (Ang & Haard, 1985), which in turn will have major effects on meat quality (Einen & Thomassen, 1998).

In general, fish muscle contains a relatively low level of glycogen compared to mammals, thus far less lactic acid is generated after death. Also, the nutritional status of the fish and the amount of stress and exercise encountered before death will have a dramatic effect on the levels of stored glycogen and consequently on the ultimate post mortem pH. As a rule, well-rested, well- fed fish contain more glycogen than exhausted fish. In a study with Japanese loach (Chiba et al.,

1991), it was shown that only minutes of pre-capture stress resulted in a decrease of 0.5 pH units in 3 hours as compared to non-struggling fish whose pH dropped only 0.1 units in the same time period. In addition, the same authors showed that bleeding of fish significantly affected the post mortem production of lactic acid.

Muscle pH is critically important because both the rate and extent of pH decline greatly affect flesh properties. When a fish is killed, the pH will remain higher, due to the early end of post mortem anaerobic glycolysis caused by the energy source shortage. Starvation before slaughter, if correctly carried out (i.e. for 1–3 days according to water temperature), will not affect this pH development. But during pre-slaughter management, the capture process becomes a stressful situation and therefore fish use more white muscle. Consequently, anaerobic glycolysis occurs, lowering the pH (Poli et al., 2005). Since fish are slaughtered in a short time, the recovery is not possible and muscle pH will remain low. On the contrary, if the traumatic pre-slaughter situation lasts for a longer time (i.e. repeated catching), the lactic acid produced will be gradually cleared from the blood and muscle, but the energy sources will become gradually exhausted. This is even more accentuated if fish are starved for a long time before slaughter.

5.2. Rigor mortis

As mentioned above, glycolysis is the only possible pathway for the production of energy once the heart stops beating. This more inefficient process has principally lactic and pyruvic acids as its end-products. In addition, ATP is produced, but only 2 moles for each mole of glucose oxidized as compared to 36 moles ATP produced for each mole of glucose if the glycolytic end products are oxidized aerobically in the mitochondrion in the living animal.

Once the ATP concentrations start to decrease, a process called rigor mortis occurs. Rigor mortis means the stiffening of the muscles of an animal shortly after death. Immediately after death the muscles of an animal are soft and limp, and can easily be flexed; at this time the flesh is said to be in the pre-rigor condition, and it is possible to make the muscles contract by stimulation. Eventually the muscles begin to stiffen and harden, and the animal is then said to be in rigor. The muscles will no longer contract when stimulated, and they never regain this property. After some hours or days the muscles gradually begin to soften and become limp again. The animal has now passed through rigor, and the muscle is in the post-rigor condition. Sometimes rigor is said to be resolved; this is simply another way of saying that the muscle has passed through rigor to the post-rigor stage.

Rigor in fish usually starts at the tail, and the muscles harden gradually along the body towards the head until the whole fish is quite stiff. The fish remains rigid for a period which can vary from an hour or so to three days, depending on a number of factors, and then the muscles soften again.

Rigor results from a series of complicated chemical changes in the muscle of a fish after death; the process is not yet fully understood, and research is still going on, but it is known that factors like the physical condition of the fish at death, and the temperature at which it is kept after death, can markedly affect the time a fish takes to go into, and pass through, rigor. The time line from start to onset to completion of rigor mortis for carcasses with normal quality and pH decline is illustrated in Fig. 8. When the fish are killed, creatine phosphate (CP) is degraded prior to the breakdown of ATP. When the CP and ATP reach about the same concentration as ATP, this starts to decrease (Watabe et al., 1991). Once the intracellular level of ATP declines from 10 µmoles/g to 1.0 µmoles/g tissue then the muscle starts to get stiff and pre-rigor processing can start with less likelihood of toughening due to shorter than normal sarcomeres. During the decline in ATP levels at the onset of rigor mortis and when the ATP concentration is about 1/3 to 1/2 of the original ATP level, some calcium starts to leak out of the mitochondrion and sarcoplasmic reticulum (SR), mainly due to pH fall and osmotic pressure changes. Muscle contraction per se is controlled by calcium and an enzyme, ATP-ase which is found in every muscle cell. When intracellular Ca+2 levels are 1 µM, Ca+2, activated ATP-ase reduces the amount of free muscle ATP which results in the interaction between the major contractile proteins, actin and myosin, forming the inextensible actomyosin. This ultimately results in the shortening of the muscle, making it stiff and inextensible. The rigor generally begins one to six hours after death in fish; in particular, it is maximal one day and a half after death for sea bass muscle stored at 0 °C. This condition usually lasts for a day and then the resolution of rigor mortis makes the muscle less rigid and no longer elastic.

At the bottom of the above figure, there are three sarcomeres representing the contractile states of the muscle at the beginning, onset, and completion of rigor. The left, pre-rigor sarcomere can still contract and relax, has a limited number of cross bridges, and is very sensitive to some meat processing conditions. The centre sarcomere is shorter indicating that the muscle is shortening just as far as the skeletal restraint will allow. Calcium is leaking out of the SR and stimulates muscle contraction and formation of cross bridges between muscle filaments (Pate & Brokaw, 1980). The sarcomere on the right represents the final contractile state of post-rigor muscle where the number of cross bridges between the filaments is greatest, the distance between muscle protein filaments is the lowest and the meat is the least tender.

Figure 8. Rigor mortis as a function of meat chemistry and time (www.qpc.adm.slu.se)

Some time after death, an opposing process called tenderization begins within hours post mortem and continues during post-mortem storage. Several studies showed that tenderization begins in the early stage of post mortem storage of fish and mammals. Possible causes of post mortem tenderization include a weakening of Z-discs of myofibrils, a degradation of connective tissue or a weakening of myosin-actin junctions (Wang et al., 1998). Tenderization rate and extent vary depending on species and other factors.

The time fish meat takes to go into, and pass through, rigor depends on several factors, among which the most important are the species, the physical condition, the degree of exhaustion before death, size, the amount of handling during rigor and the temperature at which it is kept.

 Species: Some species take longer than others to go into rigor, because of differences in their chemical composition. Whiting, for example, go into rigor very quickly and may be completely stiff one hour after death, whereas redfish stored under the same conditions may take as long as 22 hours to develop full rigor.  Condition: The poorer the physical condition of a fish or the less well-nourished it is before capture, the shorter will be the time it takes to go into rigor; this is because there is very little reserve of energy in the muscle to keep it pliable. Fish that are spent after spawning are an example.

 Degree of exhaustion: In the same way, fish that have struggled in the net for a long time before they are hauled aboard and gutted will have much less reserve of energy than those that entered the net just before hauling, and thus will go into rigor more quickly.  Size: Small fish usually go into rigor faster than large fish of the same species.  Handling: Manipulation of pre-rigor fish does not appear to affect the time of onset of rigor, but manipulation, or flexing, of the fish while in rigor can shorten the time they remain stiff.  Temperature: This is perhaps the most important factor governing the time a fish takes to go into, and pass through, rigor because the temperature at which the fish is kept can be controlled. The warmer the fish, the sooner it will go into rigor and pass through rigor. For example, gutted cod kept at 0 °C may take about 60 hours to pass through rigor, whereas the same fish kept at 30 °C may take less than 2 hours.

To sum up, small fish with low reserves of energy, that is exhausted and in poor condition, and kept at a high temperature will enter and pass through rigor very quickly. On the other hand, large, rested, well-fed fish kept at a low temperature will take a long time to enter and pass through rigor.

5.3. Water holding capacity

Water holding capacity (WHC) of fresh meat (ability to retain inherent water) is an important property of fresh meat as it affects both the yield and the quality of the end product. It can be defined as the ability of the post mortem muscle (meat) to retain water even though external pressures (e.g. gravity, heating) are applied to it.

Muscle contains approximately 75% water. The other main components include protein (approximately 20%), lipids or fat (approximately 5%), carbohydrates (approximately 1%) and vitamins and minerals (often analysed as ash, approximately 1%). The majority of water in muscle is held within the structure of the muscle itself, either within the myofibrils, between the myofibrils themselves and between the myofibrils and the cell membrane (sarcolemma), between muscle cells and between muscle bundles (groups of muscle cells).

Water is a dipolar molecule and as such is attracted to charged species like proteins. In fact, some of the water in muscle cells is very closely bound to protein. By definition, bound water is water that exists in the vicinity of non-aqueous constituents (like proteins) and has reduced

30 mobility, i.e. does not easily move to other compartments. This water is very resistant to freezing and to being driven off by conventional heating. True bound water is a very small fraction of the total water in muscle cells (approximately 0.5 g of water per g of protein). Another fraction of water that can be found in muscles is termed entrapped (also referred to as immobilized) water (Fennema, 1985). This water represents 80% of the total water in fresh meat and is held within the structure of the muscle but is not bound per se to protein and it is most affected by the rigor process and the conversion of muscle to meat. The last fraction is the free water, which is water whose flow from the tissue is unimpeded. Weak surface forces mainly hold this fraction of water in meat.

During the conversion of muscle to meat, lactic acid builds up in the tissue leading to a reduction in pH of the meat. Proteins have an electronic charge that changes as the pH changes. The rule of thumb is that the higher the meat pH, the higher the net negative charge on the proteins and the higher the WHC (Toldra, 2003). All proteins have a characteristic pH where the net electronic charge on the protein is zero (isoelectric point). Once the pH has reached the isoelectric point of the major proteins, especially myosin, the net charge of the protein is zero, meaning the numbers of positive and negative charges on the proteins are essentially equal. These positive and negative groups within the protein are attracted to each other and can result in a reduction in the amount of water that can be attracted and held by that protein. Additionally, since like charges repel, as the net charge of the proteins that make up the myofibril approaches zero (diminished net negative or positive charge) repulsion of structures in the myofibril is reduced allowing those structures to pack more closely together. The end result of this is a reduction of space within the myofibril (Huff-Lonergan & Lonergan, 2005).

Once muscle is harvested the amount of water in meat can change depending on numerous factors related to the tissue itself and how the product is handled. In the immediate pre-slaughter period, stresses on the animal such as fasting, and different stunning methods are likely to influence meat WHC (Cheng & Sun, 2008), provoking a more rapid decline in muscle pH. However, the main factor is the rigor mortis, which in turn will be affected by the stress suffered by the fish before the slaughter process. As muscle goes into rigor, cross-bridges form between the thick and thin filaments within myofibrils, reducing available space for water to reside. Additionally, during rigor development sarcomeres can shorten; this also reduces the space available for water within the myofibril. Also as mentioned previously, the net charge of the protein structures of the myofibril can also influence the amount of space available in the myofibril for water (Huff-Lonergan & Lonergan, 2005).

Water holding capacity is a very important quality attribute which has an influence on product yield, which in turn has economic implications, but is also important in terms of eating quality. Several traits are strongly correlated with high WHC, most of which are directly related to consumer-desired parameters. Among these traits firmer flesh, less drip loss, better processing and cooking, more juiciness and tenderness or better protein functionality are the most important.

5.4. Meat colour

Colour can be easily measured by means of the CIELAB (CIE, 1978), which describes all the colours visible to the human eye (380-760 nm) and was created to serve as a device-independent model to be used as a reference. The three coordinates of CIELAB represent the lightness of the colour (L* = 0 yields black and L* = 100 indicates diffuse white), its position between red/magenta and green (a*, negative values indicate green while positive values indicate magenta) and its position between yellow and blue (b*, negative values indicate blue and positive values indicate yellow).

Figure 9. CIE L*a*b* colour space (from www.linocolor.com)

Since flesh colour is one of the most important quality criteria (Koteng, 1992) and contributes substantially to the elite image of salmonids, pigmentation and flesh colour are important when evaluating the consequences of starvation on slaughter quality (Einen & Thomassen, 1998). However, there is no clear evidence in this area of research. In rainbow trout, no significant decrease in pigment concentration during starvation has been found (Foss et al., 1984; Choubert, 1985). In contrast, Wathne (1995) found significantly decreased CIE L*, a* and b* values in starved Atlantic salmon (pigment content was not measured in that study) and Mørkøre et al. (2008) found that colour was more intense 2 h after slaughter, with a subsequent decrease until 24 h in starved fish and until 72 h in fed fish and a consistently higher colour score (more intense) in starved salmon compared to fed salmon. Robb et al. (2000) report an increase in lightness (L*) in rainbow trout during 72 h of storage after the slaughter and it seems

32 that this increase was more marked in stressed animals. However, this disagrees with Einen & Thomassen (1998), who found no differences in L*, a* and b* between salmon fasted for 3 days and those regularly fed after 4 days of storage.

Acidification of anaerobic glycolysis, which is related to slaughter stress, has been reported as the cause of alteration of the flesh colour, as microstructural changes in muscle fibres dependent on pH may influence the colour perception due to altered refraction through the muscle fibres (Swatland, 2003). The rapid drop in pH post mortem is considered as the main cause for the decreased level of soluble muscle proteins in the flesh of stressed animals compared with the level in unstressed animals. The proteins denature, becoming insoluble and causing a loss of water from the flesh (water holding capacity), and resulting in changes in the reflection of light from the surface, hence changing the colour perception (Warriss, 1996).

6. Objectives

The general aim of the present Thesis is to evaluate the effect of different pre-slaughter handling procedures that have a great importance to a commercial level on welfare and product quality during storage in rainbow trout (Oncorhynchus mykiss). The specific objectives developed in the four trials are:

1. To determine the effect of short-term fasting prior to slaughter, based on degree days, on biometric, haematological and metabolic indicators and product quality development.

2. To determine the best time in the day to slaughter rainbow trout based on the results of the indicators mentioned above.

3. To evaluate the possible effect of water temperature at which fish are held before the slaughter on haematological indicators.

CHAPTER II

MATERIAL AND METHODS

1. Installations and fish material

The trials were carried out at the aquaculture facilities of the School of Forestry Engineering, Polytechnic University of Madrid (Madrid, Spain), which lies on a small slope divided into terraces, with several raceways. The terrace arrangement allows the downward water flow to be distributed among the different raceways by means of channels. For experimental purposes, two raceways of 5.16 m3 volume were filled with freshwater from an underground well, supplying a constant water flow.

Rainbow trout were purchased and transported from two fish farms in Spain to the School of Forestry Engineering: Truchas de la Alcarria S.L. (Valderrebollo, Guadalajara) and Piscifactoría Cien Fuentes (Cifuentes, Guadalajara). The former supplied fish for TRIALS 1 and 2, while the latter for TRIALS 3 and 4. The Truchas de la Alcarria farm provides trout certified as “ecological” (by the Sohiscert® Agency) and sells trout that have not been treated hormonally (both female and male individuals) and they are given a special feed. The Cien Fuentes farm sells commercial trout that are treated hormonally (only females) that have been fed commercial feed with astaxanthin (which gives the trout flesh a salmon colour).

When the fish arrived to the School of Forestry, they were randomly divided into two groups (fasted and controls) and then stocked into two independent raceways for a 1-4 week adaptation period depending on the trial (TRIAL 1: 1 week; TRIAL 2: 2 weeks; TRIAL 3: 4 weeks; TRIAL 4: 4 weeks). In all trials, each raceway was split into three different sections using net separators with the same water volume (1.72 m3), each housing 30 fish (stocking densities in Table 1). In TRIALS 1 and 2, the fish were placed directly into the raceways with the net separators in place, but in TRIALS 3 and 4, net separators were placed 2 weeks after the arrival of the fish. During this period, fish were subjected to a natural photoperiod at a constant water temperature depending on the season of the year (Table 1) and fed the same feed given on the source farm (Ecological feed: 45% crude protein, 24% fat, 11.7% ash and 1.1% crude fibre; feeding rate: 1%; Commercial feed: 42% crude protein, 23% fat, 4.1% ash and 2% crude fibre; astaxanthin: 30 ppm; feeding rate: 1%). We used belt-feeders that provided feed continuously for 24 h.

During the adaptation and experimental periods, water temperature was recorded once every 5 min using underwater temperature sensors in both raceways (Hobo® U-11). With these data we then calculated the average temperature per day (Table 1) and degree days.

Table 1. Average initial parameters and analyses in the four trials 1 2 3 4 Dates February 2011 October 2011 March 2012 May 2012 Sunrise/sunset times 08h00/19h58 08h20/19h42 07h40/19h12 06h52/21h31 Fish initial weight (g) 306.7±38.0 214.9±23.2 417.7±61.3 368.8±63.7 Stocking density (kg/m-3) 5.35 3.75 7.29 6.43 Water temperature (ºC) 11.75±1.60 19.15±0.59 11.12±0.57 22.65±0.94 Degree days (ºC d)1 11.8-23.6-35.3 19.2-38.3-57.5 11.1-22.2-33.4 22.7-45.3-68.0 Data are presented as mean ± s.d. 1Indicate mean degree days during the first, second and third day of fasting.

2. Experimental design

Once the two raceways were divided into three sections, each of them was assigned to one hour of slaughter: 08h00 (morning), 14h00 (afternoon) and 20h00 (night). The former group was located upstream and the latter downstream to avoid feed pellets passing from one group to another, according to the direction of water flow (see Figure 12). Only the fish from the first raceway were fasted the day before the trial started. Within each section, fish were fasted for 1, 2 or 3 days depending on the day they were slaughtered (first, second or third day of the trial, three consecutive days). Thus, there were nine sampling groups in each raceway (3 sections x 3 hours of slaughter), with 10 fish each (Figure 12). Fish from the control group were divided likewise and slaughtered at the same times as the fasted fish but fed every day during the experiments.

Figure 11. Photo of the raceway holding treatment fish (fasted) with three sections. Water entered through the main channel at the top of the photo and exited through a drain at the bottom. Note that the sections were covered with a net (not present in photo) to avoid fish jumping from one section to another.

Figure 12. Setup of the raceways and sections at the School of Forestry Engineering. Each day (1, 2 and 3) represents a batch of 10 fish

3. Slaughtering and sampling procedures

At the end of each fasting period, ten treatment fish and ten control fish were alternatively captured and immediately anesthetized in clove oil for 2 min (60 mg/l). Care was taken to avoid unnecessary stress during capture to minimize possible sampling effects. They were then weighed and measured individually and finally sampled to evaluate haematological variables. Blood samples were withdrawn from the caudal vein with 2 ml syringes (BD Plastipak), 1 ml were centrifuged (Biofuge Heraeus, Thermo Electron Corporation) at 1500xg for 12 min (4 ºC) and plasma was extracted to measure cortisol, glucose and lactate. The other 1 ml of blood was used for haematocrit and leucocyte count and immediately stored at 4 ºC until analysis. Just after blood collection, fish were killed by sectioning the spinal cord at the base of the head. After the blood sampling, half of the fish (fasted and control) were gutted and either the whole gut (TRIALS 1 and 2) or gut and stomach contents and liver were weighed separately (TRIALS 3 and 4). Thereafter trout were filleted by hand, placing the fillets on smooth solid trays, covered with aluminium foil and kept in refrigeration (under 4 ºC). The right fillets were used for pH and colour measurements and the left fillets for water holding capacity (WHC) analyses. The remaining fish from each group were used for recording progression of rigor mortis. All the work described has been carried out in accordance with the EU Directive 2010/63/EU for animal experiments and approved by the Animal Ethics Committee of the Polytechnic University of Madrid (Spain), in compliance with the Spanish guidelines for the care and use of animals in research (BOE, 2005). The experimental protocol followed during the trials is in agreement with the recommendations described by EFSA (2009) on the pathways for the pre- slaughter and slaughter in rainbow trout (Pathway 13).

3.1. Biometric indicators

With the data on body weight, body length, weight of the whole gut, weight of the gut and stomach content (gut content) and weight of the liver we calculated the following parameters:

3.2. Haematological indicators

As previously described, blood samples were collected in two 1.5 ml centrifuge tubes. The first was filled with potassium fluoride (FK) and used for plasma analysis (cortisol, glucose and lactate). The second tube was filled with ethylenediaminetetraacetic acid (EDTA) and used for haematocrit and leucocyte count analyses.

3.2.1. Plasma cortisol

Plasma cortisol was assessed by the enzyme-linked immunosorbent assay technique (ELISA). First, 10 µl of each sample were pipetted into the corresponding wells, (coated with anti-cortisol antiserum of rabbit). 200 µl of enzyme conjugate were then added into all wells, except for the blank well. These wells were incubated for 60 min at 37 ºC. During this first incubation, the sample cortisol competes with the cortisol conjugated to horseradish peroxidase (HRPO) for the specific sites of the antiserum coated on the wells. Following the incubation, all unbound material is removed by aspiration and washed four times with 300 µl of diluted washing solution (phosphate buffered saline, PBS). The enzyme activity which is bound to the solid phase is inversely proportional to the cortisol concentrations in calibrators and samples, which is evidenced by incubating the wells with 200 µl of a Chromogen solution (Tetramethylbenzidine, TMB) in a substrate buffer (citrate-phosphate) for 15 min at 37 ºC.

Finally, 100 µl of a blocking reagent (sulphuric acid, H2SO4) were pipetted into the wells to stop the reaction. Calibrators were prepared with vials of cortisol in PBS and BSA and lyophilized at the concentrations of 0, 10, 30, 100, 300 and 900 ng/ml. Colorimetric reading of samples, blank and calibrators was performed within 20 min from the end of the assay by using a spectrophotometer (Hitachi 717®) at 405 (for concentrations below 30 ng/ml) and 450 nm (for concentrations between 30 and 900 ng/ml).

3.2.2. Plasma glucose

Plasma glucose was determined by the enzymatic colorimetric method (GOD/PAP). Glucose oxidase (GOD) catalyses the oxidation of glucose to gluconic acid. The formed hydrogen peroxide (H2O2) is detected by a chromogenic oxygen acceptor, phenol 4-aminophenazone (4- AP) in the presence of peroxidase (POD):

→

The intensity of the colour formed is proportional to the glucose concentration in the sample.10 µl of the plasma sample and 1 ml of a reagent solution with phenol (0.3 mmol/l), GOD (15,000 U/l), POD (1000 U/l) and 4-AP (2.6 mmol/l) were mixed. A standard solution of glucose (100 mg/dl) was also prepared and mixed with 1 ml of the reagent solution. Finally, 1 ml of the reagent solution was taken as the blank. All the solutions were pipetted into a cuvette and incubated for 20 min at room temperature (15-25 ºC). Absorbance of samples and standard was read against the blank using a spectrophotometer (Hitachi 717®) at 505 nm.

The glucose concentration (mg/dl) is calculated as:

3.2.3. Plasma lactate

Plasma lactate was determined by the enzymatic colorimetric method (LO-POD). The principle of this technique is based on the oxidation of lactate by lactate oxidase (LO) to pyruvate and hydrogen peroxide (H2O2), which under the influence of peroxidase (POD), 4-aminophenazone (4-AP) and 4-chlorophenol form a red quinone compound:

→

10 µL of the plasma samples were added to 1 ml of reagent solution (800 U/L of LO, 2000 U/l of POD and 0.4 mmol/l of 4-AP dissolved in 10 ml of 4 mmol/l of 4-chlorophenol and 50 mmol/l of a buffer solution). 10 µl of a standard solution (lactate aequous, 10 mg/dl) was also mixed with 1 ml of the reagent solution. Finally, 1 ml of the reagent solution was taken as the blank. All the solutions were pipetted into cuvettes and incubated for 10 min at room temperature (15-25 ºC). Absorbance of samples and standards was read against the blank using a spectrophotometer (Hitachi 717®) at 505 nm. The lactate concentration (mmol/l) was calculated as:

3.2.4. Red and white cells

Haematocrit was automatically assessed using a Coulter STKR® analyser.

For the leucocyte count analysis, the blood sample was diluted in a saline phosphate buffer (1/40), adding 0.2 ml of methylene blue. Subsequently, the erythrocytes were lysed and the leucocyte nucleus was stained. Cells were counted in a Neubauer chamber, which is immediately filled after mixing. After 2 minutes, one may begin counting the leucocytes in the 4 large squares (1x1 mm).

The leucocyte count (103/µl) is calculated as:

3.3. Metabolic indicators

3.3.1. Liver and muscle glycogen

Liver and muscle glycogen analyses were performed on samples taken at 14h00 on days 1, 2 and 3 only in TRIAL 3. The concentrations of glycogen in liver and muscle were assessed according to the technique described by Dreiling et al. (1987). This assay assumes that iodine binds to glycogen to form a coloured complex. The inorganic salts increase the stability and sensitivity of iodine binding to branched polysaccharides and the intensity and hue of the glycogen-iodine complex reflect the average binding of iodine to glycogen.

Two muscle and liver samples from each gutted fish were collected from the fillets immediately after the slaughter, frozen in liquid nitrogen and kept until analysis. Approximately 2 g of liver and muscle were minced and then suspended in 10 ml of cold perchloric acid (CPA, 8.4%), thoroughly homogenized for 30-45 s with a Ultraturrax® homogenizer (13000 rpm) and subsequently centrifuged at 4500 rpm for 10 min (4 ºC). The supernatants were decanted and diluted depending on the type of sample: 0.2 ml of supernatant and 4.6 ml of CPA and 0.5 ml of supernatant and 1.5 ml of CPA for the liver and 0.4 ml (direct) and 0.5 ml and 1.5 ml of CPA for muscle. Iodine colour reagent was prepared by combining 1.3 ml of a solution containing 0.26 g of iodine and 2.6 g of potassium iodide (in 10 ml distilled water) with 100 ml saturated calcium chloride (CaCl2). The reagent (2.6 ml) was added to glycogen standards or tissue extracts (0.4 ml). Glycogen standards were prepared using purified oyster glycogen (Type II, Sigma-Aldrich Chemicals) and diluting them six times (160, 80, 40, 20, 10 and 5 µg/ml). After the colour reagent addition, colour developed immediately and was stable for up to 2 h. Samples were then stabilized for 20 min in complete darkness and spectrophotometrically analysed at 460 nm, adding CPA (8.4%) to the colour reagent as the blank.

3.3.2. R-value (IMP/ATP ratio)

The IMP/ATP ratio was assessed using the technique described by Honikel & Fischer (1977). Two muscle samples from each gutted fish were collected (TRIAL 3) from the fillets immediately after the slaughter, frozen in liquid nitrogen and kept until analysis. Approximately 2 g of the sample were minced and thoroughly homogenized with 15 ml of CPA (0.85 M) in a Ultraturrax® homogenizer (13000 rpm) for 30-45 s. they were then centrifuged for 10 min at 4500 rpm (4 ºC). The supernatant was filtered, collecting 0.2 ml and diluting them into 3.8 ml phosphate buffer pH 6.5. We used the phosphate buffer as a blank. Finally, the absorbance was

43 measured at 248, 250, 258 and 260 nm. The R-value was calculated using the absorbance at 250/260 nm.

3.4. Product quality

Muscle pH and rigor mortis measurements were performed at 0, 2, 9, 24, 48 and 72 h post mortem in all trials (except in TRIAL 4 with no measurements at 72 h). Percentage of water released was measured at 0, 24, 48 and 72 h in all trials (except in TRIAL 4 with no measurements at 72 h). Colour measurements were performed at 0, 24, 48 and 72 h post mortem in TRIAL 2 and 0, 24 and 48 h post mortem in TRIAL 3.

3.4.1. Muscle pH

Once the fish were filleted, the right fillet was used to measure pH by pulling an electrode into the cut surface of the muscle using a pH-meter (HANNA, mod. HI 9125) connected to a temperature probe. Before the measurements, the pH-meter was standardized at pH 4.01 and 7.00.

Figure 13. Muscle pH sampling.

3.4.2. Rigor mortis

The recording progression of rigor mortis was measured following the Cuttinger’s method (Korhonen et al., 1990). Measurements were performed by placing the trout on a solid flat

44 surface so that the body part behind the posterior end of the dorsal fin was hanging over the edge, unsupported. The rigor angle was calculated as:

where X is the length (cm) of the horizontal leg of the right-angled triangle and Y is the length (cm) of the vertical leg of the right-angled triangle.

Figure 14. Rigor mortis sampling. The length of the horizontal (X) and vertical (Y) legs of the right- angled triangle and the rigor angle (α) are represented.

3.4.3. Water holding capacity

Water holding capacity was determined as the amount of water lost by the meat sample (PWR or percentage of water released) using the filter paper press method (Grau & Hamm, 1957). Each minced muscle sample (5 g) was placed between two filter papers and pressed with a 2.25 kg load for five minutes. The amount of water lost by was calculated as:

where M0 is the muscle weight (g) before pressing and M1 is the muscle weight (g) after pressing.

3.4.4. Meat colour

Colour measurements were performed using a Minolta Spectrophotometer CM-2500c (Minolta, Osaka, Japan). The CIE 1976 L*a*b* system recommended by the International Commission Illumination (CIE, 1978) was chosen as the colour scale. In this system, L* represents the lightness of the colour (L* = 0 yields black and L* = 100 indicates diffuse white), a* denotes its position between red/magenta and green (a*, negative values indicate green while positive values indicate magenta) and b* indicates its position between yellow and blue (b*, negative values indicate blue and positive values indicate yellow). Two colour measurements were performed on the muscle dorsal area of each right fillet, just behind the dorsal fin, as described by Pavlidis et al. (2006).

Samplings and analyses carried out in the different trials are summarized in Table 2.

Table 2. Summary of measurements and analyses carried out in the four trials 1 2 3 4 Body weight & length / CC x x x x Whole gut weight / RWWG x Gut content weight / RWGC x x Liver weight / HSI x x Metabolic indicators x Haematological indicators x x x x Muscle pH, rigor mortis and PWR x x x x Meat colour x x

4. Statistics

Mean ± standard error of the means (SEM) of each sampling group on the parameters analysed were calculated using the SAS software ver. 9.0 (Statistical Analysis System Institute Inc., Cary, NC, USA). The data were tested for normality (Shapiro-Wilks) and homogeneity (Levene).

4.1. Biometric and haematological indicators

An analysis of variance using the MIXED procedure of SAS was run to compare biometric and haematological indicators both in fasted and fed trout. Prior to analysis, all the fish samples

46 were classified as fasted or not fasted and this new nominal variable (type of fasting) was nested to the day of sampling. The nested ANOVA was performed with the day of sampling, hour of slaughter and type of fasting nested to day of sampling (hereafter, type of fasting (day of slaughter)) as the main factors, according to the following model:

[1]

Where: yijkl = dependent variables (body measurements and haematological indicators).

µ0 = population mean.

Hi = hour of slaughter (08h00, 14h00 and 20h00).

Dj = day of slaughter (1st, 2nd and 3rd).

T(D)k(j) = type of fasting nested to day of slaughter (fasted 1, 2 or 3 days and fed and slaughtered on days 1, 2 and 3).

HDij = interaction hour and day of slaughter.

HT(D)ik(j )= interaction hour of slaughter and type of fasting nested to day of slaughter.

εl(ijk) = residual error

The coefficient of condition was initially included in the model as a covariable to determine the effect of this parameter on the haematological indicators but its effect was not significant and it was therefore removed from the model.

To observe the effect of water temperature on haematic indicators, an analysis of variance was conducted as in model [1] including the average water temperature of TRIALS 1 and 3 (11.4 ºC) and TRIALS 2 and 4 (20.9 ºC) as a fixed effect and the fish sample as a random effect.

[2] Where:

Ml = fish sample (random effect)

TEm = water temperature (11.4 ºC and 20.9 ºC)

HTEim = interaction hour of slaughter and water temperature

DTEjm = interaction day of slaughter and water temperature

TET(D)mk(j) = interaction water temperature and type of fasting nested to day of slaughter

Means ± SEM were compared using the LSD test with 5% as the level of significance (P<0.05) for the main effects as well as for the interactions that were significant within the model. Means

47 were also compared between fasted and fed groups at every sampling point (one-way ANOVA, level of significance: 5%) when the interaction hour of slaughter*type of fasting (day of slaughter) or water temperature*type of fasting (day of slaughter) were significant.

4.2. Metabolic indicators

Data on liver and muscle glycogen and the R-value (IMP/ATP ratio) were analysed by performing a nested analysis of variance (as in model [1]) with the day of slaughter (1st, 2nd and 3rd) and the type of fasting nested to the day of slaughter (fasted for 1, 2 or 3 days and fed and slaughtered on day 1, 2 or 3) as the main factors.

[3]

4.3. Product quality

Mean ± SEM of each sampling group on muscle pH, rigor mortis, PWR and colour were subjected to repeated measures analysis using the MIXED procedure of SAS. In order to simplify the analysis, the three control groups (slaughtered on days 1, 2 and 3) were combined into one single group (0 days of fasting), as no significant differences among the three groups were found (P>0.05) for any of the meat indicators. The meat sample was considered as the subject, hour of slaughter and days of fasting were used as the main factors and hours post mortem as the within-subject factor:

[4]

Where: yijkl = dependent variables (meat indicators)

µ0 = population mean

Hi = hour of slaughter (08h00, 14h00 and 20h00)

Dj = days of fasting (0, 1, 2 and 3)

Tk = hours post mortem (see details above for each variable)

HiDj = interaction hour of slaughter and days of fasting

HTik = interaction hour of slaughter and hours post mortem

DTjk = interaction days of fasting and hours post mortem

HDTijk = interaction hour of slaughter, days of fasting and hours post mortem

εl(ijk) = residual error

Means were compared using the LSD test with 5% as the level of significance (P<0.05) for both the main effects as well as for their interactions. When the triple interaction was significant, differences among days of fasting for every hour of slaughter were displayed. In this case, the four durations of fasting (0, 1, 2 and 3 days) were compared at every hour post mortem (one- way ANOVA, level of significance: 5%).

CHAPTER III

RESULTS

1. TRIAL 1

1.1. Biometric indicators

Mean body weight of fasted trout at slaughter was 303.2 g and 310.6 g for control fish (SEM=4.0). No differences were found in any of the factors studied. Mean body weight at 08h00 was 307.2 g, at 14h00 was 308.6 and at 20h00 was 304.8 g (SEM=4.9). Body length averaged 32.2 cm and 32.1 cm (SEM=0.1) in fasted and control trout, respectively. We found a significant effect of the day of slaughter (Table 3), with fish slaughtered on days 2 and 3 having a higher length (32.2 and 32.5 cm, respectively) than those slaughtered on day 1 (31.7 cm) (SEM=0.2). With regard to the coefficient of condition, it was 0.91 cm in fasted trout and 0.94 cm in control trout (SEM=0.1). There was a significant effect of the day of slaughter (Table 3), with animals killed on 1 day showing the highest CC (0.99) and trout slaughtered on day 3 the lowest (0.88), with the group slaughtered on day 2 having intermediate values (0.91) (SEM=0.01).

Table 3. Summary of P-values of biometric indicators in TRIAL 1 for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Body weight NS NS NS NS NS Body length NS 0.010 NS NS NS CC NS <0.001 NS NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

1.2. Haematological indicators

Haematological indicators were markedly influenced by at least one of the factors included in the statistical model used. P-values are presented in Table 4.

Plasma cortisol was found to be significantly different among hours of slaughter: it was higher at 08h00 (121 ng/ml) and 14h00 (130 ng/ml) compared to 20h00 (92 ng/ml) (SEM=6). No effect of other factors included in the statistical model was observed. Mean concentrations in fasted and fed trout were 115 and 113 ng/ml, respectively (SEM=5).

Plasma glucose showed variation among types of fasting (days of slaughter), with fish fed regularly showing the highest concentrations on days 2 and 3, and these in turn were significantly higher than trout fasted on those days (by 14% on average). However,

53 concentrations on day 1 were similar (Fig. 15). No differences were found among different hours of slaughter: mean concentrations at 08h00, 14h00 and 20h00 were 98, 92 and 94 mg/100 ml, respectively (SEM=2).

120 a 110 a

100 b b b b 90

80 Fasted Fed 70

Plasma Plasma glucose (mg/100 ml) 60

50 1 2 3 Day of slaughter

Figure 15. Changes in plasma glucose concentrations of rainbow trout during three days of fasting in TRIAL 1. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

All the factors included in the model had a significant effect on plasma lactate concentration (P<0.05). To better observe the differences among groups, the interaction hour of slaughter and type of fasting (day of slaughter) is displayed in Fig. 16, where a dramatic increase in lactate of fasted fish can be seen at 20h00 on day 1 (40 mmol/l) and at 08h00 on day 3 (65 mmol/l). It was also reported an increase in lactate of fed fish at 08h00 on day 2 (53 mmol/l). Differences between fasted and fed fish are also presented (one-way ANOVA).

80 a 70 60 a

50 a 40 30

20 b b b Plasma Plasma lactate(mmol/l) Fasted 10 Fed 0

Figure 16. Changes in plasma lactate concentrations of rainbow trout during three days of fasting in TRIAL 1. Data are means ± SEM (P=0.004; n=10). Different letters indicate significant differences between fasted and fed fish at sampling points (P<0.05).

Haematocrit was significantly lower at 20h00 (averaging 33.0%) compared to 08h00 (35.5%) and 14h00 (34.5%) (SEM=1.2). No other factor significantly affected this variable in the analysis. Mean values of fasted and fed fish were 35.2% and 35.3%, respectively (SEM=0.4).

Finally, leucocytes were higher at 14h00 in fasted and fed fish (6.3x103/µl) in comparison to 08h00 (5.0 x103/µl), with fish slaughtered at 20h00 having intermediate values (5.8 x103/µl) (SEM=1.2).

Table 4. Summary of P-values of haematological variables in TRIAL 1 for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Cortisol <0.001 NS NS NS NS Glucose NS NS <0.001 NS NS Lactate 0.042 <0.001 0.008 <0.001 0.004 Haematocrit 0.002 NS NS NS NS Leucocytes <0.001 NS NS NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

1.3. Product quality

The interaction hour of slaughter*day of slaughter*hours post mortem was significant (P<0.001) in muscle pH, rigor mortis and PWR, thus this interaction has been depicted in figures to better interpret changes in the evolution of those indicators.

1.3.1. Muscle pH

Overall, muscle pH showed very little variation between 0 and 2 h following slaughter, with a clear tendency to decrease until 24 h post mortem (ranging from 6.03 to 6.40) and then remained stable or slightly increased.

Initial muscle pH was similar among days of fasting with animals slaughtered at 08h00, ranging from 6.78 to 6.94. However, differences were observed at 2 h post mortem (P<0.05), with 0 and 1 days of fasting having the highest values (6.72 and 6.63, respectively), 2 days with intermediate values (average of 6.59) and 3 days showing the lowest pH (6.59). Subsequent sampling points did not show any significant difference, reaching a minimum after 24 h of slaughter and then levelling off until the last sampling point (+72 h).

Initial muscle pH was again similar among groups at 14h00 but differences were observed from 2 h to 24 h post mortem. At 2 h after slaughter, the group subjected to 3 days of fasting had the highest values (6.76), with controls and 1 day having intermediate values and the group fasted for 2 days with the lowest (6.44). At 9 h post mortem there was a clear reduction of muscle pH with respect to 0 h and 2 h, with the group of 3 days of fasting having the highest pH again (6.48). At 24 h post mortem, trout subjected to 3 days of fasting showed a significantly higher pH (6.40) compared to the other three groups (average of 6.15). No differences were found at 48 h or 72 h (see Fig. 17b).

7.40 (a) 0 days 7.20 1 day 7.00 2 days 3 days 6.80 a a

6.60 ab Muscle pH Muscle 6.40 b 6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 (b) 0 days 7.20 1 day 7.00 2 days 3 days 6.80 a 6.60 ab

MusclepH ab a a 6.40 b b b b 6.20 b b 6.00 b 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 (c) 0 days 7.20 1 day

7.00 2 days a ab 3 days 6.80 b a 6.60 a c a MusclepH a a 6.40 a a a a b 6.20 b b b ab b b 6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 17. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 1. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=0.07). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

Finally, at 20h00, muscle pH was the lowest for trout fasted 1 day at 0, 2, 9 and 24 h post mortem (see Fig. 17c). At 0 h, controls and 2 days of fasting were higher (6.95 and 6.90, respectively) than 3 days (6.79) and 1 day (6.61). After 2 h and 9 h, the first three groups had higher values than 1 day (averaging 6.67 vs. 6.26 at +2h and 6.50 vs. 6.28 at +9h). During the period 2-9 h, the pH decreased in all fish groups except in 1 day of fasting, and this reduction was more pronounced in controls and 3 days of fasting. At 24 h post mortem, muscle pH reached the lowest levels (except in controls, see Fig. 17c), with significant differences among controls and trout fasted 3 days (6.25 and 6.28, respectively) and 1 day (6.02). After 48 h, trout fasted for 3 days had a significantly higher muscle pH (6.39) than the other three groups (6.21, 6.08 and 6.14 for 0, 1 and 2 days, respectively). No significant differences were found 72 h post mortem (excluding the group fasted for 3 days since we had no data).

1.3.2. Rigor mortis

Rigor stiffening (expressed as the rigor angle) showed a similar development in the four groups, starting about 2 h post mortem (average of 32.1º for all the groups) and reaching maximum stiffness (full rigor) after 24 h (averaging 67.6º).

Trout slaughtered at 08h00 did not show significant differences at 0 or 2 h post mortem, although at 9 h the rigor angle was higher in the 3 day-fasted fish (55.8º) in comparison to those fasted for 0 and 2 days (45.6º and 43.2º) and these in turn higher than fish fasted 1 day (35.5º). After 24 h, the group fasted 2 days had significantly higher stiffening (72.9º) compared to controls (61.8º) and fish fasted 1 day (52.0º). After 48 h, differences were observed between controls and 2 days of fasting and 1 and 3 days, with the first two groups having highest rigor angles. All groups had similar values at the last sampling point.

However, the effect of fasting was less marked at 14h00, where differences were only found at 9 h post mortem, with the group fasted 2 days showing the lowest rigor angle (39.9º vs. 53.9º, 59.6º and 54.1º in 0, 1 and 2 days of fasting, respectively).

Finally, at 20h00, differences were not observed among treatments from 0 to 9 h post mortem, while controls and the 1 day group were highest at the moment of maximum stiffness (75.7º and 69.9º, respectively) and the group of fish fasted 3 days having the lowest (56.9º; see Fig. 18b). At 48 h, controls and fish fasted 2 days had a significantly greater angle compared to 1 and 3 days (62.7º and 62.8º vs. 45.6º and 51.0º), whereas the same situation was observed after 72 h (51.3º in controls and 55.1º in 2 days of fasting vs. 41.4º in 1 day of fasting, respectively; no data for the 3 days fasting group).

90.0 (a) 0 days 80.0 a 1 day 70.0 ab 2 days 60.0 a b a 3 days 50.0 b c b a 40.0 b c 30.0 Rigorangle (º) b 20.0 10.0 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (b) 0 days 80.0 1 day 70.0 a 2 days 60.0 a 3 days a 50.0 40.0 b

Rigorangle (º) 30.0 20.0 10.0 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (c) a 80.0 70.0 a a 60.0 a a 50.0 b b a 40.0 b b 0 days

Rigor angleRigor(ª) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 18. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 1. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=4.4 in 08h00 and 3.86 in 14h00 and 20h00). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

1.3.3. Percentage of water released

In general terms, no specific pattern regarding PWR could be discerned, since variability among groups was very high. In most cases, the maximum PWR was reached at 24 or 48 h post mortem.

At 08h00, differences were found at every sampling point. Initial PWR was higher in 1 day group (14.38% vs. 7.70%, 7.85% and 6.59% in controls, 2 days and 3 day, respectively). However, at 24 h post mortem, fish fasted 3 days showed the highest PWR (23.18%), reaching

58 the maximum for the whole period, in comparison to 1 and 2 days (15.91% and 16.34%), with controls having intermediate values (average of 18.91%). Trout subjected to 2 days of fasting had the highest value at 48 h after slaughter (32.67%, which is the peak for this group). Finally, after 72 h, trout fasted for 1 day and those fed normally had the highest PWR (27.33% and 21.31%; see Fig. 19a). These two groups reached the maximum PWR at this point.

35.00 0 days a (a) 30.00 1 day a 2 days 25.00 a 3 days b ab 20.00 ab a b 15.00 b

PWR (%) b b b b 10.00 b b 5.00 b

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 0 days (b) 30.00 1 day 2 days a 25.00 a 3 days a a 20.00 ab a ab ab

15.00 ab PWR (%) b 10.00 b b 5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 0 days (c) 30.00 1 day a 2 days 25.00 a a 3 days ab 20.00 a ab

15.00 b PWR (%) 10.00 b

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 19. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 1. Data represent means (n=3 in fasted groups and n=9 in controls; SEM=2.55). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

Initial PWR values were not different among groups at 14h00, ranging from 8.94% to 13.18%. However, all subsequent values were lower for fish fasted 3 days compared with those fasted 2 days. Maximum values were observed 24 h after slaughter except in the group fasted 2 days. Controls and the groups fasted 1 day showed intermediate values (except at 24 h, where they were both different from the 3 days group).

At 20h00, no differences were found for initial PWR either, but 24 h and 48 h post mortem fish fasted 1 or 2 days showed the highest levels compared to the other two groups. After 24 h, the group fasted 3 days exhibited the lowest PWR (12.90%), with controls having intermediate values. After 48 h, controls had the lowest level (17.79%). Finally, no differences were found 72 h after slaughter.

2. TRIAL 2

2.1. Biometric indicators

No differences were found on body weight for any of the factors included in the model (Table 5). Mean values of fasted trout at slaughter were 210.7 g and 219.3 g (SEM=2.4) for control fish, while fish weighed 214.2, 214.9 and 215.9 g (SEM=2.9) at 08h00, 14h00 and 20h00, respectively. Regarding body length, it averaged 29.3±0.1 cm (±SEM) both in fasted and control trout. The coefficient of condition was found to be significantly affected by the type of fasting (day of slaughter) (Fig. 20), with higher values in control fish on days 2 and 3 (by 5% on average).

1.00 0.95 a b a 0.90 b b b 0.85 0.80 0.75 CC Fasted 0.70 Fed 0.65 0.60 0.55 0.50 1 2 3 Day of slaughter

Figure 20. Changes in the coefficient of condition (CC) of rainbow trout during three days of fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

Finally, RWWG followed the same trend, being significantly different among types of fasting (days of slaughter) (Fig. 21). In this case, all the control groups showed a higher RWWG compared to fasted trout (averaging 11.6% vs. 8.9%, respectively).

16.0

14.0 a a a 12.0 b 10.0 b b

8.0 Fasted 6.0 RWWG RWWG (%) Fed 4.0

2.0

0.0 1 2 3 Day of slaughter

Figure 21. Changes in the relative weight of the whole gut (RWWG) of rainbow trout during three days of fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

Table 5. Summary of P-values of biometric indicators in TRIAL 2 (three days of fasting) for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Body weight NS NS NS NS NS Body length NS NS NS NS NS CC NS NS <0.001 NS NS RWWG NS NS <0.001 NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

2.2. Haematological indicators

Plasma cortisol was markedly influenced by the hour of slaughter, with no other significant relationships in the model described (P>0.05). Fish slaughtered at 08h00 showed the highest cortisol levels, averaging 89 ng/ml, whereas at 14h00 it was 42 ng/ml and at 20h00, 55 ng/ml (SEM=7). Cortisol concentrations were similar between fasted (63 ng/ml) and control fish (61 ng/ml) (SEM=6).

However, plasma glucose was only influenced by the type of fasting (day of slaughter), with differences between fasted and fed fish only on day 1 (Fig. 22). Subsequently, glucose levels significantly decreased on day 2 and slightly increased again on day 3 (with intermediate values

61 between days 1 and 2) in fed trout. In fasted fish, levels significantly increased on day 2 and decreased on day 3, reaching levels similar to initial concentrations. In this case, differences between the three hours of slaughter were not significant (70, 67 and 70 mg/100 ml in fish slaughtered at 08h00, 14h00 and 20h00, respectively; SEM=6).

90 a ab 80 bc abc 70 c c 60 50 40 Fasted 30 Fed 20

Plasma Plasma glucose (mg/100 ml) 10 0 1 2 3 Day of slaughter

Figure 22. Changes in plasma glucose concentrations of rainbow trout during three days of fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P=0.010).

Regarding plasma lactate, only the interaction between hour and day of slaughter was significant (Table 6), although differences were small. Mean concentration at 20h00 on day 1 (20 mmol/l) was significantly higher compared to the rest of groups, with an average concentration of 16 mmol/l (SEM=1). Mean concentration in fasted and fed fish were 16 and 17 mmol/l (SEM=1).

Haematocrit was found to be significantly influenced by the interaction hour and day of slaughter, with the highest value at 20h00 on day 1. In fish slaughtered at 08h00 there was a slight rise on day 3 with respect to days 1 and 2 (by 13% on average) but in fish slaughtered at 14h00, levels remained quite stable (averaging 30.4%). Finally, in fish killed at 20h00, a small reduction was observed, with the following sequence: day 1>day 2>day 3. Mean levels in fasted and fed fish were 29.6% and 30.9%, respectively (SEM=0.5).

40.0 a 35.0 b bc b bc bcd 30.0 cd cd d 25.0

20.0 08h00

15.0 14h00

Haematocrit (%) 20h00 10.0

5.0

0.0 1 2 3 Day of slaughter

Figure 23. Changes in haematocrit levels of rainbow trout during three days of fasting in TRIAL 2. Data are means ± SEM (n=20). Different letters indicate significant differences among groups (P<0.001).

Finally, the leucocyte count was markedly affected by the type of fasting (day of slaughter). As seen in Fig. 24, levels in fasted fish remained stable over the experimental period (averaging 4.6x103/µl) but there was an important increase in fish fed regularly from day 1 compared to day 3 (approximately 24%), with the group slaughtered on day 2 showing intermediate values. No differences were found regarding the hour of slaughter: mean level at 08h00 was 5.1x103/µl, at 14h00 4.8x103/µl and at 20h00 4.6x103/µl (SEM=0.2).

7.0 a 6.0 b ab b b

5.0 b

/µl) 3 4.0

3.0 Fasted Fed

2.0 Leucocytes(x10 1.0

0.0 1 2 3 Day of slaughter

Figure 24. Changes in leucocytes levels of rainbow trout during three days of fasting in TRIAL 2. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P=0.023).

Table 6. Summary of P-values of haematological variables in TRIAL 2 (three days of fasting) for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Cortisol <0.001 NS NS NS NS Glucose NS NS 0.010 NS NS Lactate NS NS NS 0.007 NS Haematocrit NS NS NS <0.001 NS Leucocytes NS NS 0.023 NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

2.3. Product quality

The interaction hour of slaughter*day of slaughter*hours post mortem was significant (P<0.001) in muscle pH, rigor mortis and PWR, thus this interaction was displayed in figures to represent the changes in the evolution of these indicators. Regarding colour measurements, only the hours post mortem was significant within the model (see model [4] in Chapter II).

2.3.1. Muscle pH

Overall, fasting significantly influenced muscle pH development in the period from 24 h to 72 h after slaughter, showing no differences until after 9 h post mortem. Minimum pH occurred between 9 and 48 h post mortem depending on the treatments, while maximum pH was reached at the moment of slaughter or 2 h later.

Fish slaughtered at 08h00 did not show any significant difference among groups between 0 and 9 h post mortem. During the period from 0 to 2 h after slaughter, muscle pH remained quite stable, except in the group fasted 1 day, which sharply increased from 6.80 to 7.15. From to +2 h to +9 h, pH dropped in all the treatments (by 0.27 units on average). After 24 h, trout subjected to 3 days of fasting showed a significantly higher pH (6.80) with respect to controls (6.43), with fish fasted for 1 and 2 days with intermediate levels (6.64 in both groups). From +24 h to +48 h, the group fasted for 2 days increased muscle pH, having the highest level at that point, along with the group of 3 days of fasting (6.69 and 6.70, respectively), whereas controls showed the lowest pH (6.43). Lastly, at +72 h the situation was similar to that in the +24 h sampling point, with fish fasted 3 days with the highest pH (6.83) and those fed normally with the lowest (6.54; see Fig. 25a).

7.40 0 days (a) 7.20 1 day 2 days 7.00 a 3 days a 6.80 a a ab ab 6.60 ab ab

MusclepH ab 6.40 b b b 6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 0 days (b) 7.20 1 day 2 days 7.00 a 3 days a ab 6.80 b a ab ab b 6.60 a MusclepH b 6.40 b b 6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 0 days (c) 7.20 1 day 2 days a 7.00 a a 3 days a 6.80 b 6.60

MusclepH b b b 6.40

6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 25. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 2. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=0.09). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

There were no significant differences among groups between 0 and 9 h post mortem at 14h00. Muscle pH slightly decreased in all the groups with the exception of the group of 1 day (rising from 6.96 to 7.04). However, from +2 h to +9 h sampling, pH was reduced in all the groups (averaging 0.25 units). During the period from +9 h to +24 h, fish fasted 3 days were the only group which increased the pH in fillets, showing a higher level (6.58) together with the group of 1 day of fasting (6.64), significantly different from controls (6.40) and those fasted 2 days (6.42). At +48 h, pH was higher in trout fasted 3 days (6.82), with significant differences compared to the 1 day group (6.58), which was the only treatment where pH was reduced

65 compared to the previous sampling point. Finally, at +72 h the group fasted 2 days increased the most and reached the highest pH (7.04).

Initial muscle pH at 20h00 averaged 6.97, with statistical analysis revealing no differences among groups. From then to +9 h pH levels were very similar, slightly increasing or decreasing (average of 7.04). During the period from +2 h to +24 h no differences were found either and in all the treatments the pH decreased (by 0.22 units from +2 h to +9 h and by 0.24 from +9 h to +24 h on average). After 48 h, pH increased to a greater extent in fish fasted 3 days in comparison with 24 h post mortem. At this point, this group was significantly higher than the other groups (6.91 vs. 6.62 on average; see Fig. 25c). However, at +72 h the group fasted 1 day differed from the others (6.54 vs. 6.93 on average), as it was the only group with a lower muscle pH with respect to 24 h earlier.

2.3.2. Rigor mortis

The onset of rigor mortis occurred at some point between 2 and 9 h post mortem, while full rigor is reached between 24 h and 48 h after slaughter in almost all the groups. Statistical differences were observed only at 9 h, 24 h and 48 h post mortem.

At 08h00, rigor angles were very similar at the point of slaughter and 2 h later among groups (averaging 24.1º and 21.7º, respectively). Thereafter, stiffness dramatically increased in all the groups (by 31.7º on average, expressed as the rigor angle), with fish fasted 3 days having the lowest angle (36.3º vs. 53.9º in the other three groups on average). No more statistical differences were observed throughout the study period. Maximum rigor angle was reached at +24 h in all the groups (averaging 78.6º). During the period 24-48 h post mortem, rigor angles slightly decreased and after 72 h, rigor angles decreased again to similar values thatn those observed at +9h (averaging 50.3º).

Initial rigor angles at 14h00 averaged 23.2º, while at +2 h were 27.3º, showing no statistical differences among groups. After 9 h, trout fasted 1 day displayed significantly lower angles (42.3º), compared to the other three groups. Stiffening was more marked at 24 h post mortem, with an average angle of 80.1º and reaching the full rigor at this point. At +48 h and +72 h, rigor angles appeared in the following sequence: 3 days > controls / 2 days > 1 day of fasting (see Fig. 26b). The angle remained quite stable in the fed group as well as in those fasted 2 and 3 days during the period, while sharply decreased in the group fasted 1 day.

At 20h00, initial rigor angle averaged 23.5º and 24.7º 2 h later, observing no differences among groups. At +9 h fish fasted 2 days showed the higher tail angle (78.8º, which is the point of full rigor for this group) and those fasted 1 day the lowest (65.4º), with controls and 3 days having intermediate values. After 24 h, all groups reached maximum stiffness except that with 3 days of fasting. Thereafter, the rigor angle in the control group and in those fasted 1 and 2 days followed the same development, with a clear reduction during the period +48h - 72h after slaughter (averaging 62.2º and 50.0º in these groups, respectively), while rigor angle in trout fasted 3 days was significantly higher at those sampling points, remaining stable or even increasing, reaching full stiffening at +72h (79.6º).

90.0 (a) 80.0 70.0 60.0 a a 50.0 a 40.0 b 0 days

Rigorangle (º) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (b) 80.0 a a 70.0 b a b a 60.0 a c b 50.0 b 40.0 b 0 days

Rigorangle (º) 30.0 1 day c 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (c) 80.0 a ab a a 70.0 ab b b 60.0 b b b 50.0 b b 40.0 0 days

Rigorangle (º) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 26. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 2. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=4.0). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

2.3.3. Percentage of water released

Overall, very little variation among groups was observed, with values remaining very stable throughout the experimental period. These PWR values were between 15% and 25% in most cases, with slight variations at some sampling points, especially at 08h00 (Fig. 27).

35.00 (a) 0 days 30.00 a 1 day 2 days 25.00 b 3 days 20.00 b

15.00 b PWR (%) 10.00

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 0 days (b) 30.00 1 day 2 days 25.00 a a 3 days 20.00 a

15.00 PWR (%) 10.00 b

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 (c) 0 days 30.00 1 day

25.00 2 days 3 days 20.00

15.00 PWR (%) 10.00

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 27. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 2. Data represent means (n=3 in fasted groups and n=9 in controls; SEM=2.04). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

At 08h00, initial PWR was 15.57%, with no differences observed among groups. However, treatments differed significantly 24 h later, with trout fasted 2 days showing the highest levels (30.52%) in comparison to the other three groups (averaging 18.68%). Statistical analysis

68 showed no more differences during the study, with PWR averaging 15.36% at +48 h and 16.83% at +72 h.

At 14h00, the initial PWR of fish fasted 1 day was consistently lower than the rest of the groups (10.19% vs. 18.60% in average). After 24 h, PWR levels remained quite stable, except in the group fasted 1 day, where increased dramatically up to 24.07%. Mean PWR value at this point was 21.49%. At +48 and +72 h, PWR did not undergo major changes, averaging 19.30% and 17.3%, respectively.

Finally, at 20h00 no differences among groups were observed at any sampling point. Mean PWR values were 15.29%, 17.92%, 15.78% and 15.80%.

2.3.4. Meat colour

Overall, a significant effect of hours post mortem (time of storage) was found for L* and a*- values (P<0.001), without any effect of the rest of the variables involved in the statistical model (including hours of slaughter or days of fasting; P>0.05). No effect was found for the b*-value.

The evolution of both L* and a* coordinates (CIE Lab) throughout the experimental period is displayed in Fig. 28. Mean L*-values (and thus lightness) significantly increased at 24, 48 and 72 h post mortem (by 4% on average) compared to initial levels. On the other hand, a*-values were quite similar, excluding mean levels at 24h post mortem, when lower levels were observed.

52.00 2.50 L* 51.00 a* a 2.00 50.00 1.50 b b 49.00 c

1.00 a* L* 48.00 a a 0.50 47.00 a b 46.00 0.00

45.00 -0.50 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 28. Colour evolution (L* and a* coordinates, CIELab) of rainbow trout flesh during 72 h of storage in TRIAL 2. Data represent means±SEM (n=90). Significant differences among sampling points within a coordinate are denoted with different letters (PL* <0.001 and Pa* =0.030).

The b*-value was not significantly affected by any of the factors (P>0.05) in the model. Mean values per hour of slaughter, days of fasting and hours post mortem are presented in Table 7.

Table 7. Mean values of the b* coordinate of rainbow trout flesh during 72 h of storage in TRIAL 2 for the statistical model used ( ) Hour of slaughter Days of fasting Hours post mortem (h) 08h00 14h00 20h00 SEM 0 1 2 3 SEM 0 24 48 72 SEM 3.03 2.70 2.58 0.25 2.89 2.78 3.38 2.03 0.33 3.08 2.56 2.75 2.69 0.25 n (hour of slaughter)=120; n (days of fasting)=30 in 1-2-3 days and 90 in 0 days; n (hours post mortem)=90. All the P-values are >0.05 (including interactions, not shown)

3. TRIAL 3

3.1. Biometric indicators

The P-values obtained in the statistical model for every measurement are presented in Table 8.

Fasted trout weighed 407.6 g and controls 427.8 g on average (SEM=8.5) and 422.1, 414.8 and 416.3 g (SEM=8.5) at 08h00, 14h00 and 20h00, respectively. Body length was 32.3 cm and 32.6 cm in control trout (SEM=0.2). No statistical relationship could be found with any of the factors studied in both parameters.

1.50 1.40 a 1.30 a a a b 1.20 b 1.10 1.00 CC Fasted 0.90 Fed 0.80 0.70 0.60 0.50 1 2 3 Day of slaughter

Figure 29. Changes in the coefficient of condition (CC) of rainbow trout during three days of fasting in TRIAL 3. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P=0.009).

However, the hour of slaughter and type of fasting (day of slaughter) had a significant effect on the coefficient of condition (Fig. 29). Regarding the type of fasting (day of slaughter), fed fish showed significantly higher CC values on days 2 and 3 compared to fasted fish (although differences were as small as 4% on average between both groups) but those were not different from the coefficients reported on day 1 both in fasted and fed animals. On the other hand, fish slaughtered at 20h00 had a higher CC than those slaughtered at 08h00 and 14h00 (1.25 vs. 1.19 and 1.20, respectively) (SEM=0.01).

The relative weight of the gut content (RWGC) was significantly influenced by the day of slaughter and the type of fasting (day of slaughter) in the statistical model used. Fish slaughtered on day 1 showed a higher RWGC than those slaughtered on day 3 (3.4% vs. 1.5%), while trout slaughtered at 14h00 had intermediate values (2.3%) (SEM=0.5). On the other hand, the RWGC in fed fish was always higher in comparison to those fasted (by 74% on average), whereas it was lower on days 2 and 3 in fasted fish compared to day 1, whereas it was also lower on day 3 in fed fish (Fig. 30). No differences were found regarding the hour of slaughter: mean values were 2.3%, 2.3% and 2.6% at 08h00, 14h00 and 20h00, respectively (SEM=0.3).

6.00 a 5.00 a 4.00 b 3.00 b Fasted

RWGC RWGC (%) 2.00 Fed c 1.00 c

0.00 1 2 3 Day of slaughter

Figure 30. Changes in the relative weight of the gut content (RWGC) of rainbow trout during three days of fasting in TRIAL 3. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

The hepatosomatic index (HSI) was only affected by the type of fasting (day of slaughter), with all the groups having similar values (averaging 1.51%) except the group of fish fasted 3 days, which showed the lowest HSI (1.32%) (SEM=0.05). Mean values at 08h00, 14h00 and 20h00 were 1.50%, 1.48% and 1.46% (SEM=0.04).

Table 8. Summary of P-values of biometric indicators in TRIAL 3 for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Body weight NS NS NS NS NS Body length NS NS NS NS NS CC <0.001 NS 0.009 NS NS RWGC NS <0.001 <0.001 NS NS HSI NS NS 0.041 NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

3.2. Haematological indicators

Plasma cortisol had a significant influence of the hour of slaughter. Fish slaughtered at 20h00 had a lower mean concentration (120 ng/ml) compared to the other two hours (162 ng/ml at 08h00 and 152 mg/ml at 14h00; SEM=7). No other significant effects in the statistical model were found. Mean concentrations in fasted fish was 140 ng/ml and in fed fish 149 ng/ml (SEM=6).

Plasma glucose was not influenced by any of the factors studied (P>0.05), with no apparent effect of days of fasting or hour of slaughter. Mean level in fasted fish was 85 mg/100 ml and in fed fish 81 mg/100 ml (SEM=3). Regarding hours of slaughter, average concentrations were 82, 86 and 82 mg/100 ml at 08h00, 14h00 and 20h00, respectively (SEM=3).

The hour of slaughter, day of slaughter and their interaction (hour of slaughter*day of slaughter) had a consistent effect on plasma lactate levels throughout the experimental period (Table 9). The average level of the groups was 21 mmol/l, except for fish slaughtered at 14h00 on day 3 (33 mmol/l) (SEM=2). Mean values in fasted and fed fish were 22 and 21 mmol/l, respectively (SEM=1).

Haematocrit however was markedly influenced by the hour of slaughter (Table 9), with levels being higher at 08h00 and 14h00 (31.0% and 32.1%, respectively) and lower at 20h00 (27.8%) (SEM=0.8). No additional significant effects were observed for this variable. In fasted fish the average level was 30.9%, whereas in fish fed regularly was 29.7% (SEM=0.7).

Leucocyte count was only influenced by the interaction hour of slaughter and day of slaughter (Fig. 31). The largest variation corresponded to fish slaughtered at 08h00, with highest levels on day 1 and lowest on day 2 (7.1 vs. 4.1x103/µl). The other groups remained quite stable during the experimental period, showing intermediate values.

8.0 a

7.0 ab bc bc bc

l) 6.0 bc bc bc

/ 3 5.0 c

4.0 08h00

3.0 14h00 20h00

Leucocytes(x10 2.0

1.0

0.0 1 2 3 Day of slaughter

Figure 31. Changes in the leucocytes levels of rainbow trout during three days of fasting in TRIAL 3. Data are means ± SEM (n=20). Different letters indicate significant differences among groups (P=0.003).

Table 9. Summary of P-values of haematological variables in TRIAL 3 for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Cortisol <0.001 NS NS NS NS Glucose NS NS NS NS NS Lactate 0.032 <0.001 NS 0.016 NS Haematocrit 0.002 NS NS NS NS Leucocytes NS NS NS 0.003 NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

3.3. Metabolic indicators

Liver glycogen, muscle glycogen and R-value did not show any effect of the factors included in the model (P>0.05). Mean values according to the type of fasting (day of slaughter) are displayed in Table 10.

Table 10. Changes in liver and muscle glycogen and R-values (IMP/ATP ratio) of rainbow trout during three days of fasting in TRIAL 3. P refers to the factor type of fasting (day of slaughter). Day 1 Day 2 Day 3 Fasted Fed Fasted Fed Fasted Fed SEM (n=5) P Liver glycogen (µg/g) 28.1 25.1 28.6 39.3 38.5 76.1 16.6 NS Muscle glycogen (µg/g) 0.90 0.30 1.82 1.30 1.73 2.32 0.60 NS IMP/ATP ratio 1.31 1.40 1.46 1.29 1.38 1.38 0.08 NS NS: not significant (P>0.05)

3.4. Product quality

The interaction hour of slaughter*day of slaughter*hours post mortem was significant (P<0.001) for muscle pH, rigor mortis and PWR, thus this interaction was displayed in figures to represent the changes in the evolution of these indicators. Regarding colour measurements, only the hours post mortem was significant within the model (see model [4] in Chapter II).

3.4.1. Muscle pH

In general terms, muscle pH showed a marked tendency to decrease up to 24 h post mortem and then level off, with very small variations. The minimum pH was reached just at that point or later on.

Initial muscle pH at 08h00 did not differ among groups, averaging 7.21. At +2 h, pH decreased in all the treatments, with fish fasted 2 days showing a greater drop, although no differences were found either. At this point, mean pH was 7.01. Muscle pH at 9 h post mortem appeared in the following sequence: 2 days > 1 day > controls > 3 days (see Fig. 32a). From that moment on, muscle pH stabilized, with mean values of 6.35, 6.33 and 6.33 at 24, 48 and 72 h after slaughter.

A very similar situation occurred at 14h00, with initial and +2 h values not differing among groups (averaging 7.19 and 7.04, respectively). After 9 h, fish fasted 1 day showed the highest pH (6.82), differing from controls (6.65) and these in turn from the group fasted 3 days (6.33). The group fasted 2 days showed intermediate values between the 1 day group and controls (averaging 6.73). The pH decreased in all the groups 24 h following slaughter, averaging 6.34. After 48 and 72 h, pH remained stable, with mean values of 6.34 for both sampling points.

At 20h00 differences were found at every sampling point except at the moment of slaughter, where all the groups showed similar values (average of 7.23). After 2 h, fish fasted 3 days had the highest muscle pH (7.24) compared with controls (6.97) and those two groups were in turn higher than those fasted 2 days (6.69). In trout fasted 1 day, the observed pH was 7.07, with intermediate values between 1 day and controls. Nine hours after slaughter, fish fasted 3 days and controls had a significantly higher pH (6.70 and 6.69, respectively) compared with those fasted 1 and 2 days (6.38 and 6.51). After 24 h, muscle pH followed the sequence 3 days >1 day > controls > 2 days (see Fig. 32c). Finally, at 48 and 72 h post mortem, pH remained stable compared to those values observed at +24 h. In both cases the group fasted 3 days showed the

74 highest pH (6.44 and 6.49, respectively) in comparison to the group fasted 2 days (6.23 and 6.24). Controls and trout fasted for 1 day showed intermediate values.

7.40 (a) 0 days 7.20 1 day 7.00 2 days

6.80 3 days a 6.60 b

MusclepH c 6.40 d

6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 (b) 0 days 7.20 1 day 7.00 2 days 3 days 6.80 a ab b

6.60 MusclepH 6.40 c

6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 (c) 0 days 7.20 a 1 day ab 2 days 7.00 b 3 days 6.80 c a 6.60 MusclepH b a a a 6.40 ab ab ab ab b ab bc 6.20 b c b 6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 32. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 3. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=0.07). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

3.4.2. Rigor mortis

Overall, the onset of rigor mortis occurred during the period 2-9 h post mortem and full rigor (maximum stiffening) took place 24 h after slaughter in all the groups (averaging 75.7º). Subsequently, rigor angle decreased, having minimum levels 72 h following slaughter.

At 08h00, rigor angles were not different among groups at the moment of slaughter and after 2, 9 and 24 h. Average levels were 35.5º, 36.6º, 64.5º and 77.9º. Thus, the onset of rigor occurs during the period between 2 and 9 h post mortem and maximum stiffening (full rigor) was at 24 h following slaughter. From that moment on, angles sharply decreased until +48 h (averaging 52.3º) and then continued decreasing, with a greater reduction in fish fasted 2 days. At this point the group fasted 3 days showed the highest angle (52.1º), followed by those fasted 1 day (45.6º), controls (38.9º) and trout fasted 2 days (29.8º).

90.0 (a) 80.0 70.0 60.0 50.0 a ab 40.0 0 days b

30.0 1 day c Rigor angleRigor(º) 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (b) a a 80.0 ab 70.0 a b a 60.0 a a 50.0 ab ab 40.0 0 days b b

Rigorangle (º) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (c) 80.0 a a 70.0 b 60.0 b 50.0 40.0 0 days

Rigorangle (º) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 33. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 3. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=3.5). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

At 14h00, initial rigor angle was 36.7º and at +2 h it was 38.0º on average. The rigor phase started from 2 h after slaughter, since at +9 h rigor angles were elevated (averaging 60.2º). No statistical differences were found until 24 h post mortem. At this point, the highest rigor angles were reached. Tail angles from fish fasted 1 and 3 days were higher compared with those fasted 2 days (82.7º, 82.3º and 71.5º, respectively), whereas controls had intermediate values (76.8º). Between 24 and 48 h, the 2 days group lost their stiffening state more quickly than the rest of the groups (the rigor angle was 44.2º vs. 60.0º on average; see Fig. 33b). Finally, at +72 h, fish fasted 3 days exhibited the highest angle again (53.8º) in comparison to the trout fasted 2 days (44.5º).

The onset and resolution of rigor was very similar at 20h00. We observed significant differences at 24 h post mortem only. Mean rigor angles were 38.3º, 38.7º and 62.6º at +0 h, +2 h and +9 h. After 24 h, stiffening was higher in controls and fish fasted 1 day, reaching full rigor at this point (except in fish fasted 2 days, where the angle was very similar compared to that observed at +9 h). During the period 24-72 h, rigor angles decreased (with an average of 51.0º and 44.0º at 48 and 72 h post mortem).

3.4.3. Percentage of water released

PWR did not show large variations throughout the trial. The minimum value measured was 9.36%, whereas the maximum was 18.33%.

The PWR values at 08h00 tended to increase starting from the moment just after slaughter up to 48 h later. Initial PWR was 10.85% and 13.90% at +24 h. Significant differences following 48 h after slaughter were observed, with controls and the groups fasted 1 and 2 days showing higher values (15.78% on average) than trout fasted 3 days (11.30%). A very similar pattern was observed at +72 h (see Fig. 34a), with groups fasted 1 and 2 days having higher PWR (14.79% and 14.39%, respectively) in comparison to those fasted 3 days (9.87%).

The PWR values at 14h00 were very similar among groups at the moment of slaughter, averaging 11.58%. However, at +24 h the group fasted 2 days differed significantly from the group fasted 3 days (18.33% vs. 13.76%), with the other two groups having intermediate values. Subsequent PWR values underwent a very small variation, averaging 13.93% and 12.63% 48 h and 72 h after slaughter, respectively.

Initial PWR at 20h00 differed significantly among groups, with fish fasted 3 days having the highest mean level (14.99%) compared to those fed regularly and fasted 1 day (10.67% and 10.17%). No differences were found 24 h post mortem (mean PWR was 14.32%), whereas at +48 h fish fasted 1 day showed the highest PWR (15.80%) compared to the group fasted 3 days (9.36%), with controls and trout fasted 2 days with intermediate values (12.84% and 11.93%). Finally, 72 h post mortem, fish fasted 3 days sharply increased the mean PWR up to 16.43%, significantly differing from the rest of the groups, which had an average PWR of 11.53%.

35.00 (a) 0 days 30.00 1 day 25.00 2 days 3 days 20.00 a a a 15.00 a a PWR (%) ab 10.00 b b 5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 (b) 0 days 30.00 1 day 25.00 2 days 3 days 20.00 a

15.00 ab ab PWR (%) b 10.00

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 (c) 0 days 30.00 1 day

25.00 2 days 3 days 20.00 a a a 15.00

PWR (%) ab ab ab b b 10.00 b b b b 5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 34. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 72 h of storage in TRIAL 3. Data represent means (n=3 in fasted groups and n=9 in controls; SEM=1.27). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

3.4.4. Meat colour

Overall, changes in flesh colour of trout meat throughout the experimental period in this trial were significantly affected by the hours post mortem (P>0.001), without any effect of the rest of the variables involved in the statistical model (including hours of slaughter or days of fasting; P>0.05).

The three coordinates had different patterns (see Fig. 35), with mean L*-values showing a marked increase as the fish spent more days in storage (by 8% on average). However, the a*- values were higher at the moment of slaughter, then decreased at 24h post mortem and levelled off at +48h. The b*-values displayed a fluctuating evolution, with lower values at 24h post mortem (approximately 24% in comparison to mean values at 0h and 48h).

52.00 18.00 L* 51.00 a* 16.00 a 50.00 b* 14.00

49.00 12.00 a

L* a b 48.00 10.00 a*,b* a 47.00 b b 8.00 c b 46.00 6.00

45.00 4.00 0 10 20 30 40 50 60 Hours post mortem

Figure 35. Colour evolution (L*a*b* coordinates, CIELab) of rainbow trout flesh during 48 h of storage in TRIAL 3. Data represent means±SEM (n=90). Significant differences among sampling points within a coordinate are denoted with different letters (PL* and Pb* <0.001; Pa*=0.0013).

4. TRIAL 4

4.1. Biometric indicators

Mean body weight was significantly affected by the hour of slaughter, with trout slaughtered at 08h00 having the highest weight (390.4 g) in comparison to those slaughtered at 14h00 and 20h00 (366.6 and 356.5 g, respectively; SEM=7.3). As seen in Table 11, body weight was also affected by type of fasting (day of slaughter). In this case, fasted fish showed lower weights than those fed and slaughtered on the same days (approximately 13% lower; see Fig. 36).

450.0 a ab ab 400.0 bc c 350.0 c

300.0

250.0 Fasted

200.0 Fed Bodyweight (g) 150.0

100.0

50.0 1 2 3 Day of slaughter

Figure 36. Changes in the body weight of rainbow trout during three days of fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

However, no effect of hour of slaughter or days of fasting was found on body length. Mean value of fasted fish was 30.9 cm, while in fed fish was 31.5 cm (SEM=0.2).

A clear effect of type of fasting (day of slaughter) was found on the coefficient of condition, relative weight of the gut content and the hepatosomatic index. Firstly, CC in fed fish was higher than in fasted fish on days 2 and 3 (by 8% on average) but not on day 1. No significant differences were found among hours of slaughter (averaging 1.25%, 1.21% and 1.19% at 08h00, 14h00 and 20h00, respectively; SEM=0.02).

1.50

1.40 a ab ab 1.30 bc c c 1.20

1.10 CC Fasted 1.00 Fed 0.90

0.80

0.70 1 2 3 Day of slaughter

Figure 37. Changes in the coefficient of condition (CC) of rainbow trout during three days of fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P=0.013).

Secondly, mean RWGC in fasted fish was significantly lower on day 1 compared to fed trout (approximately 58%) and was 0 on the following days in that group. No other differences were found in the statistical model for this parameter (mean values at 08h00, 14h00 and 20h00 were 1.5%, 1.0% and 0.8%, respectively; SEM=0.2).

3.00 a a 2.50 a 2.00

1.50 b Fasted

RWGC (%) 1.00 Fed c c 0.50

0.00 1 2 3 Day of slaughter

Figure 38. Changes in the relative weight of the gut content (RWGC) of rainbow trout during three days of fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

Finally, HSI was significantly higher in fed fish over the whole experimental period in comparison to fasted fish, averaging 1.49% and 1.19%, respectively.

1.80 a 1.60 a a 1.40 b b b 1.20 1.00

0.80 Fasted HSI HSI (%) 0.60 Fed 0.40 0.20 0.00 1 2 3 Day of salughter

Figure 39. Changes in the hepatosomatic index (HSI) of rainbow trout during three days of fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P<0.001).

Table 11. Summary of P-values of biometric indicators in TRIAL 4 for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Body weight 0.004 NS <0.001 NS NS Body length NS NS NS NS NS CC NS NS 0.013 NS NS RWGC NS NS <0.001 NS NS HSI NS NS <0.001 NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

4.2. Haematological indicators

Plasma cortisol concentrations were significantly influenced by both the hour of slaughter and the day of slaughter, but not by their interaction (P>0.05, see Table 12). Regarding the hour of slaughter, the sequence was the following: 08h00 (247 ng/ml) > 14h00 (208 ng/ml) > 20h00 (172 ng/ml) (SEM=14). With respect to the day of slaughter, fish slaughtered on day 3 showed lower concentrations (149 ng/ml) than those slaughtered on days 1 (256 ng/ml) and 2 (222 ng/ml) (SEM=14). The duration of fasting did not affect this variable, with fish fasted having a mean value of 208 ng/ml and those fed every day of 210 ng/ml (SEM=12).

Only the type of slaughter (day of fasting) had a significant influence on plasma glucose (Fig. 40). Glucose concentrations in fasted trout significantly decreased on days 2 and 3 with respect to day 1 (by approximately 13% on average). However, in fed fish, these levels remained stable, averaging 83mg/100 ml. Mean concentrations at 08h00, 14h00 and 20h00 were 78, 76 and 78 mg/100 ml, respectively (SEM=3).

100 a a 90 a a 80 b b 70 60 50 Fasted 40 Fed 30

20 Plasma Plasma glucose (mg/100 ml) 10 0 1 2 3 Day of slaughter

Figure 40. Changes in plasma glucose levels of rainbow trout during three days of fasting in TRIAL 4. Data are means ± SEM (n=30). Different letters indicate significant differences among groups (P=0.003).

The hour of slaughter, type of fasting (day of slaughter) and their interaction all had a significant effect on plasma lactate levels. However, differences were mainly due to very high values at the last sampling point (20h00 on day 3) in fed fish, where this group clearly had higher concentrations compared to fed trout (42 vs. 26 mmol/l). The mean values of that interaction are shown in Fig. 41.

50 45 a 40 35 30 b 25 20 15

Plasma Plasma lactate(mmol/l) 10 Fasted 5 Fed 0

Figure 41. Changes in plasma lactate levels of rainbow trout during three days of fasting in TRIAL 4. Data are means ± SEM (P=0.047; n=10). Different letters indicate significant differences between fasted and fed fish at sampling points (P<0.05).

Haematocrit was not influenced by any of the factors included in the statistical model. Mean levels in fasted trout were 36.2% and in fed trout 36.0% (SEM=0.6). On the other hand, mean levels at 08h00, 14h00 and 20h00 were 36.9%, 35.0% and 36.6%, respectively (SEM=0.7).

Leucocytes were only affected by the day of slaughter (Table 12), with higher values on day 1 (6.1x103/µl) compared to days 2 (4.7 x103/µl) and 3 (4.9 x103/µl) (SEM=0.2).

Table 12. Summary of P-values of haematological variables in TRIAL 4 for the statistical model used ( ). H D T(D) HD HT(D) (n=60) (n=60) (n=30) (n=20) (n=10) Cortisol <0.001 <0.001 NS NS NS Glucose NS NS 0.003 NS NS Lactate 0.010 NS 0.006 NS 0.028 Haematocrit NS NS NS NS NS Leucocytes NS <0.001 NS NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). NS: not significant (P>0.05)

4.3. Product quality

4.3.1. Muscle pH

In general, very small differences among groups were observed and they occurred within the first 9 h after slaughter. Muscle pH followed the same development in all the groups at every time of slaughter, with a reduction starting during the period 2-9 h and reaching a minimum between 48 and 72 h post mortem.

Trout slaughtered at 08h00 showed very small variations throughout the experimental period. Mean pH were 7.22, 7.14, 6.65, 6.54 and 6.51 (see Fig. 42a). The minimum pH was reached about 48 h post mortem and then it levelled off.

Initial pH at 14h00 was similar among groups (averaging 7.23) but controls and trout fasted 2 days showed the highest pH 2 h post mortem (7.03 and 7.13, respectively). Fish fasted 1 day had a marked reduction during the period 0-2 h and it was the lowest pH at that point (6.80), while the group fasted 3 days had an intermediate mean level (6.94). No more statistical differences were found, with pH averaging 6.50, 6.43 and 6.42 in the following sampling points (see Fig. 42b).

Fish fasted 2 days showed the highest muscle pH during the first 9 h post mortem at 20h00 (see Fig. 42c). Mean values of this group were 7.40, 7.19 and 6.81 at 0 h, 2 h and 9 h after slaughter. On the other hand, the group fed regularly had the lowest pH (7.16, 6.89 and 6.52). The other two groups showed intermediate values. From that moment on, pH stabilized and averaged 6.52 at +48 h and +72 h. No differences were observed among groups.

7.40 (a) 0 days 7.20 1 day 7.00 2 days 3 days 6.80

6.60 MusclepH 6.40

6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 (b) 0 days 7.20 a 1 day a 7.00 2 days ab 6.80 3 days b

6.60 MusclepH 6.40

6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

7.40 a ab (c) 0 days ab 7.20 a 1 day b ab 7.00 ab 2 days 3 days 6.80 b a ab

6.60 ab MusclepH 6.40 b

6.20

6.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 42. Muscle pH evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 48 h of storage in TRIAL 4. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=0.09). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

4.3.2. Rigor mortis

Overall, the onset of rigor occurred between 2 and 9 h after slaughter, reaching full rigor 24 h post mortem (with angles very close to 90.0º) and then maintaining that rigor (compared to +24 h) until +48 h. Most of the statistical differences among groups took place at +9 h.

The first two sampling points at 08h00 showed very similar rigor angles among groups, with a slight increase between both hours (averaging 36.4º and 38.4º, respectively). Stiffening sharply

85 increased within the period +2-9 h (when the onset of rigor seemed to take place, averaging 72.5º), excluding fish fasted 1 day, which had a steady development up to +24 h (the average angle at +9 h was 52.4º). At +24 h no differences were found, with a mean angle of 88.7º. During the period 24-48 h, significant differences were observed. The rigor angle in controls and fish fasted 3 days were consistently higher and remained stable (81.5º and 88.3º, respectively), whereas in trout fasted 1 and 2 days markedly decreased (with an average of 72.3º; see Fig. 43c).

90.0 (a) a 80.0 a a a 70.0 b a b 60.0 50.0 b 40.0 0 days

Rigorangle (º) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 a (b) 80.0 70.0 b 60.0 b 50.0 c 40.0 0 days

Rigorangle (º) 30.0 1 day 20.0 2 days 10.0 3 days 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

90.0 (c) a 80.0 a

70.0 b )

° 60.0 b 50.0 0 days 40.0 1 day 2 days

Rigorangle ( 30.0 20.0 3 days 10.0 0.0 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 43. Rigor angle evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 48 h of storage in TRIAL 4. Data represent means (n=5 in fasted groups and n=15 in controls; SEM=3.3). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

A very similar pattern was observed at 14h00. Initial rigor angles averaged 37.1º and 2 h later were 40.7 º. Following 9 h post mortem, fish fasted 2 days had the highest mean angle (86.0 º), almost reaching full rigor. At that point, this group was significantly different from controls and trout fasted 3 days, which averaged 61.6º and 58.3º, respectively, and these in turn showed significantly higher stiffening than those fasted 1 day (mean angle of 47.1º). Full rigor was reached 24 h post mortem, with an average angle of 87.2º. At +48 h mean angle was 77.6º.

Finally, at 20h00, mean rigor angle at the moment of slaughter was 36.5º, while 2 h later was 42.3º. Statistical differences were found again at +9 h, with controls and fish fasted 1 day having the highest angles (78.0º and 76.0º) in comparison to fish fasted 3 days (67.7º) and 2 days (58.9º). Stiffening increased then up to maximum levels (average of 85.3º) and slightly decreased to an average of 74.9º.

4.3.3. Percentage of water released

Initial PWR differed among groups at every time of slaughter, while it tended to decrease or remain stable during both periods from 0-24 h and 24-48 h.

At 08h00, initial PWR was consistently higher in fish fasted 1 day (19.85%) compared to the other three groups (averaging 15.14%). Thereafter, PWR levels decreased until the end of the period studied, with an average value of 13.91% at 24 h post mortem and 12.02% at the last sampling point (+48 h).

A very similar situation was observed at 14h00, when the sequence at the moment of slaughter was the following: 1 day > controls > 2 days / 3 days. Subsequently, PWR decreased except in the group fasted 3 days (where increased 1.39% units with respect to the initial value), although no differences were found, averaging 15.50%. Finally, during the period 24-48 h, PWR decreased again in all the groups, excluding those fasted 3 days (where slightly increased) and the mean PWR at that point was 14.10%.

PWR in groups fasted 1 and 3 days at 20h00 were significantly higher than those fasted 2 days at the moment of slaughter, while they significantly differed from controls and the group fasted 2 days 24 h later (see Fig. 44c). At the last sampling point, no differences were found and groups averaged 12.69%.

35.00 (a) 0 days 30.00 1 day 25.00 2 days 20.00 a 3 days b 15.00 b PWR (%) b 10.00

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 (b) 0 days 30.00 1 day 25.00 2 days a 3 days 20.00 b

15.00 c PWR (%) c 10.00

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

35.00 (c) 0 days 30.00 1 day 25.00 2 days a a 20.00 3 days a a

15.00 ab b PWR (%) 10.00 b b

5.00

0.00 0 10 20 30 40 50 60 70 80 Hours post mortem

Figure 44. Percentage of water released evolution of rainbow trout slaughtered at 08h00 (a), 14h00 (b) and 20h00 (c) during 48 h of storage in TRIAL 4. Data represent means (n=3 in fasted groups and n=9 in controls; SEM=1.61). Significant differences among days of fasting at the same sampling points are denoted with different letters (P<0.05).

5. Overall analysis of haematological indicators including water temperature

Overall, regarding physiological stress parameters, water temperature had a significant effect on plasma glucose, lactate and leucocytes.

Water temperature had no significant effect on plasma cortisol (P>0.05). Mean cortisol concentrations in both fish held at high and low temperatures were 129 ng/ml. As seen in Table 13, the whole scale analysis of the four trials together using water temperature as another factor in the statistical model, indicates that hour of slaughter was the only factor that significantly

88 affected cortisol concentrations in the four trials, where the average concentration at 08h00 (151 ng/ml) was significantly higher compared to 14h00 (130 ng/ml) and both in turn higher than at 20h00 (107 ng/ml) (SEM=6).

Plasma glucose was highly influenced by the type of fasting (day of slaughter), water temperature and their interaction in the statistical model (see Table 13). Mean concentrations for the interaction are displayed in Fig. 45. Highest levels were observed in TRIALS 1 and 3 (average temperature 11.4 ºC) on days 2 and 3 and lowest in TRIALS 2 and 4 (average temperature 20.9 ºC) on the same days.

120

100 a a a ab b b b c c 80 cd c d Fasted-20.9º C 60 Fed-20.9º C 40 Fasted-11.4º C Fed-11.4º C

20 Plasma Plasma glucose (mg/100 ml)

0 1 2 3 Day of slaughter

Figure 45. Overall changes in plasma glucose levels of rainbow trout during three days of fasting according to water temperature. Data are means ± SEM (n=60). Different letters indicate significant differences among groups (P=0.006).

The day of slaughter, water temperature and their interaction significantly influenced plasma lactate variability in the overall analysis of the four trials. Changes in levels throughout the experimental period and differences between low and high temperatures are displayed in Fig. 46. Lactate levels were similar on the first day of fasting in both groups (averaging 22 mmol/l). However, in low temperatures it significantly increased on days 2 and 3 (up to 32 mmol/l), whereas it remained stable in high temperatures.

35 a 30 b

25 c c c c 20

15 11.4 ºC 20.9 ºC

10 Plasma Plasma lactate(mmol/l) 5

0 1 2 3 Day of slaughter

Figure 46. Overall changes in plasma lactate levels of rainbow trout during three days of fasting according to water temperature. Data are means ± SEM (n=120). Different letters indicate significant differences among groups (P<0.001).

Haematocrit showed a significant effect of the hour of slaughter and the interaction hour of slaughter*water temperature in the overall analysis, with similar values at 08h00 and 14h00 among trout held at 11.4 ºC and 20.9 ºC (average temperatures) and lower levels at 20h00 with low water temperature.

35.0 a a 34.0 a a a 33.0

32.0 b 31.0 11.4º C 20.9º C

Haematocrit (%) 30.0

29.0

28.0 08h00 14h00 20h00 Hour of slaughter

Figure 47. Overall changes in haematocrit levels of rainbow trout at three different hours of slaughter according to water temperature. Data are means ± SEM (n=120). Different letters indicate significant differences among groups (P<0.001).

Finally, changes in leucocytes can be explained by two factors in the statistical model: day of slaughter and water temperature. Fish slaughtered on day 2 had significantly lower levels (5.0x106/µl) compared to day 1 (5.5 x106/µl) and day 3 (5.4 x106/µl) (SEM=0.1), while fish held in high temperatures showed lower levels (5.0 x106/µl) than those in low temperatures (5.5 x106/µl) (SEM=0.1).

Table 13. Summary of P-values of haematological variables in the overall analysis including water temperature for the statistical model used ( ). H D T(D) HD HT(D) TE HTE DTE TET(D) n=240 n=240 n=120 n=80 n=40 n=360 n=120 n=120 n=60 Cortisol <0.001 NS NS NS NS NS NS NS NS Glucose NS NS 0.003 NS NS <0.001 NS NS 0.006 Lactate NS <0.001 NS NS NS <0.001 NS <0.001 NS Haematocrit 0.048 NS NS NS NS NS <0.001 NS NS Leucocytes NS 0.008 NS NS NS <0.001 NS NS NS H: hour of slaughter; D: day of slaughter; T(D): type of fasting (day of slaughter); HD: interaction hour of slaughter*day of slaughter; HT(D): interaction hour of slaughter*type of fasting (day of slaughter). TE: water temperature; HTE: interaction hour of slaughter*water temperature; DTE: interaction day of slaughter*water temperature: TET(D): interaction water temperature*type of fasting (day of slaughter). NS: not significant (P>0.05)

CHAPTER IV

DISCUSSION

1. Biometric indicators

Despite the lack of food in the treatment group, it seems that three days of fasting was not enough to decrease live weight in rainbow trout, as reported by Pottinger et al. (2003) and Sumpter et al. (1991), who only found differences in live weight after four and six weeks of fasting, respectively. Nikki et al. (2004) concluded that rainbow trout can be subjected to short term fasting (up to 14 days) without any significant effect on body weight. Similar results have been found by McMillan & Houlihan (1992) after 6 days of food deprivation.

Regarding body length, we found no significant differences in any of the factors included in the statistical model. However, in TRIAL 1, the fish slaughtered on the first day (n=60), had a smaller average length than on days 2 and 3. That may be a sampling bias where despite a random assignment of all fish into the six separations at the beginning of the adaptation period, more smaller fish were captured on day 1 in each separation (i.e., the first 10 fish caught per separation were the smallest of the group). In any case, the difference in length was at most 8 mm, so differences between caught fish were difficult to observe with the naked eye. Other authors have also found little effect of fasting on fish length until after several weeks (Sumpter et al., 1991).

The coefficient of condition (an index of fish volume) has no units (as compared to weight or length) and is widely used in aquaculture to evaluate the nutritional status of fish. It is more independent of differences in fish weight among treatments or even among trials. According to our study, the CC of fed fish was higher than fasted fish after 2 days of fasting, implying that CC is more sensitive than live weight as an indicator of nutritional status. In TRIALS 2, 3 and 4, CC was always higher in control fish than fasted fish on days 2 and 3. In TRIAL 2, the CC of the fed fish increased, while in TRIAL 3 the CC of fasted fish decreased. In TRIAL 4 the CC was more similar but decreased slightly in the fasted fish. No changes in CC with fasting were observed in TRIAL 1, probably because of the bias in body length mentioned above.

Comparing among trials, the average CCs were higher in TRIALS 3 and 4, probably because fish were heavier (417.7 g and 368.8 g vs. 306.7 g and 214.9 g in TRIALS 1 and 2, respectively) and due to some other reasons including nutritional status, age or sex. The finding that CC was lower in fasted fish just after two days is quite novel, since most other studies found in the literature suggest that CC only find differences after 1-2 weeks of fasting (Sumpter et al., 1991; Pottinger et al., 2003). Similar results are reported for other species such as sea bass (Caruso et al., 2011), yellow perch (Foster & Moon, 1991) and catfish (Peterson & Small, 2004).

The weight of the digestive system and its contents always decreased with increasing days of fasting. The difference in gut content between fasted fish and fed fish is apparent after the first 24 hours. As more trials were performed, we obtained more details about stomach weight and content. Thus, no data were taken during TRIAL 1, and in TRIAL 2 we only weighed the whole gut. The changes in RWWG of fasted trout in TRIAL 2 suggested that they emptied their stomach within the first 24 h after the cessation of feeding. That tendency is seen in TRIAL 4 (where actual gut content was measured, i.e., RWGC), but most fish still had about 1% of their body weight in food still in their stomach after 24 hours of fasting, with little or no food left by 48 hours fasting. Interestingly, in TRIAL 3, held at a lower temperature than TRIALS 2 or 4, there are still some food remains after 48 and even 72 hours, but they could be considered negligible (representing approximately 0.5% and 0.3%) of live weight. Gut clearance is a time and temperature dependent equation, which requires a longer time at low temperatures (Usher et al., 1991). According to our results, commercial producers should consider that at least one day of fasting is necessary to empty the gut completely and at least one more day if temperature is colder. The duration of fasting needed to empty the gut depends on the species and water temperature, but may be expected to be from 1 to 5 days (Lines & Spence, 2012). Independent of the temperature, the process is complete by three days in Atlantic salmon (Robb, 2008). Our results agree with McMillan & Houlihan (1992), who demonstrated that only small traces of digesta were observed after 2 days of fasting in rainbow trout weighing 102.9 g, although no specific data are provided. The reduction in food intake in control fish during TRIAL 3 on the last day could be due to the accumulation of stress after 3 days of sampling and capturing fish, which is corroborated by significantly higher lactate levels in the same group.

The hepatosomatic index, an indicator of body reserves, was lower in fasted fish. The HSI in TRIAL 3 was significantly lower than fed fish after three days of fasting and in TRIAL 4 after only one day. Similar results have been reported for rainbow trout by Farbridge & Leatherland (1992b) but their HSI was lower than our data, probably since they used smaller fish (140 g). Lefevre et al. (2008) used 400 g rainbow trout and found very similar HSI values as ours. The HSI should be lower in fasted fish since trout begin to use body reserves to supply their caloric needs, while minimizing tissue loss. However, the rate of change of HSI may vary. In rainbow trout, Farbridge & Leatherland (1992b) found a decrease with fasting after 48 h, but McMillan & Houlihan (1992) report significantly lower HSI only after 6 days of fasting (with fish weighing approximately 100 g). In other species, HSI was also found to vary (Meton et al., 2003; Pérez-Jiménez et al., 2007). The differences in the rate of change of HSI between TRIALS 3 and 4 may be related to the findings commented above regarding the total amount of feed remaining in the stomach, which was lower in fasted fish in TRIAL 4.

2. Haematological indicators

2.1. Plasma cortisol

Plasma cortisol was found to vary at different hours during the day but not with fasting. Cortisol levels have been extensively studied in fish and are perhaps the best characterized physiological variable with respect to temporal rhythms (Boujard & Leatherland, 1992). Overall, cortisol levels were lowest for trout slaughtered at 20h00, and highest at 08h00, while trout slaughtered at 14h00 had intermediate values depending on the trial. Although three sampling points are not quite enough to demonstrate a clear diurnal rhythm, the differences reported in our study may be associated with diurnal rhythms, with lower concentrations of cortisol typically occurring at night in fish, in most of the literature reviewed. Similar results have been found in rainbow trout (Reddy & Leatherland, 1995; Polakof et al., 2007) and gilthead seabream (Montoya et al., 2010) when 12L:12D photoperiods were used. However, results are different in other species. Pavlidis et al. (1999a) did not find a cortisol rhythm in common dentex with the same photoperiod and higher values at night were observed in tench (De Pedro et al., 1998). Thus, according to our results, rainbow trout are less stressed later in the day. The effect of netting, stunning and slaughter should be worse in fish that have higher basal cortisol levels (in this case in the morning). Slaughtering at night, when plasma cortisol levels are lower, may help to avoid excessive stress or negative effects on meat quality.

In the scientific literature, data about the effect of fasting on cortisol in rainbow trout are inconsistent. There are results reporting an increase in cortisol in fasted rainbow trout depending on the degree of food deprivation (Sumpter et al., 1991). Other studies report no effect (Reddy et al., 1995; Holloway et al., 1994) or even decreased levels in fasted fish (Farbridge & Leatherland, 1992a). In other species, the evidence is also contradictory, with cortisol being higher (Barcellos et al., 2010), lower (Barton et al., 1988) or even no effect (Weber & Bosworth, 2005) in fasted jundia, salmon and channel catfish compared to fed fish, respectively. In the present study, changes in cortisol among the three hours of slaughter were observed in both fasted and fed trout, suggesting it was independent of fasting. That agrees with Polakof et al. (2007) who analysed trout, although other authors report higher nychthemeral concentrations in fed vs. fasted catfish (Small, 2005). Sumpter et al., (1991) and Pottinger et al. (2003) found no differences in the concentration of cortisol between fasted and fed fish after 6-9 weeks of fasting in trout, but contradict Blom et al. (2000) who found higher cortisol after three weeks of starvation. All these studies were carried out during longer periods of starvation compared to our experiment. Three days of fasting may be too short a period to provoke any physiological

effect on cortisol, but we were unable to find studies assessing cortisol using shorter periods of fasting.

Overall cortisol concentrations were higher than those reported by other authors in rainbow trout (Jentoft et al., 2005) and may correspond to animals with a high level of stress (Pottinger & Carrick, 1999; Olsen et al., 2005). It is important to consider other exogenous environmental factors or confounding effects of handling the fish pre-slaughter that may have played a role in increasing the cortisol levels above normal, such as the transport from the farm (Barton, 2000; Lines & Spence, 2012), adaptation to the new housing conditions (Kristiansen & Juell, 2002), changes in stocking density (Ewing & Ewing, 1995; Wall, 2001), social hierarchy adjustments (Wedemeyer, 1997; Mussa & Gilmour, 2012) and netting stress (Pickering, 1998; Parisi et al., 2001; Poli et al., 2005). However, because of one or several of the stressors mentioned above, cortisol levels were high in both fasted an fed fish, which contradicts Olsen et al. (2005), who found a more pronounced cortisol and lactate response in food deprived trout (3 days) subjected to acute stress.

2.2. Plasma glucose

Under different stress conditions plasma glucose levels can reach up to 150 mg/100 ml in rainbow trout (Pottinger & Carrick, 1999), with basal levels around 70-90 mg/100 ml (Jentoft et al., 2005). Thus, mean levels in the four trials studied can be considered normal, with slightly lower levels in TRIAL 2 (average glucose level was 65 mg/100 ml), which may be a reflection of the low nutritional status mentioned above, as fish probably had lower body reserves.

Regarding diurnal variations in glucose concentrations, we did not find differences among the three hours sampled. Fasting is the most frequently used approach to study whether a rhythm in glucose metabolism is independent of the disturbing/masking effect of feeding, since those daily rhythms dependent on feeding should disappear in food deprived animals. In rainbow trout, plasma glucose has been found to be higher after feeding (Figueroa et al., 2000) and at night (Holloway et al., 1994; Polakof et al., 2007), but this rhythm is abolished by food deprivation, suggesting the dependency on feeding. In our study, there no differences in glucose levels between sampling hours for control or fasted fish (i.e., no clear diurnal variation in glucose). A possible reason for this is that food was supplied continuously to control trout and not at different hours as in the studies previously mentioned, so the post-prandial peak reported by other authors did not appear.

Regarding differences between fasted and fed trout, the former normally had lower glucose levels, although some differences were observed among trials. In TRIALS 1 and 4, the average glucose concentrations in both fasted and fed trout were similar on day 1 and significantly different by days 2 and 3. In TRIAL 1, glucose levels in fed fish significantly increased on the second day. The reason for this is unclear but the most likely reason is that the control fish may have eaten more on the second and third days (higher feed intake) than on the first day, although we carefully measured and equalled the amount of feed given per fish on each day. In TRIAL 4, control fish maintained similar glucose levels throughout the trial but plasma glucose decreased significantly in fasted fish by day 2 and 3. This finding is quite logical and implies that the use of liver metabolites was not enough to maintain glucose levels (Soengas et al., 1996; Pérez- Jiménez et al., 2012). In addition, the stomach contents of fasted fish were significantly lower on days 2 and 3, compared to day 1. Glucose concentrations in rainbow trout may vary with diet and nutritional status but in general decrease with fasting since carbohydrates are the primary and earliest source of energy when fish are food deprived (Davis & Gaylord, 2011), although no variations have also been found in rainbow trout after 14 days of starvation (Soengas et al., 2006). The time that glucose levels take to start decreasing varies with species but normally not before the first 5-7 days of food deprivation in salmonids (Navarro et al., 1992; Holloway et al., 1994; Furné et al., 2012) and in many other species (tench: De Pedro et al., 2003; channel catfish: Peterson & Small, 2004; sunshine bass: Davis & Gaylord, 2011). One exception is Soengas et al. (1996), who found glucose decreased after 1 day of fasting in Atlantic salmon.

In TRIAL 2 there were significant differences in plasma glucose between fasted and fed fish only on day 1 and similar values on days 2 and 3. This may be a consequence of the low weight of fish (about 100 and 150 g lower on average than TRIALS 1 and 4, respectively) and low glucose levels in general (approximately 32% and 18% lower on average than TRIALS 1 and 4, respectively). Thus, the fish in TRIAL 2 had an overall lower nutritional status than what can be considered normal in the other trials, which may have partially masked the physiological response to fasting (Vijayan & Moon, 1994; Martínez-Porchas et al., 2009). In any case, plasma glucose levels even in the fasted fish were never below the basal levels established for most ﬁsh (65-70 mg/100 ml; Echevarría et al., 1997; Pérez-Jiménez et al., 2007; Pérez-Jiménez et al., 2012). In TRIAL 3 no changes were observed between fasted and fed fish, probably because fish were large (417.7 g on average) compared to other trials and therefore they had more reserves in liver, so glucose depletion would have taken longer.

2.3. Plasma lactate

Plasma lactate levels were quite high and quite variable between treatments and among trials. It appeared that this variation was characterized by single peaks at certain sampling points, affecting both fasted and fed fish, with no apparent overall pattern. In TRIAL 1, there were three peaks and average lactate levels were higher than TRIALS 2, 3 and 4, which also had fewer peaks. These results make it difficult to discern a circadian-like pattern, as in Polakof et al. (2007), who found no changes in plasma lactate in rainbow trout throughout the day. However, Lanford et al. (2003) suggest higher lactate concentration during the night in green sturgeon.

Regarding fasting, plasma lactate levels appear to decrease in several species of fish that have been fasted (Hemre et al., 1990; Blasco et al., 1992). That has been related to increased use of lactate as a substrate for hepatic gluconeogenesis (Vijayan et al., 1996; Polakof et al., 2006). However, our results indicate that fasting did not decrease lactate levels, as has also been reported in other species and after different periods of food deprivation. Soengas et al. (1996) found no variation in plasma lactate over 7 days of fasting in Atlantic salmon parr. Costas et al. (2011) report no changes in plasma lactate in Senegalese sole after fasting (24 h) or starvation (21 days). In cod, Black & Love (1986) did not find any obvious decrease in lactate with length of starvation, although their food deprivation period was longer (77 days). The latter authors indicate that a great variability was observed among fish and possibly reflecting different degrees of capture stress. Along those lines, the peaks in lactate levels in our study may be more related to specific bouts of anaerobic glycolysis due to strenuous exercise (with little oxygen supply to muscle, see Pagnotta et al., 1994), than to the effect of fasting. Similarly, Blasco et al. (1992) suggest that mobilization of lactate as a fuel is dependent on the stressor, and is not only related to fasting.

Moments of strenuous exercise in aquaculture are often associated with capture or even the establishment of hierarchies in a group of fish. Driedzic & Hochachka (1978), for example, report that “burst” swimming in fish increases anaerobic metabolism in fast glycolytic white muscle fibres, leading to increased levels of plasma lactate. Burst swimming is common during fish netting so presumably anaerobic glycolysis increased when we caught the fish in our trials, but the production of lactate is not immediate, taking 2-4 h to increase in plasma after a stressful event (Milligan, 1996; Olsen et al., 2005). Thus, the peaks were probably not due to capture since it took only a few minutes, shortly after which fish were given anaesthesia. Therefore, some other kind of exogenous environmental factor pre-capture, such as a change in stocking

100

density or social hierarchy adjustments as fewer fish remained in the sections as the trial continued, increased anaerobic muscular activity in few groups at different time points.

Overall, mean lactate levels in our study were quite high for rainbow trout, compared to other authors (Trenzado et al., 2003; Chowdhury et al., 2004; Morrow et al., 2004), which could indicate a more chronic stress on most of the fish such as adjustment to new housing conditions. In our study it could be that the effect of muscle exhaustion increased lactate above normal levels, which minimized any decrease in lactate that could have been observed by fasting, as found in other fish (Liew et al., 2012).

2.4. Haematocrit

Mean haematocrit levels were within normal ranges in our trials compared to other studies in rainbow trout and not significantly different in fasted fish. Average values were 35.3%, 30.2% and 36.1% and 36.1 % in TRIALS 1, 2, 3 and 4, respectively, similar to those observed by Benfrey & Biron (2000), who found average values of 38.4% in pre-stressed trout and 45.0% after the stress and Farbridge & Leatherland (1992b), with an average level of 34.7%. Chowdhury et al. (2004) report resting values of 19.2% on average and Rehulka et al. (2004) report mean levels of 30.4%-54.6%.

Strenuous activity will increase the haematocrit value (Reddy & Leatherland, 1988), since there is an increase in metabolic rate, oxygen uptake and release of erythrocytes from the spleen (Wendelaar Bonga, 1997; Gallaugher & Farrell, 1998) but the effect of hour of the day is unclear. In TRIALS 1 and 3, trout slaughtered at 20h00 had lower values, while there were no clear patterns in TRIALS 2 and 4. At 20h00 on day 1 in TRIAL 2, the overall haematocrit in fasted and fed fish was higher than at other times, which also corresponds with a lactate peak (20 mmol/l) in the same trial (see above) but the haematocrit decreased with days of slaughter. There is relatively little information about daily rhythms in haematocrit for fish. However, higher levels during the day could be related with higher activity during the photophase in rainbow trout (Young et al., 1997), as it may impose a high respiratory demand (De Pedro et al, 2005). Similar results have been found in killfish (Fundulus heteroclitus), with peak haematocrit levels 8 h after dawn (Hannah & Pickford, 1981). The lower haematocrit at night in TRIALS 1 and 3 (water temperature: 11.4 ºC) compared with TRIALS 2 and 4 (20.9 ºC), may indicate that trout are less active during the night at low temperatures, when differences between day and night are more evident.

101

Regarding the effect of fasting on haematocrit, it is not affected by food deprivation in sunshine bass (Davis & Gaylord, 2011) or sea bass (Caruso et al., 2011), and there is very little information in rainbow trout. Farbridge & Leatherland (1992b) found very small differences between fasted and fed trout after 4 weeks of food deprivation but Olsen et al (2005) report slight differences after 3 days of fasting. Our results suggest that fasting has little effect on haematocrit since there were no significant differences among fasted and fed fish after even three days of fasting or among trials.

2.5. Leucocytes

Few studies that have considered the effect of stress on immune responses in fish, but what little is known suggests a negative effect (Tort, 2011). There is little information about the effect of fasting stress on the immune response, but according to previous regarding other types of stress, it should plausibly have a negative effect. To our knowledge, this is the first experiment studying the effect of fasting on leucocytes in rainbow trout.

Regarding diurnal variations in leucocyte count, no clear pattern can be discerned from the data but it seems that time of day had little effect. In TRIAL 1 the count was higher at 14h00 compared to 08h00, in both fasted and fed fish. In TRIAL 2 there was no apparent change in leucocyte count with time of day. In TRIAL 3, results were similar to TRIAL 2 but with a peak on day 1 at 08h00. Finally, in TRIAL 4 leucocyte count was also higher on day 1 (all hours). Changes within the day may be related to diurnal rhythms, although no pattern has been described in rainbow trout. The factors that can entrain these rhythms include photoperiod (Melingen et al., 2002), temperature changes (Langston et al., 2002) or feeding (Lim & Klesius, 2003). For example, stress in general, and physical stress induced by manipulation in particular, alters the immune system in fish (De Pedro et al., 2005).

Regarding fasting effect, the only apparent effect found was in TRIAL 2, with a progressive increase in leucocytes in fed trout throughout the experimental period (with maximum values by day 3), whereas in fasted fish the levels remained unchanged. The reason for this trend is linked with the above discussion about body condition. The fish in TRIAL 2 started at a poorer overall nutritional status than the other trials, with controls improving their coefficient of condition in three days. Thus, it seems logical that leucocyte count would increase with better feeding and that leucocyte counts in the fasted fish were low. There were no significant differences in leucocyte counts in the other three trials. In any case, overall, the mean leucocyte count values were lower (averaging 5,000 cells/µl) than previous studies in rainbow trout (average of 20,000

102

cells/µl in controls Valenzuela et al., 2008; average of 50,000 cells/µl in controls, Bektas & Ayik, 2009). It has been demonstrated that glucocorticoids affect the immune response by decreasing the number of lymphocytes and eosinophils (Squires, 2003) and hence the number and proliferation of leucocytes. Thus, low leucocyte count in the present study can be related to high cortisol levels, which were already quite high (see above). It seems that low levels of leucocytes in our fish indicate a depression of the immune system due to a chronic stress (Tort, 2011).

The leucocyte response to stress is complex and involves different types of cells but it seems that stress may decrease the relative number of blood lymphocytes in the total leucocyte population. Ellsaesser & Clem (1986) showed that the total number of lymphocytes in channel catfish decreases due to transport stress. Engelsma et al. (2003) report a reduction in the number of lymphocytes after exposure to cold water temperature in common carp. Increases in the number of leucocytes due to chronic stress has also been reported in salmonids (Wedemeyer at al., 1990) as a consequence of bacterial infection (Bektas & Ayik, 2009), or confinement stress that led to neutrophilia (Morgan et al., 1993) and chronic exposure to metals in tilapia (Nussey et al., 1995).

A reduction in the number of white blood cells has been described in channel catfish after four weeks of starvation (Lim & Klesius, 2003). The results from TRIAL 2 suggests that fish with lower levels of stress have higher leucocyte counts, which partly corroborates the finding that fasted fish would be more stressed and tend to have lower leucocyte counts. However, although our results suggest that different exogenous factors played an important role in the variability of leucocytes, the high variability in results and the low overall counts do not allow to make strong conclusions using this indicator.

3. Metabolic indicators

3.1. Liver and muscle glycogen

Overall, liver and muscle glycogen were not affected by fasting, although the results were quite variable among days of slaughter. Fasting mobilizes nutrients stored in the liver to support maintenance (Davis & Gaylord, 2011) and the liver plays an important role as a source of metabolites especially during the early stages of fasting. In our study, this was evidenced by the significant changes in liver weight (HSI) mentioned above. That loss of weight may have multiple causes but the most accepted is glycogen depletion, as it is normally the first source of energy used during food deprivation (Davis & Gaylord, 2011). Reduction of liver glycogen in

103

response to feed deprivation has been documented for multiple fish species (Figueiredo-Garutti et al., 2002; Furné et al., 2012), but the time it takes to start depleting liver reserves in fish is unclear. Although it is a continuous process from the beginning of fasting (Navarro & Gutiérrez, 1995), it is widely accepted that glycogen is mobilized 5-20 days after fasting begins (Barcellos et al., 2010). Soengas et al. (1996) report a decrease in liver glycogen only after 4 days of fasting in Atlantic salmon parr, while other studies report that glycogen depletion did not begin until 7 days after the cessation of feeding (De Pedro et al., 2003; Sangiao-Alvarellos et al., 2005; Barcellos et al., 2010). In rainbow trout, Furné et al (2012) found that liver glycogen significantly decreased after 5 days of fasting. However, in our study (glycogen depletion was only measured in TRIAL 3), the loss of liver weight was not accompanied by a decrease in glycogen reserves, suggesting that a different strategy is being operative to preserve the liver reserves (Moon et al., 1989).

Salmonids have been reported to belong to the group of fish that partially protect glycogen reserves during food deprivation at the expense of lipids and proteins (Sheridan & Mommsen, 1991), whereas some omnivorous species (Gadus morhua: Hemre et al. 1993; Oncorhynchus kisutch: Larsen et al. 2001) conserve protein and lipids while partially depleting glycogen stores. Increased rates of gluconeogenesis may be required to maintain glucose levels (Pereira et al., 1995; Gillis & Ballantyne, 1996), breaking down protein and lipids in liver. The use of liver lipids (Farbridge & Leatherland, 1992b) and liver protein (McMillan & Houlihan, 1992) have been reported after only 1-2 days of fasting in rainbow trout and longer periods (Storebakken et al., 1991). In addition, the high levels of cortisol may have played an important role in stimulating gluconeogenesis (Squires, 2003). Although protein and lipids in liver have not been measured in the present study, we suggest that our trout used this mechanism to maintain plasma glucose levels, which would explain the lack of changes in glycogen.

Regarding muscle glycogen, it makes only a small potential contribution to total energy expenditure and is directly related to muscular activity. According to Navarro & Gutiérrez (1995), glycogen mobilization in fish is more related to an increase in muscular activity than with fasting itself. In contrast, Furné et al. (2012) report a reduction in muscle glycogen only after 2 days of food deprivation of rainbow trout and Navarro et al. (1992) after 8 days of fasting in brown trout (S. trutta). Moreover, the depletion can occur after longer periods of starvation (cod, 77 days: Black & Love, 1986; carp, 19 days: Blasco et al., 1992). The overall conclusions of Navarro & Gutiérrez (1995) are veriﬁed in the present study, as shown by the absence of the effect of fasting on muscle glycogen content, as also reported in rainbow trout (Kieffer & Tufts, 1998; Harmon et al., 2011) and other species (sea bass: Gutiérrez et al., 1991; jundia: Barcellos et al., 2010).

104

3.2. R-value

To our knowledge, this is the first experiment determining R-values in fasted trout. The IMP- ATP ratio is widely used as an indicator of PSE and DFD meats in terrestrial animals, as stressed individuals prior to slaughter will have a faster degradation of the ATP and thus a lower ratio. In fish, the R-value is a useful indicator of the amount of stress exposure just prior to death (Korhonen et al., 1990), with ratios of 0.80-0.90 are typical of unstressed fish (Giuffrida et al., 2007). In our study, the average levels (approximately 1.37) in both fasted and fed fish indicate that the amount of IMP was higher than ATP and thus that ATP was rapidly split into IMP at the moment of slaughter. Although we observed higher R-values when compared to the aforementioned authors, these levels were similar in both fasted and fed fish. Thus, it appears that fasting did not affect the rate of ATP degradation but these results, along with plasma cortisol and lactate, suggest that some kind of stress may have increased energy use in the muscle of all fish before the slaughter in TRIAL 3.

4. Product quality

4.1. Muscle pH

Initial pH values were slightly lower in TRIALS 1 and 2 (6.80 and 7.00, respectively) compared to TRIALS 3 and 4 (7.20 and 7.30, respectively). Meat pH always fell after slaughter until 24- 48 h post mortem (ultimate pH), after which values were quite stable, as previously reported in rainbow trout (Duran et al., 2008; Lefevre et al., 2008), Atlantic cod (Kristoffersen et al., 2006) or sea bass (Bagni et al., 2007). In other studies, the ultimate pH varies from 6.5 (Lefevre et al., 2008) to 7.0 (Robb et al., 2000), whereas in our trials it was slightly lower, ranging from 6.02 to 6.70. In fish, an initial pH reduction right after death is attributed to the concentration of lactic acid produced by the anaerobic metabolism of glycogen in the muscle (Grigorakis et al., 2003), and this pH reduction is accompanied by the onset of rigor mortis (Huss, 1995).

Meat pH values varied with hour of slaughter, but the results are more difficult to interpret than most blood plasma indicators since there was a significant interaction between hour of slaughter and the duration of fasting in all trials and we must also consider the effect of storage time. The results suggest that the fall in pH values (in the early stages of storage, i.e. between 2 and 9 h post mortem) were more homogeneous when the fish were slaughtered at 08h00 and 14h00 and then there was more variability among the different fasting durations (within each hour of slaughter) as the day progressed, with the greatest variability at 20h00. The only exception to this general trend was TRIAL 2, which as mentioned above, had fish with a poorer nutritional

105

status. It was the only trial where meat pH increased from 0 hours to 2 hours post-mortem, where there was more variability among fasting groups at 08h00, where the difference between initial pH and ultimate is smaller compared to other trials (a drop of 0.3-0.4 compared to 0.6- 1.0).

Regarding fasting stress, in several cases pH values at 24, 48, 72 hours were higher for fish fasted 3 days compared to 1 or 2 days, and controls. The clearest example was for TRIAL 3 at 20h00, where all pH values post-mortem for fish fasted 3 days were higher than the other groups. This may indicate that these fish were stressed and the glycogen stores were depleted and thus less lactic acid was produced, causing pH to remain higher (Newton & Gill, 1981). A similar trend is visible in the other trials as well with, again, the highest variability observed in TRIAL 2, with greater differences among fasting groups at every hour of slaughter. However, lower pH in trout fasted 1 day and slaughtered at 20h00 (TRIAL 1) may be associated with PSE meats, which could indicate a high activity and thus a higher stress prior to slaughter (Lowe et al., 1993; Thomas et al., 1999). The only haematological indicator that suggests that this group of trout had higher levels of activity or stress was lactate, which peaked in this group (20h00 day 1). The only other peak in lactate in TRIAL 1 is also associated with a faster decrease in pH on day 3 at 8h00. Both of these results suggest a relationship between stress, lactate levels and an early pH fall related to PSE meats. In the literature, evidence on the effect of starvation on pH levels during storage is contradictory: Einen & Thomassen (1998) found no differences between 0 and 3 days of fasting in Atlantic salmon and Álvarez et al. (2008) report higher pH with 24 and 48 h of starvation compared to 72 h in gilthead sea bream.

4.2. Rigor mortis

Overall, initial rigor angles were clearly higher in TRIALS 3 and 4 (averaging 37º in both trials) compared to TRIALS 1 and 2 (25º and 24º, respectively). This indicates greater stiffening in TRIALS 3 and 4, even though the method used to store the fillets was always similar. These differences may correspond with differences observed in muscle pH, since initial pH was higher in TRIALS 1 and 2, but also with the body condition of fish (CC), which was clearly lower in those trials. Thus, trout in TRIALS 1 and 2 probably had lower reserves and less lactic acid produced just after death, leading to a lower pH decrease compared to TRIALS 3 and 4. Since high pH environments tend to decrease rigor rate in fish (Love & Haq, 2007), it follows that fish from TRIALS 1 and 2 had a lower initial rigor rate, so rigor mortis began later than TRIALS 3 and 4. In addition, we must consider the different sources of the fish, which may have had an effect.

106

Initial rigor angles in the four trials are low compared to other authors for rainbow trout (Robb et al., 2000) but those observed in TRIALS 3 and 4 are similar to Atlantic salmon (Mørkøre et al., 2008). However, rigor angles varied with the treatments. The onset of rigor occurs between 2 and 9 h post mortem, while maximum stiffening (rigor angle) was observed 24 h after slaughter in most of the trials, which coincides with the moment of minimum pH (Love et al., 1974), as low muscle pH (as a consequence of glycolysis) interrupts the production of ATP through several processes, which in turn leads to the onset of rigor. These results coincide with other studies reporting similar rigor angles in rainbow trout (Robb et al., 2000; Merkin et al., 2010), salmon (Roth et al., 2006) and cod (Kristoffersen et al., 2006). However, the resolution of rigor (i.e. the moment when the rigor starts to decline and the muscles soften) is less clear, as a great variability among trials was found. Overall, in TRIALS 1 and 3, it seems that the resolution of rigor occurs after 24 h post mortem, which makes the muscle soften and lower angles are observed. Contrarily, in TRIALS 2 and 4 fillets remains in rigor for a longer time, with very similar angles at +24 and +48 h and a final decrease after 72 h (only in TRIAL 2; this parameter was not measured in TRIAL 4 after 72 h of storage). A similar tendency was reported by Mørkøre et al. (2008). Thus, water temperature probably plays a major role, as it is the only source of variation between TRIALS 1-3 and 2-4. We suggest that high temperatures (around 20 ºC) delay the resolution of rigor compared to normal temperatures.

Regarding the effect of hour of slaughter on rigor, no clear pattern could be discerned, as the interaction with the duration of fasting make results difficult to interpret. However, it seems that fasting had no clear effect on rigor development. Contrary to pH evolution, differences among treatments were only evident at and after 9 h post mortem, with very similar initial values (0 and +2 h). This implies that the biochemical processes taking place in post-mortem meat that affect rigor always begin at least 2 hours after slaughter. Between 2 and 9 hours post-mortem some changes occur to then allow differences at 9 h post mortem, which can be more related to fasting. As at 2 h, rigor values are similar in fish by 24 h post mortem in most of the trials, indicating that there is a window between those hours when changes between treatments may be more obvious. In future trials, it may be appropriate to measure rigor more frequently, possibly at 4, 6, 12 or 18 hours.

In relation to fasting, overall differences among groups were small and rigor development was variable, thus suggesting that fasting had no major role on rigor. Similar results have been reported by Mørkøre et al. (2008), who show similar rigor values in both fed and starved Atlantic salmon during 72 h of storage and after a 35 days fasting period. However, pre- slaughter conditions and stress (transport, crowding, water quality or slaughter method) seem to be the determinant factors that affect rigor mortis more than food deprivation, especially with

107

short-term fasting (Mørkøre et al., 2008; Bjørnevik & Solbakken, 2010; Merkin et al., 2010). On the other hand, Ginés et al. (2002) report an increase in firmness in sea bream fillets with increasing fasting (4 days vs. 2 days) after 2 days of storage. Even so, we cannot strictly compare with other studies because of the difficulties of separating temperature and species effects.

4.3. Water holding capacity-percentage of water released

Overall, the PWR seemed to remain quite stable throughout the storage period or even slightly decrease, with values between 10% and 20% in relation to meat wet weight. The exceptions are TRIAL 1 and fish slaughtered at 08h00 in TRIAL 2, where a great variability was observed, reaching values up to 32.67%.%. As muscle pH decreases, the net surface charge on the muscle proteins is reduced, causing them to partially denature and lose some of their water holding capacity (Toldra, 2003). Thus, low pH values in TRIAL 1 probably explains the high PWR (and low WHC) of fillets, as very low ultimate pH can result in meat that has relatively greater drip loss than product with a normal ultimate pH in pork muscle (Lonergan et al., 2001). Thus, minimum WHC (and maximum PWR) should be reached at the same point that minimum pH was observed. However, overall PWR was very stable, suggesting that it was independent of pH or even that some other factors played an important role. A similar trend was found by Wilkinson et al. (2008), who also observed stable drip loss values in frozen barramundi fillets after 24 h of storage. Indeed, the relationship between pH and WHC was not supported in a report which suggests that the water holding capacity was more related to structural degradation of the flesh and less to do with pH in halibut (Olsson et al., 2003).

Although there were significant differences in PWR in terms of hour of slaughter, the interactions with the duration of fasting and the storage time make results difficult to interpret. Regarding fasting stress, it did not seem to have a major effect since there were few significant differences in PWR or WHC among fasted and fed fish and the overall range of values was relatively small (around 5-10%). Evidence on the effect of fasting on WHC is unclear. Mørkøre et al. (2008) suggest that starving Atlantic salmon for 35 days significantly increases WHC compared to fed fish, while Suárez et al. (2010) showed no differences in starved (5 weeks) or fed dentex during post mortem storage. In any case, there is a general lack of a uniform method for measuring WHC (Ofstad et al., 1993; Roth et al., 2006; Ganesh et al., 2006) and it also depends on the storage temperature (Sigholt et al., 1997), so conclusions should be taken with care. In our study, trout fasted 2 days had higher PWR in TRIALS 1 and 2, at certain sampling points, suggesting that pre-slaughter stress was higher in those fish (possibly related to social

108

hierarchy, changes in stocking density or netting, see above). Roth et al. (2006) suggest that physical stress in muscle fibres or connective tissue releases proteases and accelerates the degeneration of the structure in salmon, with stressed fish having softer flesh texture and a drip loss almost 3-fold greater than rested fish. Overall, few strong conclusions can be made about the effect of fasting stress on PWR and WHC.

4.4. Meat colour

Changes in muscle colour (only measured in TRIALS 2 and 3) were independent of fasting or the hour of slaughter but clearly affected by storage time. Lightness increased significantly with storage time in both trials (average of 6% increase in TRIAL 2 and 8% in TRIAL 3), representing a discolouration of the flesh, as previously reported in salmonids (Robb et al., 2000). Contrary to lightness, a*-values had a tendency to decrease, indicating a colour change to more greenish, but more so in TRIAL 3 which included pigmented trout. The fall in a*-values, although significant, was so small (around 0.5-1.0 units) that it may be difficult for the consumer to distinguish. Regarding b*-values, they decreased after 24 h of storage in TRIAL 3 (by 24% on average) and then returned to initial levels, while in TRIAL 2 they were not affected by storage time. Similar reductions in a* and b*-values have been reported by Álvarez et al. (2008) in sea bream but they took place only after 7 days of storage. Also, flesh colouration may be affected by acidification from anaerobic glycolysis during storage. Swatland (2003) found substantial evidence that fibre diameters of pork muscle are inversely related to muscle pH, and that pH dependent microstructural changes significantly influence the colour perception due to altered refraction through the muscle fibres. In fish, it has been demonstrated in halibut that muscle becomes more insoluble with increasing storage time and decreasing pH (Tomlinson et al., 1965). In our study, the increasing L*-values and decreasing a* and b*-values may correspond with the normal development of meat just after slaughter, with a decrease in pH and increases in the stiffening and water released from the muscles cells as a consequence of the loss of protein solubility (Warriss & Brown, 1987), which lowers liquid holding capacity (Bendall & Wismer-Pederson, 1962).

We found little effect of fasting on colour values. In any case, in white flesh fish the link between fasting and changes in lightness are less important than in pigmented fish (like those used in TRIAL 2), where long periods of starvation may change the carotenoid content in flesh (Ginés et al., 2002). In Atlantic salmon no changes in colour were reported within the first three days of fasting (Einen & Thomassen, 1998) but after 12 days the L* decreased, which is in agreement with Álvarez et al. (2008), who found that the L* significantly decreased after 21

109

days of storage in gilthead seabream. Regarding colour (a* and b*), they do not vary with fasting or starvation in both pigmented (Einen & Thomassen, 1998) and non-pigmented fish (Ginés et al., 2002). In rainbow trout, no significant decrease in pigment concentration during starvation has been found (Foss et al., 1984; Choubert, 1985). In contrast, Wathne (1995) found significantly decreased CIE L*, a* and b* values in starved Atlantic salmon (pigment content was not measured in that study).

As expected, a*-values and b*-values were markedly higher in TRIAL 3 compared to TRIAL 2 (approximately 25 times and 3 times, respectively), indicating a more reddish and yellowish flesh, as the trout were fed with feeds containing 30 ppm of astaxanthin. However, L*-values were very similar between TRIALS 2 and 3, indicating that pigmentation did not greatly affect this value. The values obtained in TRIAL 2 agree with colour development in white flesh fish (Ginés et al., 2002; Álvarez et al., 2008), while in TRIAL 3 agree with pigmented flesh in salmonids (Einen & Thomassen, 1998; Lefevre et al., 2008).

5. Overall analysis of haematological indicators including water temperature

Overall, water temperature had a clear effect on plasma glucose, plasma lactate and haematocrit, especially in trout held at low water temperatures, with different responses depending on the indicator.

It is clear that water temperature did not affect plasma cortisol, as it did not have a significant influence on the overall analysis. The pattern seems to be similar in TRIALS 1-3 and 2-4, with lower values at 20h00 and higher at 08h00, thus confirming the results obtained in the trials individually. Although 19.2 and 22.7 ºC (TRIALS 2 and 4, respectively) are at the high-end of the optimal range for rainbow trout (FAO, 2011; Skov et al., 2011) and salmonids (Brett, 1971), similar results have been reported by Barton & Schreck (1987), who found that chinook salmon reared at 12.5 ºC and 21.0 ºC had similar cortisol responses after handling stress. In contrast, Davis (2004) report higher cortisol levels in sunshine bass (which has a very similar optimal temperature range to rainbow trout) held below 20 ºC compared to 20 ºC and above. It seems that daily rhythms and social stress and/or pre-slaughter handling were the responsible for changes in plasma cortisol, while water temperature had no or little effect.

Overall, it seems that water temperature had an obvious effect on plasma glucose concentrations, with a clear increase in trout held at low temperatures on days 2 and 3 in both fasted and fed fish. This implies that water temperature played a more important role than

110

fasting in plasma glucose variations during the experimental periods, probably because 3 days is a too short period to provoke important changes. Differences between fasted and fed fish are more evident at high temperatures, possibly because higher temperatures cause a further increase in metabolism (Fry, 1971), depleting glycogen and glucose reserves more rapidly, as also previously reported by Sherstneva & Shabalina (1973) and Hilton (1982) in muscle and liver of fasted rainbow trout, respectively.

Plasma lactate was also affected by water temperature, with mean concentrations stable at high temperatures but increasing on days 2 and 3 in fish held at low temperatures. This may indicate that lactate accumulated over the experimental periods due to repeated pre-slaughter stress (e.g. netting or changes in stocking density) but only at low temperatures, for unknown reasons. Although this finding disagrees with the results for haematocrit, we believe that lactate changes are more related to the type and degree of pre-stress slaughter than to water temperature, which agrees with Rapp et al. (2012), who found that changes in lactate in common carp were dependent only on the time of stress more than on temperature (12 ºC vs. 22 ºC).

Water temperature had an overall significant effect on haematocrit throughout the experimental periods, with lower values at 20h00 in fish held at low temperatures (average of 11.4 ºC). This corroborates the findings mentioned above, with trout being less active during the night at low temperatures. In relation to this, De Pedro et al. (2005) found that tench (which has a strictly nocturnal pattern of activity), showed higher haematocrit values at night with high temperatures compared to day. Thus, the higher activity during the scotophase (Young et al., 1997) in rainbow trout should explain the lower values in low temperatures at night. This finding disagrees with Valenzuela et al. (2008), who found that haematocrit increased after 14 days of exposure to high temperatures (18 ºC) in rainbow trout, although no data are available on previous days. Also, Denton & Yousef (1975) found that water temperature does not seem to be responsible for changes in haematocrit, with diet, metabolic adaptations and activity as probable causes of seasonal changes. However, these results are consistent with results on plasma glucose, suggesting that increased metabolism is present in fish held at high temperatures.

Leucocytes count was globally affected by water temperature (in both fasted and fed fish) according to the statistical analysis, with higher levels in fish held at high temperatures. In the literature reviewed, high temperatures decrease the number of leucocytes. Langston et al. (2002) found lower leucocytes levels in Canadian halibut at 18 ºC than at 12 ºC and Valenzuela et al. (2008) report lower levels in rainbow trout held at 18 ºC compared to 11 ºC but only after 30 days after the exposure. Engelsma et al. (2003) and Pettersen et al. (2005) showed that thermal shock affects leucocytes by diminishing the number of B lymphocytes and increasing

111

granulocytes in common carp and Atlantic salmon, respectively. However, the day of slaughter also influenced leucocytes, with lower levels on day 2. We believe that this is a reflection of the high variability found in the trials, which is independent of the water temperature. The small range of values found in our results and the low average levels compared to other authors in rainbow trout (Valenzuela et al., 2008; Bektas & Ayik, 2009) suggest that the overall effect of water temperature on leucocytes levels is small.

In conclusion, the present Thesis shows for the first time results based on multiple welfare and product quality indicators on how pre-slaughter handling procedures affect rainbow trout, the most important freshwater species in Spain. It provides useful advice on how to handle fish prior to slaughter more properly on a commercial level, with some useful conclusions.

One of the main novelties of this Thesis is to include degree days, not just time or not only temperature to evaluate the effect of fasting, as trials were carried out in different seasons of the year and thus at different temperatures. Along those lines, analysing a higher limit, or in the future a lower limit (less temperature, less degree days) may provide more information about fish stress than carrying out trials at average temperatures (e.g., 12-14 ºC). In any case, when farmers subject fish to fasting prior to slaughter, the degree day range should be considered, not only the days of fasting, i.e., at 12 ºC, we could still starve for 3 days but the total degree days (which we think is the most important) would only be 36 ºC d, or approximately half of the maximum limit we tested (68 ºC d). Although it has been demonstrated that gut emptying occurs within the first 24-48 h after food deprivation, haematological indicators do not suggest any harmful effect of fasting. Thus, extending the fasting period would be suitable if carried out at lower temperatures, not exceeding this maximum limit.

Also, results on haematological indicators suggest that trout are less stressed at 20h00 and thus the effect of pre-slaughter procedures such as netting may have a lower impact. Hence, we suggest to slaughter later in the day. Although some other factors such as the initial nutritional status of fish or pre-slaughter stressors mentioned above may have played a major role on haematological changes, we tried to minimize their effect using of a high number of control fish.

Based on the results of this Thesis, current recommendations on fasting prior to slaughter should be reviewed in terms of degree days and not only days, as temperature plays a very important role depending on the season of the year. It is also necessary to conduct more studies on fasting in this species to better understand the strategies that rainbow trout use in response to short-term food deprivation, especially regarding glucose/glycogen metabolism. In relation to this, future

112

research should also evaluate the use of some indicators such as liver lipids and/or protein and muscle ATP and lactate to determine the extent of energy mobilization during the early stages of fasting. Finally, relationship between fasting and product quality should also be studied in more detail to determine the possible links between pre-slaughter stress and fasting at different hours in the day and with different water temperatures to better understand how the biochemical changes in meat affect handling procedures and the extent to which these are detrimental for rainbow trout. Using more indicators (firmness, sensory analyses, muscle CP) and standardizing methods of analysis is recommended, as well as evaluation of the effect of freezing storage, as it is known to affect meat quality.

113

CONCLUSIONS

1. Biometric indicators

Three days of fasting prior to slaughter is not enough to decrease live weight in rainbow trout but the body condition of fasted fish was lower after 2 days of food deprivation compared to fed fish. Stomach emptying occurs within the first 24 h of fasting in rainbow trout, with small traces of digesta remaining after 48 and 72 h at lower temperatures. Thus, commercial producers should consider that at least one day of fasting is needed to remove food from the gut completely, but at colder temperatures the fish may need at least one more day. The HSI clearly decreased in fasted trout but the rate of change may be also related to the amount of feed remaining in the stomach, with greater differences between fasted and fed trout when gut clearance is faster.

2. Haematological indicators

Plasma cortisol was not affected by fasting. However, it was lower at 20h00, suggesting that trout were less stressed later in the day. Water temperature did not affect cortisol changes. Plasma glucose response to food deprivation was highly variable in the four trials but overall differed between fasted and fed fish as the fasting progressed. Initial nutritional status of fish probably played a major role in the variability observed, as well as water temperature, which had a great impact based on the overall analysis of the four trials, with higher concentrations on days 2 and 3 and smaller differences between fasted and fed fish at lower temperatures. Three days of fasting did not affect lactate levels. Our results suggest specific bouts of anaerobic glycolysis due to strenuous exercise and that these bouts were more evident at lower temperatures. Haematocrit results indicate that trout were less active during the night, thus confirming the diurnal patter of activity in rainbow trout, but only with low water temperatures, as higher temperatures increases metabolism. Leucocytes counts were quite low in the four trials. However, the high variability observed does not allow to make strong conclusions.

3. Metabolic indicators

Liver glycogen was not affected by fasting in TRIAL 3. Thus, the decrease in liver weight (HSI) in that trial suggests that fish used another strategy to maintain plasma glucose levels, probably liver lipids and/or protein through gluconeogenesis. Muscle glycogen did not vary with fasting either, probably because glycogen mobilization in fish is more related to an increase in muscular activity than to fasting itself. The results on the R-value showed that the amount of IMP was

117

higher than ATP, suggesting that pre-slaughter stress may have increased energy use in the muscle of all fish.

4. Product quality

Overall, both fasting and the hour of slaughter did not have a major effect on product quality development under our experimental conditions, as storage seems to be the determinant factor influencing meat changes. Ultimate muscle pH was observed between 24 and 48 h post mortem. The fall in pH values during the early stages of storage (0-9 h post mortem) was more homogeneous at 08h00 and differences among durations of fasting increased as the day progressed. However, results in TRIAL 1 suggest a relationship between pre-slaughter stress, plasma lactate levels and early fall of pH associated to PSE meats. Full rigor was observed at 24 h post mortem, coinciding with the moment of minimum pH. Also, nutritional status of fish probably influenced initial development of pH and rigor mortis, as trout in TRIALS 1 and 2 had lower reserves and produced less lactic acid just after death, making that initial pH was higher and stiffening higher. However, water temperature possibly affected the resolution of rigor, as it was delayed in TRIALS 2 and 4, when fish were held at higher temperatures.

Results on rigor development suggest that biochemical changes that affect rigor probably occur after 2 h of storage and that there is a window between 2 and 24 h post mortem when differences among treatments are more evident. Water released from muscle fibres remained stable throughout the storage period in most of the trials, suggesting that changes are independent of muscle pH. Meat colour followed a typical development in meat just after slaughter and changes were influenced by the storage time more than fasting. Overall, lightness increased and a* and b*-values remained stable or slightly decreased. Differences between trials may be due to different pigmentations but only in a* and b*-values, which were clearly higher in pigmented trout (TRIAL 3).

118

REFERENCES

121

Ainsworth, A.J., Dexiang, C., Waterstrat, P.R. (1991). Changes in peripheral blood leucocyte percentages and function of neutrophils in stressed channel catfish. Journal of Aquatic Animal Health 3, 41–47. Alanara, A., Burns, M.D., Metcalfe, N.B. (2001). Intraspecific resource partitioning in brown trout: the temporal distribution of foraging is determined by social rank. Journal of Animal Ecology 70, 980–986. Alsop, D. & Vijayan, M. (2009). The zebrafish stress axis: molecular fallout from the teleost-specific genome duplication event. General and Comparative Endocrinology 161, 62- 66. Álvarez, A., García García, B., Garrido, M.D., Hernádez, M.D. (2008). The influence of starvation time prior to slaughter on the quality of commercial-sized gilthead seabream (Sparus aurata) during ice storage. Aquaculture 284, 106–114. Ang, J. & Haard, N.F. (1985). Chemical composition and postmortem changes in soft textured muscle from intensely feeding Atlantic cod, Gadus morhua. Journal of Food Biochemistry 9, 49-64. Ashley, P.J. (2007). Fish welfare: Current issues in aquaculture. Applied Animal Behaviour Science 104, 199-235. Azam, K., Mackie, M., Smith, J. (1989). The effect of slaughter method on the quality of rainbow trout (Salmo gairdneri) during storage on ice. International Journal of Food Science and Technology 24, 69–79. Azevedo, P.A., Young Cho, C., Leeson, S., Bureau, D.P. (1998). Effects of feeding level and water temperature on growth, nutrient and energy utilization and waste outputs of rainbow trout (Oncorhynchus mykiss). Aquatic Living Resources 11, 227-238. Bagni, M., Civitareale, C., Priori, A., Ballerini, A., Finoia, M., Brambilla, G., Marino, G. (2007). Pre-slaughter crowding stress and killing procedures affecting quality and welfare in sea bass (Dicentrarchus labrax) and sea bream (Sparus aurata). Aquaculture 263, 52-60. Barcellos, L., Marqueze, A., Trapp, M., Quevedo, R.M., Ferreira, D. (2010). The effects of fasting on cortisol, blood glucose and liver and muscle glycogen in adult jundiá Rhamdia quelen. Aquaculture 300, 231-236. Barry, T.P., Lapp, A.F., Kayes, T.B., Malison, J.A. (1993). Validation of a microtitre plate ELISA for measuring cortisol in fish and comparison of stress responses of rainbow trout (Oncorhynchus mykiss) and lake trout (Salvelinus namaycush). Aquaculture 117, 351–363. Barton, B.A. & Schreck, C.B. (1987). Influence of acclimation temperature on interrenal and carbohydrate stress responses in juvenile chinook salmon (Oncorhynchus tshawytscha). Aquaculture 62, 299-310.

121

Barton, B.A., Schreck, C.B., Fowler, L.G. (1988). Fasting and diet content affect stress induced changes in plasma glucose and cortisol in juvenile chinook salmon. Progressive. Fish- Culturist 50, 16–22. Barton, B.A. (2000). Salmonid fishes differ in their cortisol and glucose responses to handling and transport stress. North American Journal of Aquaculture 62, 12-18. Barton, B.A. (2002). Stress in fishes: a diversity of responses with particular references to changes in circulating corticosteroids. Integrative and Comparative Biology 42, 517-525. Barton, B. A., Ribas, L., Acerete, L., Tort, L. (2005). Effects of chronic confinement on physiological responses of the juvenile gilthead sea bream, Sparus aurata L. to acute handling. Aquaculture Research, 36, 172-179. Bavcevic, L., Klanjscek, T., Karamarko, V., Anicic, I., Legovic, T. (2010). Compensatory growth in gilthead sea bream (Sparus aurata) compensates weight, but not length. Aquaculture 301, 57-63. Behnke, R.J. (1966). Relationships of the Far Eastern Trout, Salmo mykiss Walbaum. Copeia 2, 346-348. Bektas, S. & Ayik, O. (2009). Hematological parameters and erythrocyte osmotic fragility in rainbow trout, Oncorhynchus mykiss, experimentally infected with Pseudomonas putida. Journal of Fisheries and Aquatic Science 4, 246-253. Bendall, J.R. & Wismer-Pederson (1962). Some properties of the fibrillar proteins of normal and watery pork muscle. Journal of Food Science 27, 144-158. Benfrey, T.J. & Biron, M. (2000). Acute stress response in triploid rainbow trout (Oncorhynchus mykiss) and brook trout (Salvelinus fontinalis). Aquaculture 184, 167-176. Bjørnevik, M. & Solbakken, V. (2010). Preslaughter stress and subsequent effect on flesh quality in farmed cod. Aquaculture Research 41, 467-474. Black, D. & Love, R.M. (1986). The sequential mobilization and restoration of energy reserves in tissues of Atlantic cod during starvation and refeeding. Journal of Comparative Physiology B 156, 469-479. Blanco Cachafeiro, M.C. (1995). La trucha: cría industrial. Ed. Mundi-Prensa, Madrid, Spain. Blasco, J., Fernández, J., Gutiérrez, J. (1992). Fasting and refeeding in carp, Cyprinus carpio L.: the mobilization of reserves and plasma metabolite and hormone variations. Journal of Comparative Physiology B 162, 539-546. Blom, S., Andersson, T.B., Forlin, L. (2000). Effects of food deprivation and handling stress on head kidney 17alphahydroxyprogesterone 21-hydroxylase activity, plasma cortisol and the activities of liver detoxification enzymes in rainbow trout. Aquatic Toxicology 48, 265–274.

122

Boletín Oficial del Estado, BOE. (2005). Real Decreto 1201/2005. Sobre protección de los animales utilizados para experimentación y otros fines científicos. B.O.E. 252, 34367- 34391. Boujard, T. & Leatherland, J. F. (1992). Circadian pattern of hepatosomatic index, liver glycogen and lipid content, plasma non-esterified fatty acid, glucose, T3, T4, growth hormone and cortisol concentrations in Oncorhynchus mykiss held under different photoperiod regimes and fed using demand-feeders. Fish Physiology and Biochemistry 10, 111–122. Brannas, E., Jonsson, S., Lundqvist, H. (2003). Influence of food abundance on individual behaviour strategy and growth rate in juvenile brown trout (Salmo trutta). Canadian Journal of Zoology 81, 684–691. Brett, J.R. (1971). Energetic responses of salmon to temperature. A study of some thermal relations in the physiology and freshwater ecology of Sockeye salmon (Oncorhynchus nerka). American Zoologist 11, 99–113. Buller, H. & Cesar, C. (2007). Eating well, eating fare: Farm animal welfare in France. International Journal of Sociology of Food and Agriculture 15, 45–58. Caggiano, M. (1999). Quality in harvesting and post-harvesting procedures-influence on quality. Fish freshness and quality assessment for sea bass and sea bream, in: Global quality assessment in Mediterranean aquaculture (CIHEAM-IAMZ), Zaragoza, pp. 55-61. Caldwell, C.A. & Hinshaw, J. (1994). Physiological and haematological responses in rainbow trout subjected to supplemental dissolved oxygen in fish culture. Aquaculture 126, 183- 193. Caruso, G., Denaro, M.G., Caruso, R., Mancari F., Genovese, L., Maricchiolo, G. (2011). Response to short term starvation of growth, haematological, biochemical and non-specific immune parameters in European sea bass (Dicentrarchus labrax) and blackspot sea bream (Pagellus bogaraveo). Marine Environmental Research 72, 46-52. Chang, L.L., Kau, M.M., Wun, W.S.A., Ho, L.T., Wang, P.S. (2002). Effects of fasting on corticosterone production by zona fasciculata-reticularis cells in ovariectomized rats. Journal of Investigative Medicine 50, 86–94. Chatzifotis, S., Papadaki, M., Despoti, S., Roufidou, C., Antonopoulou, E. (2011). Effect of starvation and re-feeding on reproductive indices, body weight, plasma metabolites and oxidative enzymes of sea bass (Dicentrarchus labrax). Aquaculture 316, 53-59. Cherry, D.S., Dickson, K.L., Cairns, J. Jr. (1977). Preferred, avoided and lethal temperatures of fish during rising temperature conditions. Journal of the Fisheries Research Board of Canada 34, 239-246. Chiba, A., Hamaguchi, M., Kosaka, M., Tokuno, T., Asai, T, Chichibu, S. (1991). Quality evaluation of fish meat by "phosphorus-nuclear magnetic resonance. Journal of Food Science 56, 660-664.

123

Choubert, G. (1985). Effect of starvation and feeding on canthaxanthin depletion in the muscle of rainbow trout (Salmo gairdneri Rich). Aquaculture 46, 293–298. Chowdhury, M.J., Pane, E.F., Wood, C.M. (2004). Physiological effects of dietary cadmium acclimation and waterborne cadmium challenge in rainbow trout: respiratory, ionoregulatory, and stress parameters. Comparative Biochemistry and Physiology B 139, 163- 173. Christiansen, D.G. & Klungsoyr, L. (1987). Metabolic utilization of nutrients and the effects of insulin in fish. Comparative Biochemistry and Physiology B 88, 701-711. CIE (Commission Internationale de l' Eclairage). (1978). Recommendations on uniform color spaces, color difference equations, psychometric color terms. Supplement No 2 to CIE Publication No 15, Colorimetry, Bureau Central de la CIE, Paris. Colombe, L., Fostier, A., Bury, N., Pakdel, F. & Guiguen, Y. (2000). A mineralocorticoid-like receptor in the rainbow trout, Oncorhynchus mykiss: cloning and characterization of its steroid binding domain. Steroids 65, 319-328. Compassion in World Farming, CIWF (2009). The welfare of farmed fish. Accessed July 2012.http://www.ciwf.org.uk/includes/documents/cm_docs/2009/f/farmed_fish_briefing_aug20 09.pdf Costas, B., Aragao, C., Ruiz-Jarabo, I., Vargas-Chacoff, L., Arjona, F.J., Dinis, M.T., Mancera, J.M., Conceiçao, L. (2011). Feed deprivation in Senegalese sole (Solea senegalensis Kaup, 1858) juveniles: effects on blood plasma metabolites and free amino acid levels. Fish Physiology and Biochemistry 37, 495-504. Council of Europe (2005). Standing Committee of the European Convention for the Protection of animals kept for farming purposes: Recommendation concerning farmed fish. Accessed July 2012. http://www.coe.int/t/e/legal_affairs/legal_co- operation/biological_safety_and_use_of_animals/Farming/Rec%20fish%20E.asp. Currie, R.J., Bennett, W.A., Beitinger, T.L. (1998). Critical thermal minima and maxima of three freshwater game-fish species acclimated to constant temperatures. Environmental Biology of Fishes 51, 187-200. Davis, K. B. & McEntire, M. E. (2006). Comparison of the cortisol and glucose stress response to acute confinement and resting insulin-like growth factor-I concentrations among white bass, striped bass and sunshine bass. Aquaculture America Book of Abstracts p 79. Davis, K.B. (2004). Temperature affects physiological stress responses to acute confinement in sunshine bass (Morone chrysops x Morone saxatilis). Comparative Biochemistry and Physiology A 139, 433-440. Davis, K.B. & Gaylord, T.G. (2011). Effect of fasting on body composition and responses to stress in sunshine bass. Comparative Biochemistry and Physiology A 158, 30-36.

124

Davis, A.K., Maney, D.J., Maerz, J.C. (2008). The use of leukocyte proﬁles to measure stress in vertebrates: a review for ecologists. Functional Ecology 22, 760-772. De Pedro, N., Delgado, M.J., Pinillos, M.L., Alonso-Gómez, A.L., Alonso- Bedate, M. (1998). Daily rhythms in NAT activity, cortisol, glucose, glycogen and catecholamines in tench (Tinca tinca (L.)). Polish Archives of Hydrobiology 45, 321–329. De Pedro, N., Delgado, M.J., Gancedo, B., Alonso-Bedate, M. (2003). Changes in glucose, glycogen, thyroid activity, and hypothalamic catecholamines in tench by starvation and refeeding. Journal of Comparative Physiology B 173, 475–481. De Pedro, N., Guijarro, A.I., López-Patiño, M.A., Mártinez-Álvarez, R., Delgado, M.J. (2005). Daily and seasonal variations in haematological and blood biochemical parameters in the tench, Tinca tinca Linnaeus, 1758. Aquaculture Research 36, 1185–1196. Denton, J.E. & Yousef, M.K. (1975). Seasonal changes in haematology of rainbow trout, Salmo gairdneri. Comparative Biochemistry and Physiology A 51, 151-153. Dethloff, G.M., Schlenk, D., Hamm, J.T., Bailey, H.C. (1999). Alterations in physiological parameters of rainbow trout (Oncorhynchus mykiss) with exposure to copper and copper/zinc mixtures. Ecotoxicology and Environmental Safety, 42, 253–264. Dreiling, C.E., Brown, D.E., Casale, L., Kelly, L. (1987). Muscle glycogen: comparison of iodine binding and enzyme digestion assays and application to meat samples. Meat Science 20,167-177. Driedzic, W.R. & Hochachka, P.W. (1978). Metabolism in fish during exercise, in: Hoar, W.S., Randall, D.J. (Eds.), Fish Physiology, Vol. 8. Academic Press, New York, pp. 503–543. Duran, A., Erdemli, U., Karakaya, M., Yilmaz, M.T. (2008). Effects of slaughter methods on physical, biochemical and microbiological quality of rainbow trout Oncorhynchus mykiss and mirror carp Cyprinus carpio ﬁlleted in pre-, in- or post-rigor periods. Fisheries Science 74, 1146-1156. Echevarría, G., Martínez-Bebia, M., Zamora, S. (1997). Evolution of biometric indices and plasma metabolites during prolonged starvation in European sea bass (Dicentrarchus labrax L.). Comparative Biochemistry and Physiology A 118, 111-123. Einen, O. & Thomassen, M.S. (1998). Starvation prior to slaughter in Atlantic salmon (Salmo salar) II. White muscle composition and evaluation of freshness, texture and colour characteristics in raw and cooked fillets. Aquaculture 169, 37–53. Einen, O., Waagan, B., Thomassen, M.S. (1998). Starvation prior to slaughter in Atlantic salmon (Salmo salar) I. Effects on weight loss, body shape, slaughter- and fillet-yield, proximate and fatty acid composition. Aquaculture 166, 85–104. Ellsaesser, C.F., Clem, L.W. (1986). Haematological and immunological changes in channel catfish stressed by handling and transport. Journal of Fish Biology 28, 511-521.

125

Engelsma, M., Hougee, S., Nap, D., Hofenk, M., Rombout, J., van Muiswinkel, W.B., Verburg-van Kemenade, B.M. (2003). Multiple acute temperature stress affects leucocyte populations and antibody responses in common carp, Cyprinus carpio L. Fish & Shellfish Immunology 15, 397-410. Erikson, U., Sigholt, T., Rustad, T., Einarsdottir, I.E. and Jorgensen, L. (1999). Contribution of bleeding to total handling stress during slaughter of Atlantic salmon. Aquaculture International 7, 101–115. Escobar, C., Díaz-Muñoz, M., Encinas, F., Aguilar-Roblero, R. (1998). Persistence of metabolic rhythmicity during fasting and its entrainment by restricted feeding schedules in rats. American Journal of Physiology 274, 1309–1316. Essen-Gustavsson, B, Karlström, K, Lundström, K, Pösö, R. (1991). Adenine nucleotide breakdown products in muscle at slaughter and their relation to meat quality in pigs with different halothane genotypes. Proceedings of the International Congress of Meat Science & Technology, 37 th, Vol 1, Kulmback, Germany. European Commission (2007). Attitudes of consumers towards the welfare of farmed animals — Wave 2. Special Eurobarometer 229(2)/Wave 64.4 — TNS Opinion & Social. Accessed August 2012.http://ec.europa.eu/food/animal/welfare/survey/sp_barometer_fa_en.pdf. European Food Safety Authority, EFSA (2008). Animal welfare aspects of husbandry systems for farmed trout. Scientific Opinion of the Panel on Animal Health and Animal Welfare. EFSA Journal 796, 1-22 European Food Safety Authority, EFSA (2009). Species-specific welfare aspects of the main systems of stunning and killing of farmed fish: rainbow trout. Scientific Opinion of the Panel on Animal Health and Welfare. EFSA Journal 1013, 1-55. European Union (1997). Treaty of Amsterdam. Office for Official Publications of the European Communities, Luxembourg. Ewing, R.D. & Ewing, S.K. (1995). Review of the effects of rearing density on the survival to adulthood for Pacific salmon. Progressive Fish-Culturist 57, 1-25. Food and Agriculture Organization, FAO (2008). Glossary of Aquaculture. Accessed July 2012. http://glossary/aquaculture/pdf/glossary.pdf. Food and Agriculture Organization, FAO (2011). Small-scale rainbow trout farming. FAO Fisheries and Aquaculture Technical Paper 561, Rome, Italy. Food and Agriculture Organization, FAO (2012a). The state of world fisheries and aquaculture. Food and Agriculture Organization of the United Nations. Rome, Italy. Food and Agriculture Organization, FAO (2012b). Cultured aquatic species information program. Accessed July 2012. http://www.fao.org/fishery/culturedspecies/Oncorhynchus_mykiss/en.

126

Farbridge, K.J. & Leatherland, J.F., (1992a). Plasma growth hormone levels in fed and fasted rainbow trout (Oncorhynchus mykiss) are decreased following handling stress. Fish Physiology and Biochemistry 10, 67–73. Farbridge, K.J. & Leatherland, J.F., (1992b). Temporal changes in plasma thyroid hormone, growth hormone and free fatty acid concentrations, and hepatic 5'-monodeiodinase activity, lipid and protein content during chronic fasting and re-feeding in rainbow trout (Oncorhynchus mykiss). Fish Physiology and Biochemistry 10, 245–257. Fast, M.D., Hosoya, S., Johnson, S.C., Alfonso, L.O.B. (2008). Cortisol response and immune-related effects of Atlantic salmon (Salmo salar Linnaeus) subjected to short- and long- term stress. Fish & Shellfish Immunology 24, 194-204. Farmed Animal Welfare Council, FAWC (1996). Report on the Welfare of Farmed Fish. FAWC Report 2765. Surbiton, Surrey, UK. Farmed Animal Welfare Council, FAWC (2009). Five Freedoms. Accessed August 2012. http://www.fawc.org.uk/freedoms.htm. Fennema, O.R. (1985). Water and ice, in: Fennema, O. R. (Ed.), Food Chemistry. Second Edition Marcel Dekker, Inc. New York, pp. 23-68. Figueiredo-Garutti, M.L., Navarro, I., Capilla, E., Souza, R., Moraes, G., Gutiérrez, J., Vicentini-Paulinoe, M.L.M. (2002). Metabolic changes in Brycon cephalus (Teleostei, Characidae) during post-feeding and fasting. Comparative Biochemistry and Physiology A 132, 467-476. Figueroa, R.I., Rodriguez-Sabaris, R., Aldegunde, M., Soengas, J.L. (2000). Effects of food deprivation on 24 h-changes in brain and liver carbohydrate and ketone body metabolism of rainbow trout. Journal of Fish Biology 57, 631–646. Flodmark, L.E.W., Urke, H.A., Halleraer, J.H., Arnekleiv, J.V., Vøllestad, L.A., Poleo, A.B.S. (2001). Cortisol and glucose responses in juvenile brown trout subjected to a fluctuating flow regime in an artificial stream. Journal of Fish Biology 60, 238-248. Foss, P., Storebakken, T., Schiedt, K., Liaaen-Jensen, S., Austreng, E., Streiff, K. (1984). Carotenoids in diets for salmonids I. Pigmentation of rainbow trout with the individual optical isomers of astaxanthin in comparison with canthaxanthin. Aquaculture 41, 213–226. Foss, A., Imsland, A., Stefansson, S.O., Norberg, B., Vikingstad, E., Pedersen, S., Sandvik, T., Roth, B. (2009). Compensatory growth in Atlantic halibut: effect of starvation and subsequent feeding on growth, maturation, feed utilization and flesh quality. Aquaculture 290, 304–310. Foster, G.D. & Moon, T.W. (1991). Hypometabolism with fasting in the yellow perch (Perca flavescens): a study of enzymes, hepatocyte metabolism and tissue size. Physiological Zoology 64, 259-275.

127

From, J. & Rasmussen, G. (1991). Growth of rainbow trout, Oncorhynchus mykiss (Walbaum, 1792) related to egg size and temperature. Dana 9, 31-38. Fryer, J.L. & Lederis, K., (1986). Control of corticotropin secretion in teleost fishes. American Zoologist 26, 1017-1026. FSBI (Fisheries Society of the British Isles), (2002). Fish Welfare. Briefing Report 2. Granta Information Systems. Accessed June 2012. http://www.le.ac.uk/biology/fsbi/welfare.pdf. Fry, F.E.J. (1971). The effect of environmental factors on the physiology of ﬁsh, in W.S. Hoar, W.S. & Randall, D.J. (Eds.), Fish Physiology. Vol. VI. Academic Press, New York, pp. 1–98. Furné, M., Morales, A.E., Trenzado, C.E., García-Gallego, M., Hidalgo, M.C., Domezain, A., Sanz, A. (2012). The metabolic effects of prolonged starvation and refeeding in sturgeon and rainbow trout. Journal of Comparative Physiology B 182, 63-76. Gallaugher, P. & Farrell, A.P. (1998). Hematocrit and blood oxygen-carrying capacity, in: Perry, S.F. & Tufts, B.L. (Eds.), Fish Respiration. Academic Press, New York, USA, pp. 185-222. Gamperl, A.K., Vijayan, M.M., Boutilire, R.G. (1994). Epinephrine, norepinephrine and cortisol concentrations in cannulated sea-water acclimated rainbow trout (Oncorhynchus mykiss) following black-box confinement and epinephrine injection. Journal of Fish Biology 45, 313-324. Ganesh, A., Dileep, A.O., Shamasundar, B.A., Singh, U. (2006). Gel-forming ability of common carp fish (Cyprinus carpio) meat: effect of freezing and frozen storage. Journal of Food Biochemistry 30, 342-361. Gaylord, T.G. & Gatlin III, D.M. (2000). Assessment of compensatory growth in channel catfish Ictalurus punctatus R. and associated changes in body condition indices. Journal of the World Aquaculture Society 31, 326–336. Gill, T.A. (1992). Biochemical and chemical indices of seafood quality, in: Huss, H.H., Jacobsen, M., Liston, J. (Eds.), Quality Assurance in the Fish Industry. Proceedings of an International Conference. Copenhagen, Denmark, August 1991. Elsevier, Amsterdam, 377-388. Gillis, T.E., Ballantyne, J.S. (1996). The effects of starvation on plasma free amino acid and glucose concentrations in lake sturgeon. Journal of Fish Biology 49, 1306-1316. Ginés, R., Palicio, M., Zamorano, M.J., Argüello, A., López, J.L., Afonso, J.M. (2002). Starvation before slaughtering as a tool to keep freshness attributes in gilthead sea bream (Sparus aurata). Aquaculture International 10, 379–389. Giuffrida, A., Pennisi, L., Ziino, G., Fortino, L., Valvo, G., Marino, S., Panebianco, A. (2007). Influence of slaughtering method on some aspects of quality of gilthead seabream and smoked rainbow trout. Veterinary Research Communications 31, 437-446.

128

Grau, R. & Hamm, R. (1957). Ubre das Wasserbindungsvermogen des Saugetiermuskels. II. Mitt.Ueber die Bestimmung der Wasserbindung des muskels. Zeitschrift fur Lebensmittel- Untersuchung und Forchschung 105, 446-460. Greaser, M.L. (1986). Conversion of muscle to meat, in: Bechtel, P.J., (Ed.), Muscle as Food. Academic Press, Orlando, USA, pp. 37-102. Grigorakis, K., Taylor, K.D.A., Alexis, M.N. (2003). Seasonal patterns of spoilage of ice- stored cultured gilthead sea bream (Sparus aurata). Food Chemistry 81, 263–268. Grutter, A.S. & Pankhurst, N.W. (2000). The effects of capture, handling, confinement and ectoparasite load on plasma levels of cortisol, glucose and lactate in the coral reef fish Hemygimnus melapterus. Journal of Fish Biology 57, 391-401. Gutiérrez, J., Pérez, J., Navarro, I., Zanuy, S., Carrillo, M. (1991). Changes in plasma glucagon and insulin associated with fasting in sea bass (Dicentrarchus labrax). Fish Physiology and Biochemistry 9, 107-112. Hannah, G.S. & Pickford, G.E. (1981). Diurnal peaks in the hematocrit and in serum sodium in the killifish, Fundulus heteroclitus, and their absence in serum potassium and chloride: a lack of correlation. Comparative Biochemistry & Physiology A 70, 157-159. Harmon, K.J., Bolinger, M.T., Rodnick, K.J. (2011). Carbohydrate energy reserves and effects of food deprivation in male and female rainbow trout. Comparative Biochemistry and Physiology A 158, 423-431. Harris, J. & Bird, D.J. (2000). Modulation of the ﬁsh immune system by hormones. Veterinary Immunology and Immunopathology, 77, 163–176. Hastein, T. (2004). Animal welfare issues relating to aquaculture, in: Proceedings of the Global Conference on Animal Welfare: An Oie Initiative. Paris, France. pp. 219–227. Hemre, G.I., Lie, O., Lambertsen, G., Sundby, A. (1990). Dietary carbohydrate utilization in cod (Gadus morhua). Hormonal response of insulin, glucagon and glucagon-like peptide to diet and starvation. Comparative Biochemistry and Physiology A 97, 41-44. Hemre, G.I., Lie, O., Sundby, A. (1993). Dietary carbohydrate utilization in cod (Gadus morhua): metabolic responses to feeding and fasting. Fish Physiology and Biochemistry 10, 455-463. Hilton, J.W. (1982). The effect of pre-fasting diet and water temperature on liver glycogen and liver weight in rainbow trout, Salmo gairdneri Richardson, during fasting. Journal of Fish Biology 20, 69–78. Holloway, A.C., Reddy, P.K., Sheridan, M.A., Leatherland, J.F. (1994). Diurnal rhythms of plasma growth hormone, somatostatin, thyroid hormones, cortisol and glucose concentrations in rainbow trout, Oncorhynchus mykiss, during progressive food deprivation. Biological Rhythm Research 25, 415-432.

129

Honikel, K.O. & Fischer, C. (1977). A rapid method for the detection of PSE and DFD porcine muscles. Journal of Food Science 42, 1633-1636. Hontela, A., Rasmussen, J. B., Audet, C., Chevalier, G. (1992). Impaired cortisol stress response in fish from environments polluted by PAHs, PCBs, and mercury. Archives of Environmental Contamination and Toxicology 22, 278-283. Humane Slaughter Association, HSA (2005). Humane Harvesting of Salmon and Trout. (Guidance notes no 5). HSA & CJA. Wheathampstead. UK. Huff-Lonergan, E. & Lonergan, S.M. (2005). Mechanisms of water-holding capacity of meat: The role of postmortem biochemical and structural changes. Meat Science 71, 194-204. Huffman P.A., Arkoosh M.R., Casillas E. (1997). Characteristics of peripheral blood cells from rainbow trout evaluated by particle counter, image analysis, and hemocytometric techniques. Journal of Aquatic Animal Health 9, 239-248. Huntingford, F.A. & Kadri, S. (2008). Welfare and Fish, in: Branson, J. (Ed.), Fish welfare. Blackwell Publishing, Oxford, UK, pp. 19-31. Huss, H.H. (1995). Quality and quality changes in fresh fish. FAO Fisheries Technical paper 348. Rome, Italy. Innovación Sostenible de la Acuicultura Continnetal (IS-AC) (2010). Trayectoria del sector de la acuicultura continental. Una mirada desde 1960 hasta 2010. Accessed September 2012.http://www.isac.es/attachments/098_Trayectoria%20del%20sector%20de%20la%20acuic ultura%20continental.%201960-2010.pdf Iwama, G. K., Vijayan, M. M., Forsyth, R. B., Ackerman, P. A. (1999). Heat shock proteins and physiological stress in fish. American Zoologist, 39, 901-909. Iwama, G. K., Afonso, L. O. B., Vijayan, M. M. (2006). Stress in fishes, in: Evans, D. H. & Claiborne, J. B. (Eds.), The physiology of fishes. Taylor and Francis, USA, pp. 319-342. Jentoft, S., Aastveit, A.H., Torjesen, P.A., Andersen, O. (2005). Effects of stress on growth, cortisol and glucose levels in non-domesticated Eurasian perch (Perca fluviatilis) and domesticated rainbow trout (Oncorhynchus mykiss). Comparative Biochemistry & Physiology A 141, 353-358. Kieffer, J.D. & Tufts, B.L. (1998). Effects of food deprivation on white muscle energy reserves in rainbow trout (Oncorhynchus mykiss): the relationships with body size and temperature. Fish Physiology and Biochemistry 19, 239-245. Koed, A. (2001). The effects of meal size, body size and temperature on gastric evacuation in pikeperch. Journal of Fish Biology, 58, 281-290. Korhonen, R.W., Lanier, T.C., Giesbrecht, F. (1990). An evaluation of simple methods for following rigor development in fish. Journal of Food Science 55, 346-349. Koteng, A., 1992. Markedsundersøkelse norsk laks (Market investigation Norwegian salmon), Prosjekt God Fisk, Bergen, Norway, 165 pp.

130

Kristiansen, T.S. & Juell, J.E. (2002). How can we document the welfare of farmed fish?, in: Proceedings of the ‘‘Aquaculture Europe 2002: Sea Farming Today and Tomorrow’’. Special Publication n. 32, pp. 285–286. Kristoffersen, S., Tobiassen, T., Steinsund, V., Olsen, R.L. (2006). Slaughter stress, postmortem muscle pH and rigor development in farmed Atlantic cod (Gadus morhua L.). International Journal of Food Science and Technology 41, 861-864. Lanford, S.E., Adams, T.E., Cech, J.J. (2003). Time of day and water temperature modify the physiological stress response in green sturgeon, Acipenser medirostris. Comparative Biochemistry and Physiology A 135, 291-302. Langston, A.L., Hoare, R., Stefansson, M., Fitzgerald, R., Wergeland, H., Mulcahy, M. (2002). The effect of temperature on non-specific defence parameters of three strains of juvenile Atlantic halibut (Hippoglossus hippoglossus L.). Fish and Shellfish Immunology 12, 61-76. Larsen, D.A., Beckman, B.R., Dickhoff, W.W. (2001). The effect of low temperature and fasting during the winter on metabolic stores and endocrine physiology (insulin, insulin-like growth factor-I, and thyroxine) of coho salmon, Oncorhynchus kisutch. General and Comparative Endocrinology 123, 308-323. Lefevre, F., Bugeon, J., Auperin, B., Aubin, J. (2008). Rearing oxygen level and slaughter stress effects on rainbow trout flesh quality. Aquaculture 284, 81-89. Lemieux, H., Dutil, J.D., Guderley, H., Larocque, R. (2004). Growth, starvation and enzyme activity in white muscle of Atlantic cod: at what point do muscle metabolic capacities change? Marine & Freshwater Behaviour & Physiology 37, 287–294. Liew, H.J., Sinha, A.K., Mauro, N., Diricx, M. (2012). Fasting goldﬁsh, Carassius auratus, and common carp, Cyprinus carpio, use different metabolic strategies when swimming. Comparative Biochemistry and Physiology A 163, 327-335. Lim, C. & Klesius, P.H. (2003). Influence of feed deprivation on hematology, macrophage chemotaxis, and resistance to Edwardsiella ictaluri challenge of channel catfish. Journal of Aquatic Animal Health 15, 13-20. Lines, J.A. & Spence, J. (2012). Safeguarding the welfare of farmed fish at harvest. Fish Physiology and Biochemistry 38, 153-162. Lonergan, S.M., Huff-Lonergan, E., Rowe, J., Kuhlers, D. L., Jungst, S. B. (2001). Selection for lean growth efficiency in Duroc pigs: Influence on pork quality. Journal of Animal Science 79, 2075-2085. Love, R. M. (1975). Variability of Atlantic cod (Gadus morhua) from the northeast Atlantic: a review of seasonal and environmental influences on various attributes of fish. Journal of the Fisheries Research Board of Canada 32, 2333-2342. Love, R.M., Robertson, I., Smith, G.L., Whittle, K.J. (1974). The texture of cod muscle. Journal of Texture Studies 5, 201–212.

131

Love, R.M. & Haq, M.A. (2007). The connective tissues of fish III. the effect of pH on gaping in cod entering rigor mortis at different temperatures. International Journal of Food Science and Technology 5, 241-248. Lowe, T., Ryder, J.M., Carrager, J.F., Wells, R.M.G. (1993). Flesh quality in snapper, Pagrus auratus, affected by capture stress. Journal of Food Science 58, 770–773. Lucas, A. (1996). Physical concepts of bioenergetics, in: Lucas, A. (Ed), Bioenergetics of aquatic animals. English edition, Taylor & Francis, France, pp. 10-20. Lund, V., Coleman, G., Gunnarsson, S., Appleby, M.C., Karkinen, K. (2006). Animal welfare science—Working at the interface between the natural and social sciences. Applied Animal Behaviour Science 97, 37-49. Lymbery, P. (2002). In too deep—the welfare of intensively farmed fish. Compassion in World farming. Accessed June 2012. http://www.ciwf.org.uk/includes/documents/cm_docs/2008/i/in_too_deep_2001.pdf MacIntyre, C.M., Ellis, T., North, B.P., Turnbull, J.F. (2008). The influences of water quality on the welfare of farmed rainbow trout: a review, in: Branson, J., (Ed.), Fish Welfare. Blackwell Publishing, Oxford, UK, pp. 150-184. Martínez, M., Guderley, H., Dutil, J.D., Winger, P.D., He, P., Walsh, S.J. (2003). Condition, prolonged swimming performance and muscle metabolic capacities of cod Gadus morhua. Journal of Experimental Biology 206, 503–511. Martínez-Porchas, M., Martínez-Cordova, L.R., Ramos-Enriquez, R. (2009). Cortisol and glucose: reliable indicators of fish stress? Pan-American Journal of Aquatic Sciences 4, 158- 178. Marx, H., Brunner, B., Weinzierl, W., Hoffman, R., Stolle, A. (1997). Methods of stunning freshwater fish: impact on meat quality and aspects of animal welfare. Zeitschrift fur Lebenschmittel Untersuchung und Forschung 204, 282–286. Maule, A.G. & Schreck, C.B. (1990). Changes in the numbers of leukocytes in immune organs of juvenile coho salmon after acute stress or cortisol treatment. Journal of Aquatic Animal Health 2, 298–304. McMillan, N. & Houlihan, D.F. (1992). Protein synthesis in trout liver is stimulated by both feeding and fasting. Fish Physiology and Biochemistry 10, 23-34. Melingen, G.O., Pettersen, E.F., Wergeland, H.I (2002). Leucocyte populations and responses to immunization and photoperiod manipulation in Atlantic salmon (Salmo salar L.) 0+ smolt. Aquaculture 214, 381-396. Mench, J. (1998). Thirty years after Brambell: whither animal science? Journal of Applied Animal Welfare Science 12, 91-102.

132

Merkin, G.V., Roth, B., Gjerstad, C., Dahl-Paulsen, E., Nortvedt, R. (2010). Effect of pre- slaughter procedures on stress responses and some quality parameters in sea-farmed rainbow trout (Oncorhynchus mykiss). Aquaculture 309, 231-235. Metcalfe, N.B. & Thorpe, J.E. (1992). Anorexia and defended energy levels in overwintering juvenile salmon. Journal of Animal Ecology 61, 175–181. Meton, I., Fernandez, F., Baanante, I.V. (2003). Short- and long-term effects of refeeding on key enzyme activities in glycolysis–gluconeogenesis in the liver of gilthead seabream (Sparus aurata). Aquaculture 225, 99-107. Milligan, C.L. (1996). Metabolic recovery from exhaustive exercise in rainbow trout. Comparative Biochemistry & Physiology A 113, 51–60. Montoya, A., López-Olmeda, J.F., Garayzar, A.B.S., Sánchez-Vázquez, F.J. (2010). Synchronization of daily rhythms of locomotor activity and plasma glucose, cortisol and thyroid hormones to feeding in Gilthead seabream (Sparus aurata) under a light–dark cycle. Physiology & Behavior 101, 101-107. Mood, A. & Brooke, P., (2012). Estimating the number of farmed fish killed in global aquaculture each year. Fishcount Publication. Accessed September 2012. http://fishcount.org.uk/published/std/fishcountstudy2.pdf. Moon, T.W., Foster, G.D., Plisetskaya, E. (1989). Changes in peptide hormones and liver enzymes in the rainbow trout deprived of food for six weeks. Canadian Journal of Zoology 67, 2189-2193. Morgan, J.A.W., Pottinger, T.G., Rippon, P. (1993). Evaluation of flow cytometry as a method for quantification of circulating blood cell populations in salmonid fish. Journal of Fish Biology 42, 131-141. Mørkøre, T., Mazo, T.P., Tahirovic, V., Einen, O. (2008). Impact of starvation and handling stress on rigor development and quality of Atlantic salmon (Salmo salar L.). Aquaculture 277, 231–238. Morrow, M.D., Higgs, D., Kennedy, C.J. (2004). The effects of diet composition and ration on biotransformation enzymes and stress parameters in rainbow trout, Oncorhynchus mykiss. Comparative Biochemistry and Physiology C 137, 143-154. Mussa, B. & Gilmour, K.L. (2012). Acid–base balance during social interactions in rainbow trout (Oncorhynchus mykiss). Comparative Biochemistry and Physiology A 162, 177- 184. Narnaware, Y.K. & Baker, B.I. (1996). Evidence that cortisol may protect against the immediate effects of stress on circulating leukocytes in trout. General and Comparative Endocrinology 103, 359-366.

133

Navarro, I., Gutiérrez, J., Planas, J. (1992). Changes in plasma glucagon, insulin and tissue metabolites associated with prolonged fasting in brown trout (Salmo trutta fario) during two different seasons of the year. Comparative Biochemistry and Physiology A 102, 401-407. Navarro, I. & Gutiérrez, J. (1995). Fasting and starvation, in: Hochachka, P.W., Mommsen, T. (Eds.), Biochemistry and molecular biology of fishes. Elsevier Science B.V, pp. 393–434. Nelson, D. L. & Cox, M. M. (2005). Lehninger Principles of Biochemistry. 4th ed. WH Freeman and Co. New York, USA. Newton, K.G. & Gill, C.O. (1981). The microbiology of DFD fresh meats: A review. Meat Science 5, 223-232. Nikki, J., Pirhonen, J., Jobling, M., Karjalainen, J. (2004). Compensatory growth in juvenile rainbow trout, Oncorhynchus mykiss (Walbaum), held individually. Aquaculture 235, 285-296. Nordgreen, J., Garner, J., Janczak, A. Ranheim, B., Muir, W., Horsberg, T. (2009). Thermonociception in fish: Effects of two different doses of morphine on thermal threshold and post-test behaviour in goldfish (Carassius auratus). Applied Animal Behaviour Science 119, 101-107. Nussey, G., van Vuren, J.H.J., du Preez, H.H. (1995). Effect of copper on the haematology and osmoregulation of the Mozambique tilapia, Oreochromis mossambicus (Cichlidae). Comparative and General Biochemistry C 111, 369-380. Ofstad, R., Kidman, S., Myklebust, R., Hermansson, A.M. (1993). Liquid holding capacity and structural changes during heating fish muscle: Cod (Gadus morhua L.) and salmon (Salmo salar). Food Structure 12, 163-174. OIE (2012). Aquatic Animal Health Code. (World Organisation for Animal Health). Accessed July 2012. www.oie.int/en/international-standard-setting/aquatic-code/access-online. Olsen, R.E., Sundell, K., Mayhew, T.M., Myklebust, R., Ringø, E. (2005). Acute stress alters intestinal function of rainbow trout, Oncorhynchus mykiss (Walbaum). Aquaculture 250, 480-495. Olsson, G.B., Olsen, R.L., Ofstad, R. (2003). Post-mortem structural characteristics and water-holding capacity in Atlantic halibut muscle. Food Science and Technology 36, 125–133. Ortiz, R.M., Wade, C.E., Ortiz, C.L. (2001). Effects of prolonged fasting on plasma cortisol and TH in postweaned northern elephant seal pups. American Journal of Physiology- Regulatory, Integrative & Comparative Physiology 280, 790–795. Pagnotta, A., Brooks, L., Milligan, L. (1994). The potential regulatory roles of cortisol in recovery from exhaustive exercise in rainbow trout. Canadian Journal of Zoology 72, 2136- 2146.

134

Parisi, G., Mecatti, M., Lupi, P., Zampacavallo, G., Poli, B.M. (2001). Fish welfare and quality: experimental results on rearing and harvesting practices, in: Proceedings of the Third Congress of the European Society for Agricultural and Food Ethics (EURSAFE 2001) ‘‘Food Safety, Food Quality and Food Ethics’’, 3–5 October 2001, Florence, Italy, pp. 337–340. Pate, E.F. & Brokaw, C.J. (1980). Cross-bridge behaviour in rigor-muscle. Biophysics of Structure & Mechanism 7, 51-54. Pavlidis, M., Greenwood, L., Paalavuo, M., Mölsa, H., Laitinens, J. (1999a). The effects of photoperiod on diel rhythms in serum melatonin, cortisol, glucose and electrolytes in the common dentex, Dentex dentex. General & Comparative Endocrinology 113, 240–250. Pavlidis, M., Paspatis, M., Koistinen, M., Paavola, T., Divanach, P., Kentouri, M. (1999b). Diel rhythms of serum metabolites and thyroid hormones in red porgy held in different photoperiod regimes. Aquaculture International 7, 29-44. Pavlidis, M., Papandroulakis, N., Divanach, P. (2006). A method for the comparison of chromaticity parameters in fish skin: preliminary results for coloration pattern of red skin Sparidae. Aquaculture 258, 211–219. Pereira, C., Vijayan, M.M., Moon, T.W. (1995). In vitro hepatocyte metabolism of alanine and glucose and the reponse to insulin in fed and fasted rainbow trout. Journal of Experimental Zoology 271, 425-443. Pérez-Jiménez, A., Guedes, M.J., Morales, A.E., Oliva-Teles, A. (2007). Metabolic responses to short starvation and refeeding in Dicentrarchus labrax. Effect of dietary composition. Aquaculture 265, 325-335. Pérez-Jiménez, A., Cardenete, G., Hidalgo, M.C., Garcia-Alcazar, A., Abellan, E., Morales, A.E. (2012). Metabolic adjustments of Dentex dentex to prolonged starvation and refeeding. Fish Physiology and Biochemistry 38, 1145-1157. Peterson, B.C. & Small, B.C. (2004). Effects of fasting on circulating IGF-binding proteins, glucose, and cortisol in channel catfish (Ictalurus punctatus). Domestic Animal Endocrinology 26, 231–240. Pettersen, E.F., Bjorlow, I., Hagland, T.J., Wergeland, H.I. (2005). Effect of seawater temperature on leucocyte populations in Atlantic salmon post-smolts. Veterinary Immunology and Immunopathology 106, 65-76. Pickering, A.D., Pottinger, T.G., Christie, P. (1982). Recovery of the brown trout, Salmo trutta L., from acute handling stress: a time-course study. Journal of Fish Biology 20, 229–244. Pickering, A.D. & Pottinger, T.G. (1985). Factors influencing blood cortisol levels of brown trout under intensive culture conditions, in: Lofts, B. & Holms, W.N. (Eds.), Current Trends in Endocrinology. Hong Kong University, pp. 1239–1242. Pickering, A.D. (1998). Stress responses of farmed fish, in: Black, K.D. & Pickering, A.D. (Eds.), Biology of farmed fish. Sheffield Academic Press, Sheffield, UK, pp. 222-255.

135

Pickering, A.D. & Pottinger, T.G. (1995). Biochemical effects of stress, in: Hochachka, P.W. & Mommsen, T.P. (Eds.), Biochemistry and Molecular Biology of Fishes Vol. 5. Elsevier, Amsterdam, The Netherlands, pp. 349-379. Pirhonen, J., Schreck, C.B., Reno, P.W., Ogut, H. (2003). Effect of fasting on feed intake, growth and mortality of Chinook salmon, Oncorhynchus tshawytscha, during an induced Aeromonas salmonicida epizootic. Aquaculture 216, 31–38. Polakof, S., Arjona, F.J., Sangiao-Alvarellos, S., Míguez, J.M., Martín del Río, M.P., Mancera, J.M., Soengas, J.L. (2006). Food deprivation alters osmoregulatory and metabolic responses to salinity acclimation in gilthead sea bream Sparus aurata. Journal of Comparative Physiology B 176, 441-452. Polakof, S., Ceinos, R., Fernandez-Duran, B., Miguez, J.M., Soengas, J.L. (2007). Daily changes in parameters of energy metabolism in brain of rainbow trout: dependence on feeding. Comparative Biochemistry & Physiology A 146, 265–273. Poli, B.M., Parisi, G., Scappini, F., Zampacavallo, G. (2005). Fish welfare and quality as affected by pre-slaughter and slaughter management. Aquaculture International 13, 29-49. Pottinger, T. G. (1998). Changes in blood cortisol, glucose and lactate in carp retained in anglers’ keepnets. Journal of Fish Biology 53, 728-742. Pottinger, T.G. & Carrick, T.R. (1999). A comparison of plasma glucose and plasma cortisol as selection markers for high and low stress-responsiveness in female rainbow trout. Aquaculture 175, 351-363. Pottinger, T.G. (2001). Effects of husbandry stress on flesh quality indicators in fish, in: Kestin, S.C. & Warriss, P.D. (Eds.), Farmed Fish Quality. Fishing News Books, Oxford, pp. 145–160. Pottinger, T.G., Rand-Weaver, M., Sumpter, J.P. (2003). Overwinter fasting and refeeding in rainbow trout: plasma growth hormone and cortisol levels in relation to energy mobilization. Comparative Biochemistry & Physiology B 136, 403–417. Pulsford, A.L., Lemairegony, S., Tomlinson, M., Collingwood, N., Glynn, P.J. (1994). Effects of acute stress on the immune system of the dab, Limanda limanda. Comparative Biochemistry and Physiology C 109, 129–139. Rapp, T., Hallermann, J., Cooke, S.J., Hetz, S.K., Wuertz, S., Arlinghaus, R. (2012). Physiological and behavioural consequences of capture and retention in carp sacks on common carp (Cyprinus carpio L.), with implications for catch-and-release recreational fishing. Fisheries Research 125-126, 57-68. Ratz, H.J. & Lloret, J. (2003). Variation in fish condition between Atlantic cod (Gadus morhua) stocks, the effect on their productivity and management implications. Fisheries Research 60, 369-380.

136

Reddy, P.K. & Leatherland, J.F. (1988). Stress Physiology, in: Leatherland, J.F. & Woo, P.T.K. (Eds.), Fish Diseases, Volume III: Non-infectious disorders. C.A.B. International, pp. 279–302. Reddy, P.K. & Leatherland, J.F. (1995). Influence of the combination of time of feeding and ration level on the diurnal hormone rhythms in rainbow trout. Fish Physiology & Biochemistry 14, 25–36. Reddy, P.K., Vijayan, M.M., Leatherland, J.F., Moon, T.W. (1995). Does RU486 modify hormonal responses to handling stressor and cortisol treatment in fed and fasted rainbow trout? Journal of Fish Biology 46, 341–359. Regost, C., Arzel, J., Cardinal, M., Laroche, M., Kaushik, S.J. (2001). Fat deposition and flesh quality in seawater reared, triploid brown trout (Salmo trutta) as affected by dietary fat levels and starvation. Aquaculture 193, 325–345. Rehulka J., Minarik B., Rehulkova E. (2004). Red blood cell indices of rainbow trout Oncorhynchus mykiss (Walbaum) in aquaculture. Aquaculture Research 35, 529-546. Reid, S.D., Moon, T.W., Perry, S.F. (1992). Rainbow trout hepatocyte beta- adrenoceptors, catecholamine responsiveness, and effects of cortisol. American Journal of Physiology 262, 794-799. Reid, S.G., Bernier, N.J., Perry, S.F. (1998). The adrenergic stress response in fish: control of catecholamine storage and release. Comparative Biochemistry and Physiology C 120, 1-27. Ríos F.S., Kalinin A.L., Rantin F.T. (2002). The effects of long-term deprivation on respiration and haematology of the neotropical fish Hoplias malabaricus. Journal of Fish Biology 61, 85-95. Robb, D.H.F., Kestin, S.C., Warris, P.D. (2000). Muscle activity at slaughter: I. Changes in flesh colour and gaping in rainbow trout. Aquaculture 182, 261-269. Robb, D. (2008). Welfare of fish at harvest, in: Branson, J. (Ed.), Fish welfare. Blackwell Publishing, Oxford, UK, pp. 217-242. Robertson, L., Thomas, P., Arnold, C.R., Trant, J.M. (1987). Plasma cortisol and secondary stress responses of red drum to handling, transport, rearing density, and a disease outbreak. The Progressive Fish Culturist 49, 1-12. Roe, E. & Buller, H. (2008). Marketing farm animal welfare. Welfare quality factsheet. Wageningen, NL: Welfare Quality. Accessed July 2012. (http://www.welfarequality.net/everyone/41858/5/0/22). Rose, J.D. (2002). The neurobehavioral nature of fishes and the question of awareness and pain. Reviews in Fisheries Science 10, 1-38.

137

Roth, B., Slinde, E., Arildsen J. (2006). Pre or post mortem muscle activity in Atlantic salmon (Salmo salar). The effect on rigor mortis and the physical properties of flesh. Aquaculture 257, 504-510. Sangiao-Alvarellos, S., Guzman, J.M., Laiz-Carrion, R., Miguez, J.M., Martin Del Rio, M.P., Mancera, J.M., Soengas, J.L. (2005). Interactive effects of high stocking density and food deprivation on carbohydrate metabolism in several tissues of gilthead sea bream Sparus auratus. Journal of Experimental Zoology 303, 761–775. Schurmann, H., Steffensen, J.F., Lomholt, J.P. (1991). The influence of hypoxia on the preferred temperature of rainbow trout Oncorhynchus mykiss. Journal of Experimental Biology 157, 75–86. Sheridan, M.A. & Mommsen, T.P. (1991). Effccts of nutritional state on in vivo lipid and carbohydrate metabolism of coho salmon, Oncorhynchus kisutch. General and Comparative Endocrinology 81, 473-483. Sherstneva, T.A. & Shabalina, T.A. (1973). Aspects of carbohydrate-lipid metabolism in rainbow trout Salmo irideus. Gibb. following lengthy starvation. Journal of Ichthyology 13, 576-582. Sigholt, T., Erikson, U., Rustad, S., Johansen, S., Nordtvedt, T.S., Seland, A. (1997). Handling stress and storage temperature affect meat quality of farmed-raised Atlantic salmon (Salmo salar). Journal of Food Science 62, 898-905. Simontacchi, C., Poltronieri, C., Carraro, C., Bertotto, D., Xiccato, G., Trocino, A., Radaelli, G. (2008). Alternative stress indicators in sea bass Dicentrarchus labrax, L. Journal of Fish Biology 72, 747-752. Skov, P.V., Larsen, B.K., Frisk, M., Jokumsen, A. (2011). Effects of rearing density and water current on the respiratory physiology and haematology in rainbow trout, Oncorhynchus mykiss at high temperature. Aquaculture 319, 446-452. Small, B.C. (2005). Effect of fasting on nychthemeral concentrations of plasma growth hormone (GH), insulin-like growth factor I (IGF-I), and cortisol in channel catfish (Ictalurus punctatus). Comparative Biochemistry & Physiology B 142, 217–223. Sneddon, L.U. (2003). The evidence for pain in fish: the use of morphine as an analgesic. Applied Animal Behaviour Science 83, 153-162. Sneddon, L.U., Braithwaite, V.A., Gentle, M.J. (2003). Do fishes have nocioceptors? Evidence for the evolution of vertebrate sensory system. Proceedings of the Royal Society of London Series B-Biological Sciences 270, 1115-1121. Soengas, J.L., Strong, E.F., Fuentes, J., Veira, J.A.R., Andres, M.D. (1996). Food deprivation and refeeding in Atlantic salmon, Salmo salar: effects on brain and liver carbohydrate and ketone bodies metabolism. Fish Physiology and Biochemistry 15, 491–511.

138

Soengas, J.L., Polakof, S., Chen, X., Sangiao-Alvarellos, S., Moon, T.W. (2006). Glucokinase and hexokinase expression and activities in rainbow trout tissues: changes with food deprivation and refeeding. American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 291, 810-821. Squires, E.J. (2003). Effects on animal behaviour, health and welfare, in: Applied Animal Endocrinology. CAB International, Wallingford, UK, pp. 192-229. Storebakken, T., Hung, S.S.O., Calvert, C.C., Plisetskaya, E.M. (1991). Nutrient partitioning in rainbow trout at different feeding rates. Aquaculture 96, 191-203. Suárez, M. D., Cervera, M.A.R., Abellan, E., Morales, A.E., Gallego, M.G., Cardenete, G. (2010). Influence of starvation on flesh quality of farmed dentex, Dentex dentex. Journal of the World Aquaculture Society 41, 490-505. Sumpter, J.P., Le Bail, P.Y., Pickering, A.D., Pottinger, T.G., Carragher, J.F. (1991). The effect of starvation on growth and plasma growth hormone concentrations of rainbow trout, Oncorhynchus mykiss. General & Comparative Endocrinology 83, 94–102. Swatland, H. (2003). Ellipsometry across isolated muscle fibres indicates a refractive contribution to paleness in pork. Meat Science 63, 463-467. Thomas, P. M., Pankhurst, N. W., Bremner, H. A. (1999). The effect of stress and exercise on post-mortem biochemistry of Atlantic salmon and rainbow trout. Journal of Fish Biology 54, 1177–1196. Toldra, F. (2003). Muscle foods: water, structure and functionality. Food Science and Technology International 9, 173–177. Tomlinson, N., Geiger, S.E., Dollinger, E. (1965). Chalkiness in halibut in relation to muscle pH and protein denaturation. Journal of the Fisheries Research Board of Canada 22, 653-663. Tort, Ll. (2011). Stress and immune modulation in ﬁsh. Developmental and Comparative Immunology 35, 1366-1375. Trenzado, C.E., Carrick, T.R., Pottinger, T.G. (2003). Divergence of endocrine and metabolic responses to stress in two rainbow trout lines selected for differing cortisol responsiveness to stress. General and Comparative Endocrinology 133, 332-340. Tripathi, G. & Verma, P. (2003). Starvation-induced impairment of metabolism in a freshwater catfish. Zeitschrift Fur Naturforschung C 58, 446–451. Usher, M.L., Talbot, C., Eddy, F.B. (1991). Effects of transfer to seawater on growth and feeding in Atlantic salmon smolts (Salmo salar, L.). Aquaculture 94, 309-326. Valenzuela, A.E., Silva, V.M., Klempau, A.E. (2008). Effects of different artificial photoperiods and temperatures on haematological parameters of rainbow trout (Oncorhynchus mykiss). Fish Physiology and Biochemistry 34, 159-167.

139

van Raaij, M.T.M., van den Thillart, G.E.E.J.M., Vianen, G.J., Pit, D.S.S., Balm, P.H.M., Steffens, A.B. (1996). Substrate mobilization and hormonal changes in rainbow trout Oncorhynchus mykiss, Walbaum. and common carp Cyprinus carpio L.. during deep hypoxia and subsequent recovery. Journal of Comparative Physiology B 166, 443–452. Varnavsky, V.S., Sakamoto, T., Hirano, T. (1995). Effects of premature seawater transfer and fasting on plasma growth hormone levels of yearling coho salmon (Oncorhynchus kisutch) parr. Aquaculture 135, 141–145. Velarde, A. & Dalmau, A. (2012). Animal welfare assessment at slaughter in Europe: Moving from inputs to outputs. Meat Science 92, 244-251. Vijayan, M. M. & Moon, T. W. (1994). The stress response and the plasma disappearance of corticosteroid and glucose in a marine teleost, the sea raven. Canadian Journal of Zoology, 72, 379–386. Vijayan, M.M., Morgan, J.D., Sakamoto, T., Grau, E.G., Iwama, G.K. (1996). Food- deprivation affects seawater acclimation in tilapia: hormonal and metabolic changes. Journal of Experimental Biology 199, 2467-2475. Wall, A.J. (2001). Ethical considerations in the handling and slaughter of farmed fish, in: Kestin, S.C. & Warriss, P.D. (Eds.), Farmed Fish Quality. Fishing News Books, Oxford, UK, pp. 108–115. Wang, D., Tang, J., Correia, L.R., Gill, T.A. (1998). Postmortem changes of cultivated Atlantic salmon and their effects on salt uptake. Journal of Food Science 63, 634-637. Waring, C. P., Stagg, R. M., Poxton, M. G. (1996). Physiological responses to handling in the turbot. Journal of Fish Biology 48, 161–173. Warriss, P.D. & Brown, S.N. (1987). The relationships between initial pH, reflectance and exudation in pig muscle. Meat Science 20, 65–74. Warriss, P.D. (1996). Instrumental measurement of colour, in: Taylor, S.A., Raimundo, A., Severini, M., Smulders, F.J.M., (Eds.), Meat quality and meat packaging. ECCEAMST, Utrecht, The Netherlands, pp. 221–231. Watabe, S., Kamal, M., Hashimoto, K. (1991). Postmortem changes in ATP, creatine phosphate and lactate in sardine muscle. Journal of Food Science 56, 151-154. Wathne, E. (1995). Strategies for directing slaughter quality of farmed Atlantic salmon (Salmo salar) with emphasis on diet composition and fat deposition. Dr. Scient. Thesis 1995:6, Agricultural University of Norway, As, Norway, 230 pp. Weber, T.E. & Bosworth, B.G. (2005). Effects of 28 day exposure to cold temperature or feed restriction on growth, body composition, and expression of genes related to muscle growth and metabolism in channel catfish. Aquaculture 246, 483–492.

140

Wedemeyer, G.A., Barton, B.A., McLeay, D.J. (1990). Stress and acclimation, in: Schreck, C.B. & Moyle, P.B. (Eds.), Methods for Fish Biology. MD: American Fisheries Society, Bethesda, USA, pp. 491-527. Wedemeyer, G.A., (1996). Physiology of fish in intensive culture systems. Chapman & Hall, London, UK. Wedemeyer, G. (1997). Effects of rearing conditions on the health and physiological quality of fish in intensive culture, in: Iwama, G., Pickering, A., Sumpter, J., Schreck, C. (Eds.), Fish Stress and Health in Aquaculture. Cambridge University Press, Cambridge, pp. 35-72. Wendelaar Bonga, S.E. (1997). The stress response in fish. Physiological Reviews 77, 591– 625. Wilkinson, R.J., Paton, N., Porter, M.J.R. (2008). The effects of pre-harvest stress and harvest method on the stress response, rigor onset, muscle pH and drip loss in barramundi (Lates calcarifer). Aquaculture 282, 26-32. Witeska, M. (2005). Stress in ﬁsh: hematological and immunological effects of heavy metals. Electronic Journal of Ichthyology 1, 35–41. Young, M.K., Wilkinson, R.A., Phelps III, J.M., Griffith, J.S. (1997). Contrasting movement and activity of large brown trout and rainbow trout in Silver Creek, Idaho. Great Basin Naturalist 57, 238-244.

141