TY - JOUR
AU1 - Vitali, A.
AU2 - Lana, E.
AU3 - Amadori, M.
AU4 - Bernabucci, U.
AU5 - Nardone, A.
AU6 - Lacetera, N.
AB - Abstract The study was based on data collected during 5 yr (2003–2007) and was aimed at assessing the effects of the month, slaughter house of destination (differing for stocking density, openings, brightness, and cooling device types), length of the journey, and temperature–humidity index (THI) on mortality of heavy slaughter pigs (approximately 160 kg live weight) during transport and lairage. Data were obtained from 24,098 journeys and 3,676,153 pigs transported from 1,618 farms to 3 slaughter houses. Individual shipments were the unit of observation. The terms dead on arrival (DOA) and dead in pen (DIP) refer to pigs that died during transport and in lairage at the abattoir before slaughtering, respectively. These 2 variables were assessed as the dependent counts in separate univariate Poisson regressions. The independent variables assessed univariately in each set of regressions were month of shipment, slaughter house of destination, time traveled, and each combination of the month with the time traveled. Two separate piecewise regressions were done. One used DOA counts within THI levels over pigs transported as a dependent ratio and the second used DIP counts within THI levels over pigs from a transport kept in lairage as a dependent ratio. The THI was the sole independent variable in each case. The month with the greatest frequency of deaths was July with a risk ratio of 1.22 (confidence interval: 1.06–1.36; P < 0.05) and 1.27 (confidence interval: 1.06–1.51; P < 0.05) for DOA and DIP, respectively. The lower mortality risk ratios for DOA and DIP were recorded for January and March (P < 0.05). The aggregated data of the summer (June, July, and August) versus non-summer (January, March, September, and November) months showed a greater risk of pigs dying during the hot season when considering both transport and lairage (P < 0.05). The mortality risk ratio of DIP was lower at the slaughter house with the lowest stocking density (0.64 m2/100 kg live weight), large open windows on the roof and sidewalls, low brightness (40 lx) lights, and high-pressure sprinklers as cooling devices. The mortality risk ratio of DOA increased significantly for journeys longer than 2 h, whereas no relationship was found between length of transport and DIP. The piecewise analysis pointed out that 78.5 and 73.6 THI were the thresholds above which the mortality rate increased significantly for DOA and DIP, respectively. These results may help the pig industry to improve the welfare of heavy slaughter pigs during transport and lairage. INTRODUCTION The severity of transport stress depends on aspects of loading and unloading as well as length of the journey, stocking density, group social hierarchies, genotype, and climatic condition (Warriss, 1998; Nanni Costa, 2009). The mortality rate (i.e., number of pigs died/number of pigs transported) is a simple, objective, and useful parameter to quantify pigs' welfare during market procedures. Even if an individual does not die, its welfare is likely to be reduced when overall mortality rate increases, which is usually determined, at least in part, by transport conditions (Warriss, 1998). Pig market losses are a topic that is debated by the public because welfare of animals is very important to consumers. Additionally, stakeholders within the swine industry and regulatory agencies treat such losses with concern because of the economic impact. Different studies have been aimed at investigating the effects of season, air temperature, and length of the journey on in-transit market pig losses (Werner et al., 2007; Averos et al., 2008; Sutherland et al., 2009). Conversely, pig losses during lairage received little attention (Weeks, 2008). Finally, to the best of our knowledge, 1 study has been performed to assess the relationship between market pig losses and temperature–humidity index (THI), an index that combines the simultaneous effect of temperature and relative humidity (Fitzgerald et al., 2009). Interestingly, all these studies referred to market pigs of 100 to 120 kg live weight (LW), whereas no information was available for heavy pigs slaughtered at 160 kg LW, a typical slaughter weight to obtain meat suitable for producing Italian crude Parma's ham. Therefore, the study reported herein was aimed at assessing the effects of the month of the year, length of the journey, slaughter house of destination, and THI on mortality of heavy slaughter pigs during transport and lairage at commercial abattoirs. MATERIALS AND METHOD All journeys considered in the present study were done in respect of regulations in force for the protection of animals during transport, slaughter, and related operations (European Council, 2005). Pigs, Transports, and Slaughter Houses Data This retrospective observational study was performed by analyzing a database collected during 5 yr (2003–2007) and included 24,098 journeys and 3,676,153 heavy pigs (approximately 160 kg LW) transported by truck from 1,618 farms to 3 different slaughter houses indicated as A, B, and C. Information on truck type were not available, but the personnel of the abattoirs indicated that the most common trucks were 3-deck type that commonly load approximately 140 pigs. The journeys within 2003 through 2007 that went to the 3 abattoirs were the data we had access to. The 3 slaughter houses differed for some structural and management items, which included the floor area per head, brightness, presence of natural or fan-assisted ventilation, and cooling devices (Table 1). Table 1. Structural and management factors adopted in lairage pens for the 3 slaughter houses involved in the study   Slaughter house  Structural and management factors  A  B  C  Ventilation  Two opening windows on the roof and large openings on 1 side of the barn  Nine opening windows on 1 side of the barn and forced air extraction device  Four opening windows on the roof and large opening windows on 1 side on the barn  Floor  Concrete floor  Concrete floor  Concrete floor  Brightness, lx  60  70  40  Stocking density, m2/100 kg live weight  0.46  0.56  0.64  Box dimensions  Different size and not modular  Completely modular  Semimodular  Water supplied  By flow from the top  By troughs  By nipples  Cooling system  By water drops  By high-pressure sprinklers  By high-pressure sprinklers    Slaughter house  Structural and management factors  A  B  C  Ventilation  Two opening windows on the roof and large openings on 1 side of the barn  Nine opening windows on 1 side of the barn and forced air extraction device  Four opening windows on the roof and large opening windows on 1 side on the barn  Floor  Concrete floor  Concrete floor  Concrete floor  Brightness, lx  60  70  40  Stocking density, m2/100 kg live weight  0.46  0.56  0.64  Box dimensions  Different size and not modular  Completely modular  Semimodular  Water supplied  By flow from the top  By troughs  By nipples  Cooling system  By water drops  By high-pressure sprinklers  By high-pressure sprinklers  View Large Table 1. Structural and management factors adopted in lairage pens for the 3 slaughter houses involved in the study   Slaughter house  Structural and management factors  A  B  C  Ventilation  Two opening windows on the roof and large openings on 1 side of the barn  Nine opening windows on 1 side of the barn and forced air extraction device  Four opening windows on the roof and large opening windows on 1 side on the barn  Floor  Concrete floor  Concrete floor  Concrete floor  Brightness, lx  60  70  40  Stocking density, m2/100 kg live weight  0.46  0.56  0.64  Box dimensions  Different size and not modular  Completely modular  Semimodular  Water supplied  By flow from the top  By troughs  By nipples  Cooling system  By water drops  By high-pressure sprinklers  By high-pressure sprinklers    Slaughter house  Structural and management factors  A  B  C  Ventilation  Two opening windows on the roof and large openings on 1 side of the barn  Nine opening windows on 1 side of the barn and forced air extraction device  Four opening windows on the roof and large opening windows on 1 side on the barn  Floor  Concrete floor  Concrete floor  Concrete floor  Brightness, lx  60  70  40  Stocking density, m2/100 kg live weight  0.46  0.56  0.64  Box dimensions  Different size and not modular  Completely modular  Semimodular  Water supplied  By flow from the top  By troughs  By nipples  Cooling system  By water drops  By high-pressure sprinklers  By high-pressure sprinklers  View Large The farms as well as the 3 slaughter houses were located in northern Italy and reported in Fig. 1 (Google Earth, 2012). All pigs considered in the study were slaughtered to obtain meat suitable for producing the typical Italian crude Parma's ham or cooked ham if defects were found. For each of the 24,098 journeys within the original database, the following information was collected: date of the journey, farm code and municipality, destination (slaughter house), number of pigs transported, number of pigs dead on arrival (DOA), and number of pigs dead in pen (DIP) at the abattoir before the slaughter operations. Figure 1. View largeDownload slide Map of the studied area. The black squares indicate the 82 weather stations used to acquire temperature and relative humidity data, which were used for calculation of the temperature–humidity index. The 3 white triangles indicate the slaughter houses considered in the study and reported as A, B, and C. The black dots indicate the 477 municipalities around which the 1,618 pig farms were located. Figure 1. View largeDownload slide Map of the studied area. The black squares indicate the 82 weather stations used to acquire temperature and relative humidity data, which were used for calculation of the temperature–humidity index. The 3 white triangles indicate the slaughter houses considered in the study and reported as A, B, and C. The black dots indicate the 477 municipalities around which the 1,618 pig farms were located. This database was used to assess the effect of the month and slaughter house for DOA and DIP. Time Traveled The original database was integrated with the length of the journey. The duration of the journeys was estimated by Google Maps where farm municipality center was set as departure point and the delivery plant as arrival point (Google Maps, 2013). The journeys with no correspondence between municipality name and Italian municipality database (Istat, 2012) were eliminated. Therefore, 21,778 journeys were measured as to duration and were categorized as follows: <1 h (11,812 journeys and 140.6 ± 25.9 pigs/load), 1 to 2 h (7,840 journeys and 138.9 ± 24.1 pigs/load), 2 to 3 h (1,534 journeys and 140.1 ± 19.3 pigs/load), and 3 to 4 h (592 journeys and 136.6 ± 22.7 pigs/load). This database was used to assess the effect of the length of the journey and the combined effect of the length and month of the journey for DOA and DIP. Temperature–Humidity Index Data Because of lack of some weather data, out of 24,098 journeys within the original database, only 14,316 journeys representing 1,209 farms could have environmental information assessed. The database relative to these journeys was therefore integrated with THI data. For the analysis of mortality associated with transport (DOA), the THI assigned to each journey was the mean of the THI calculated at the weather stations closest to the farm (municipality) and to the abattoir the day of transport. For the analysis of mortality during lairage (DIP), the THI assigned to each journey was the THI value calculated only at the weather station closest to the abattoir the day of delivery. Association of farm (municipality) and abattoir to the nearest weather station was performed by transforming latitude and longitude of farms' municipality, slaughter houses, and weather stations in radians. The distances between farms municipality or slaughter houses and weather stations were therefore calculated by the formula  in which latA and latB were latitude of A and B expressed in radians and lonA and lonB were the longitude of A and B expressed in radians. The term 6,371 km referred to the radius of the earth. For each weather station enrolled in the study, the THI was calculated on the basis of average values of ambient temperature (AT; expressed as °C) and relative humidity (RH; expressed as fraction of unit) by using the following formula (Kelly and Bond, 1971):  The average monthly THI is shown in Fig. 2. The original database including the time traveled was thus merged with THI data. Figure 2. View largeDownload slide Overall average monthly values (±SD) of temperature–humidity index (THI) recorded in the studied area during 5 yr (2003–2007). Figure 2. View largeDownload slide Overall average monthly values (±SD) of temperature–humidity index (THI) recorded in the studied area during 5 yr (2003–2007). Data were grouped by THI, and for each value of THI mortality rate (MR) was calculated as  in which the THI ranged between 29 and 84. This database was used in the THI piecewise analysis. Statistical Analysis Mortality Counts. Variations of DOA and DIP in relation to month, grouped summer (June, July, and August) and non-summer (January, March, September, and November) months, slaughter house of destination, time traveled, and each combination of the month with the time traveled were assessed by univariate Poisson regression model. The models were set for each group of data in relation to transport or lairage. Individual shipments were the unit of observation. The dependent variables in the models were the outcome counts of DOA or DIP; the pigs transported or kept in lairage were used as the denominator counts. The reference group was the total outcome counts of DOA or DIP as dependent variable and the total of pigs shipped or kept in lairage as denominator counts. Pearson and G statistic tests were performed to evaluate the goodness-of-fit. Regression coefficient β was transformed into risk ratio (RR) by taking the exponent of the coefficient as follows: RR = exp(β). A RR greater than 1 means that the determinant increases the incidence rate of DOA and DIP. A RR between 0 and 1 means that the incidence rate of DOA and DIP is reduced by the determinant. A RR is statistically significant when the 95% confidence interval (CI) does not include the unit. The analyses were performed with NCSS software for Windows (Hintze, 2007). Temperature–Humidity Index Piecewise Analysis. The analysis of the relationships between DOA or DIP and THI were performed by a 2-phase linear regression procedure (Nickerson et al., 1989), which detects an inflection point, if one exists, in the relationship between average THI as the independent variable and MR as the dependent variable. The model used was  The analyses were performed with Statistica 7.0 software (StatSoft, Inc. 2004). RESULTS Descriptive statistics of the study is reported in Table 2. Table 2. Descriptive statistic of the study1   Study  Data  Month 2  Time traveled 3  THI4  Journeys, n  24,098  21,778  14,316  In-transit pigs, n  3,676,153  3,346,350  2,156,322  Pigs in pen, n  3,674,992  3,345,283  2,155,658  Farms of pig origin, n  1,618  1,431  1,209  Pigs load/journey, n  140 ± 245  139.8 ± 24.85  139.6 ± 25.25  Minimum pigs load/journey, n  50  50  50  Maximum pigs load/journey, n  175  175  175  Pigs dead on arrival, n  1,161  1,067  664  Pigs dead in pen, n  676  650  399  Average journey duration (min) or THI  –  66.9 ± 42.35  62 ± 125  Minimum journey duration (min) or THI  –  5  29  Maximum journey duration (min) or THI  –  240  84  Weather stations consulted, n  –  –  82  Distance between weather stations and municipality of pig farms, km  –  –  14.07 ± 9.125  Distance between weather stations and slaughter houses, km  –  –  10.70 ± 1.65    Study  Data  Month 2  Time traveled 3  THI4  Journeys, n  24,098  21,778  14,316  In-transit pigs, n  3,676,153  3,346,350  2,156,322  Pigs in pen, n  3,674,992  3,345,283  2,155,658  Farms of pig origin, n  1,618  1,431  1,209  Pigs load/journey, n  140 ± 245  139.8 ± 24.85  139.6 ± 25.25  Minimum pigs load/journey, n  50  50  50  Maximum pigs load/journey, n  175  175  175  Pigs dead on arrival, n  1,161  1,067  664  Pigs dead in pen, n  676  650  399  Average journey duration (min) or THI  –  66.9 ± 42.35  62 ± 125  Minimum journey duration (min) or THI  –  5  29  Maximum journey duration (min) or THI  –  240  84  Weather stations consulted, n  –  –  82  Distance between weather stations and municipality of pig farms, km  –  –  14.07 ± 9.125  Distance between weather stations and slaughter houses, km  –  –  10.70 ± 1.65  1Data referred to the years 2003 to 2007 and to northern Italy. 2The study was aimed at investigating the monthly risk of market pig losses during transport and lairage and data referred to all journeys present in the original database. 3The study was aimed at investigating the effect of transport duration on pig losses during transport and lairage and data referred only to the journeys present in the original database for which it was possible to estimate the time traveled. 4The study was aimed at investigating the effect of temperature–humidity index (THI) on pig losses during transport and lairage and data referred only to the journeys present in the original database for which it was possible to associate weather data. 5Means values ± SD. View Large Table 2. Descriptive statistic of the study1   Study  Data  Month 2  Time traveled 3  THI4  Journeys, n  24,098  21,778  14,316  In-transit pigs, n  3,676,153  3,346,350  2,156,322  Pigs in pen, n  3,674,992  3,345,283  2,155,658  Farms of pig origin, n  1,618  1,431  1,209  Pigs load/journey, n  140 ± 245  139.8 ± 24.85  139.6 ± 25.25  Minimum pigs load/journey, n  50  50  50  Maximum pigs load/journey, n  175  175  175  Pigs dead on arrival, n  1,161  1,067  664  Pigs dead in pen, n  676  650  399  Average journey duration (min) or THI  –  66.9 ± 42.35  62 ± 125  Minimum journey duration (min) or THI  –  5  29  Maximum journey duration (min) or THI  –  240  84  Weather stations consulted, n  –  –  82  Distance between weather stations and municipality of pig farms, km  –  –  14.07 ± 9.125  Distance between weather stations and slaughter houses, km  –  –  10.70 ± 1.65    Study  Data  Month 2  Time traveled 3  THI4  Journeys, n  24,098  21,778  14,316  In-transit pigs, n  3,676,153  3,346,350  2,156,322  Pigs in pen, n  3,674,992  3,345,283  2,155,658  Farms of pig origin, n  1,618  1,431  1,209  Pigs load/journey, n  140 ± 245  139.8 ± 24.85  139.6 ± 25.25  Minimum pigs load/journey, n  50  50  50  Maximum pigs load/journey, n  175  175  175  Pigs dead on arrival, n  1,161  1,067  664  Pigs dead in pen, n  676  650  399  Average journey duration (min) or THI  –  66.9 ± 42.35  62 ± 125  Minimum journey duration (min) or THI  –  5  29  Maximum journey duration (min) or THI  –  240  84  Weather stations consulted, n  –  –  82  Distance between weather stations and municipality of pig farms, km  –  –  14.07 ± 9.125  Distance between weather stations and slaughter houses, km  –  –  10.70 ± 1.65  1Data referred to the years 2003 to 2007 and to northern Italy. 2The study was aimed at investigating the monthly risk of market pig losses during transport and lairage and data referred to all journeys present in the original database. 3The study was aimed at investigating the effect of transport duration on pig losses during transport and lairage and data referred only to the journeys present in the original database for which it was possible to estimate the time traveled. 4The study was aimed at investigating the effect of temperature–humidity index (THI) on pig losses during transport and lairage and data referred only to the journeys present in the original database for which it was possible to associate weather data. 5Means values ± SD. View Large Month, Slaughter House, and Time Traveled Poisson Count Analysis The analysis of the risk indicated an association between month of the year, slaughter house of destination (for DIP analysis only), and time traveled with the deaths recorded during transport and lairage. Dead on Arrival. When the original database was considered, with 3,676,153 transported pigs and 1,161 of DOA recorded (Table 2), July was the month with the greatest risk of death with a RR of 1.22 (CI: 1.06–1.36; P < 0.05). Conversely, January and March were the months with the lowest risk to die (P < 0.05; Table 3). The RR value of pig losses considering the summer months altogether (June, July, and August) was 1.10 (CI: 1.01–1.22; P < 0.05), indicating a greater risk for pigs to die during summer (data not shown). The RR value for the aggregated data of the non-summer months (January, March, September, and November) was 0.87 (CI: 0.80–0.96; P < 0.05), indicating a lower risk for pigs to die during non-summer months (data not shown). Table 3. In-transit and lairage mortality risk ratios (RR) in relation to the month of the year In transit  Month  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  January  327,143  74  –0.340  0.71  0.56–0.90  <0.05  March  330,582  59  –0.570  0.57  0.43–0.72  <0.05  June  669,052  211  –0.001  1.00  0.86–1.13  >0.05  July  653,909  252  0.198  1.22  1.06–1.36  <0.05  August  640,354  224  0.102  1.11  0.96–1.25  >0.05  September  705,310  249  0.111  1.12  0.98–1.25  >0.05  November  349,803  92  –0.183  0.83  0.67–1.01  >0.05  Total  3,676,153  1,161          Lairage  Month  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  January  327,069  39  –0.437  0.65  0.46–0.89  <0.05  March  330,523  35  –0.552  0.58  0.41–0.80  <0.05  June  668,841  129  0.046  1.05  0.86–1.26  >0.05  July  653,657  153  0.240  1.27  1.06–1.51  <0.05  August  640,130  131  0.106  1.11  0.92–1.34  >0.05  September  705,061  133  0.024  1.02  0.85–1.23  >0.05  November  349,711  56  –0.139  0.87  0.66–1.14  >0.05  Total  3,674,992  676          In transit  Month  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  January  327,143  74  –0.340  0.71  0.56–0.90  <0.05  March  330,582  59  –0.570  0.57  0.43–0.72  <0.05  June  669,052  211  –0.001  1.00  0.86–1.13  >0.05  July  653,909  252  0.198  1.22  1.06–1.36  <0.05  August  640,354  224  0.102  1.11  0.96–1.25  >0.05  September  705,310  249  0.111  1.12  0.98–1.25  >0.05  November  349,803  92  –0.183  0.83  0.67–1.01  >0.05  Total  3,676,153  1,161          Lairage  Month  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  January  327,069  39  –0.437  0.65  0.46–0.89  <0.05  March  330,523  35  –0.552  0.58  0.41–0.80  <0.05  June  668,841  129  0.046  1.05  0.86–1.26  >0.05  July  653,657  153  0.240  1.27  1.06–1.51  <0.05  August  640,130  131  0.106  1.11  0.92–1.34  >0.05  September  705,061  133  0.024  1.02  0.85–1.23  >0.05  November  349,711  56  –0.139  0.87  0.66–1.14  >0.05  Total  3,674,992  676          1DOA = dead on arrival. 2β = regression coefficients provided by the model. 3CI = confidence interval. The RR is statistically significant when the 95% CI does not include the unit. 4DIP = dead in pen. View Large Table 3. In-transit and lairage mortality risk ratios (RR) in relation to the month of the year In transit  Month  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  January  327,143  74  –0.340  0.71  0.56–0.90  <0.05  March  330,582  59  –0.570  0.57  0.43–0.72  <0.05  June  669,052  211  –0.001  1.00  0.86–1.13  >0.05  July  653,909  252  0.198  1.22  1.06–1.36  <0.05  August  640,354  224  0.102  1.11  0.96–1.25  >0.05  September  705,310  249  0.111  1.12  0.98–1.25  >0.05  November  349,803  92  –0.183  0.83  0.67–1.01  >0.05  Total  3,676,153  1,161          Lairage  Month  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  January  327,069  39  –0.437  0.65  0.46–0.89  <0.05  March  330,523  35  –0.552  0.58  0.41–0.80  <0.05  June  668,841  129  0.046  1.05  0.86–1.26  >0.05  July  653,657  153  0.240  1.27  1.06–1.51  <0.05  August  640,130  131  0.106  1.11  0.92–1.34  >0.05  September  705,061  133  0.024  1.02  0.85–1.23  >0.05  November  349,711  56  –0.139  0.87  0.66–1.14  >0.05  Total  3,674,992  676          In transit  Month  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  January  327,143  74  –0.340  0.71  0.56–0.90  <0.05  March  330,582  59  –0.570  0.57  0.43–0.72  <0.05  June  669,052  211  –0.001  1.00  0.86–1.13  >0.05  July  653,909  252  0.198  1.22  1.06–1.36  <0.05  August  640,354  224  0.102  1.11  0.96–1.25  >0.05  September  705,310  249  0.111  1.12  0.98–1.25  >0.05  November  349,803  92  –0.183  0.83  0.67–1.01  >0.05  Total  3,676,153  1,161          Lairage  Month  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  January  327,069  39  –0.437  0.65  0.46–0.89  <0.05  March  330,523  35  –0.552  0.58  0.41–0.80  <0.05  June  668,841  129  0.046  1.05  0.86–1.26  >0.05  July  653,657  153  0.240  1.27  1.06–1.51  <0.05  August  640,130  131  0.106  1.11  0.92–1.34  >0.05  September  705,061  133  0.024  1.02  0.85–1.23  >0.05  November  349,711  56  –0.139  0.87  0.66–1.14  >0.05  Total  3,674,992  676          1DOA = dead on arrival. 2β = regression coefficients provided by the model. 3CI = confidence interval. The RR is statistically significant when the 95% CI does not include the unit. 4DIP = dead in pen. View Large When the integrated time traveled database was considered, with 3,346,350 transported pigs and 1,067 of DOA recorded (Table 2), the analysis indicated a significantly lower risk to die for very short journeys (<1 h) and pointed out significantly greater DOA for transports longer than 2 h. The RR values were 1.56 (CI: 1.29–1.89; P < 0.05) and 3.09 (CI: 2.46–3.88; P < 0.05) for journeys of 2 to 3 and 3 to 4 h, respectively (Table 4). Table 4. In-transit and lairage mortality risk ratios (RR) in relation to the time traveled In transit  Time traveled, h  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  <1  1,841,533  453  –0.259  0.77  0.69–0.86  <0.05  1–2  1,187,066  416  0.094  1.09  0.98–1.23  >0.05  2–3  236,608  118  0.447  1.56  1.29–1.89  <0.05  3–4  81,143  80  1.128  3.09  2.46–3.88  <0.05  Total  3,346,350  1,067          Lairage  Time traveled, h  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  <1  1,841,080  359  0.003  1.00  0.88–1.14  >0.05  1–2  1,186,650  242  0.048  1.05  0.90–1.21  >0.05  2–3  236,490  38  –0.190  0.83  0.59–1.14  >0.05  3–4  81,063  11  –0.359  0.70  0.38–1.26  >0.05  Total  3,345,283  650          In transit  Time traveled, h  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  <1  1,841,533  453  –0.259  0.77  0.69–0.86  <0.05  1–2  1,187,066  416  0.094  1.09  0.98–1.23  >0.05  2–3  236,608  118  0.447  1.56  1.29–1.89  <0.05  3–4  81,143  80  1.128  3.09  2.46–3.88  <0.05  Total  3,346,350  1,067          Lairage  Time traveled, h  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  <1  1,841,080  359  0.003  1.00  0.88–1.14  >0.05  1–2  1,186,650  242  0.048  1.05  0.90–1.21  >0.05  2–3  236,490  38  –0.190  0.83  0.59–1.14  >0.05  3–4  81,063  11  –0.359  0.70  0.38–1.26  >0.05  Total  3,345,283  650          1DOA = dead on arrival. 2β = regression coefficients provided by the model. 3CI = confidence interval. The RR is statistically significant when the 95% CI does not include the unit. 4DIP = dead in pen. View Large Table 4. In-transit and lairage mortality risk ratios (RR) in relation to the time traveled In transit  Time traveled, h  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  <1  1,841,533  453  –0.259  0.77  0.69–0.86  <0.05  1–2  1,187,066  416  0.094  1.09  0.98–1.23  >0.05  2–3  236,608  118  0.447  1.56  1.29–1.89  <0.05  3–4  81,143  80  1.128  3.09  2.46–3.88  <0.05  Total  3,346,350  1,067          Lairage  Time traveled, h  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  <1  1,841,080  359  0.003  1.00  0.88–1.14  >0.05  1–2  1,186,650  242  0.048  1.05  0.90–1.21  >0.05  2–3  236,490  38  –0.190  0.83  0.59–1.14  >0.05  3–4  81,063  11  –0.359  0.70  0.38–1.26  >0.05  Total  3,345,283  650          In transit  Time traveled, h  Pigs transported, n  DOA,1n  β2  RR  95% CI3  P-value  <1  1,841,533  453  –0.259  0.77  0.69–0.86  <0.05  1–2  1,187,066  416  0.094  1.09  0.98–1.23  >0.05  2–3  236,608  118  0.447  1.56  1.29–1.89  <0.05  3–4  81,143  80  1.128  3.09  2.46–3.88  <0.05  Total  3,346,350  1,067          Lairage  Time traveled, h  Pigs in pens, n  DIP,4n  β2  RR  95% CI3  P-value  <1  1,841,080  359  0.003  1.00  0.88–1.14  >0.05  1–2  1,186,650  242  0.048  1.05  0.90–1.21  >0.05  2–3  236,490  38  –0.190  0.83  0.59–1.14  >0.05  3–4  81,063  11  –0.359  0.70  0.38–1.26  >0.05  Total  3,345,283  650          1DOA = dead on arrival. 2β = regression coefficients provided by the model. 3CI = confidence interval. The RR is statistically significant when the 95% CI does not include the unit. 4DIP = dead in pen. View Large The combined effect of the month of the journey and time traveled indicated an increase of the risk for DOA in correspondence of the hottest months and the longest journeys. In particular, the greatest risk of death (RR = 6.15; CI: 4.24–8.91; P < 0.001) was found for the month of July in association with journeys of 3 to 4 h (Fig. 3). Figure 3. View largeDownload slide Risk ratios (RR) with relative 95% confidence intervals (CI) for pigs found dead on arrival (DOA) in relation to the month of the year and the time traveled (h). Figure 3. View largeDownload slide Risk ratios (RR) with relative 95% confidence intervals (CI) for pigs found dead on arrival (DOA) in relation to the month of the year and the time traveled (h). Dead in Pen. When the original database was considered, with 3,674,992 unloaded pigs and 676 of DIP recorded (Table 2), July was the month with the greatest risk of death with a RR of 1.27 (CI: 1.06–1.51; P < 0.05). Conversely, January and March were the months with the lowest risk (P < 0.05) to die (Table 3). The RR value of pig losses considering the summer months altogether (June, July, and August) was 1.14 (CI: 1.01–1.29; P < 0.05), indicating a greater risk for pigs to die during summer (data not shown). The RR value for the aggregated data of the non-summer months (January, March, September, and November) was 0.83 (CI: 0.74–0.94; P < 0.05), indicating a lower risk for pigs to die during non-summer months (data not shown). As already anticipated above, the DIP differed significantly in relation to the slaughter house of destination. In details, the RR during lairage were 1.95 (CI: 1.72–2.20; P < 0.05) and 1.21 (CI: 1.03–1.40; P < 0.05) for slaughter houses A and B, respectively, and 0.14 (CI: 0.10–0.19; P < 0.05) for pigs delivered to slaughter house C (Table 5). Table 5. Lairage mortality risk ratios (RR) in relation to the slaughter house of destination Lairage  Slaughter house  Pigs in pens, n  DIP,1n  β2  RR  95% CI3  P-value  A  1,172,154  419  0.667  1.95  1.72–2.20  <0.05  B  976,226  217  0.188  1.21  1.03–1.40  <0.05  C  1,526,612  40  –1.949  0.14  0.10–0.19  <0.05  Total  3,674,992  676          Lairage  Slaughter house  Pigs in pens, n  DIP,1n  β2  RR  95% CI3  P-value  A  1,172,154  419  0.667  1.95  1.72–2.20  <0.05  B  976,226  217  0.188  1.21  1.03–1.40  <0.05  C  1,526,612  40  –1.949  0.14  0.10–0.19  <0.05  Total  3,674,992  676          1DIP = dead in pen. 2β = regression coefficients provided by the model. 3CI = confidence interval. The RR is statistically significant when the 95% CI does not include the unit. View Large Table 5. Lairage mortality risk ratios (RR) in relation to the slaughter house of destination Lairage  Slaughter house  Pigs in pens, n  DIP,1n  β2  RR  95% CI3  P-value  A  1,172,154  419  0.667  1.95  1.72–2.20  <0.05  B  976,226  217  0.188  1.21  1.03–1.40  <0.05  C  1,526,612  40  –1.949  0.14  0.10–0.19  <0.05  Total  3,674,992  676          Lairage  Slaughter house  Pigs in pens, n  DIP,1n  β2  RR  95% CI3  P-value  A  1,172,154  419  0.667  1.95  1.72–2.20  <0.05  B  976,226  217  0.188  1.21  1.03–1.40  <0.05  C  1,526,612  40  –1.949  0.14  0.10–0.19  <0.05  Total  3,674,992  676          1DIP = dead in pen. 2β = regression coefficients provided by the model. 3CI = confidence interval. The RR is statistically significant when the 95% CI does not include the unit. View Large When the integrated time traveled database was considered, with 3,345,283 unloaded pigs and 650 of DIP recorded (Table 2), the analysis indicated a lack of association between the time traveled and DIP (Table 4). Temperature–Humidity Index Piecewise Analysis Dead on Arrival. When the database associated with climatic information was considered, with 2,156,322 transported pigs and 664 of DOA recorded (Table 2), the 2-phase regression computed between MR (calculated for DOA) as the dependent variable and THI as the independent variable indicated a break point at 78.5 THI (Fig. 4a). The model provided the slope and intercept of 2 distinct regression lines: MR1 = 1.47827 + 0.02095 × THI (R2 = 0.2772; F1,46 = 17.64; P = 0.0001) below the break point (31 to 78.5 THI range) and MR2 = –109.20 + 1.43013 × THI (R2 = 0.8516; F1,3 = 17.2; P = 0.0254) above the break point (78.5 to 83 THI range). Figure 4. View largeDownload slide In-transit (a) and lairage (b) mortality rates (MR; × 10,000) in relation to mean values of the temperature–humidity index (THI). (a) A break point was detected at 78.5 THI for dead on arrival (DOA) and the model provided slope and intercept of 2 distinct regression lines: MR1 = 1.47827 + 0.02095 × THI below the break point (31 to 78.5 THI range) and MR2 = –109.20 + 1.43013 × THI above the break point (78.5 to 83 THI range). (b) A break point was detected at 73.6 THI for dead in pen (DIP) and the model provided slope and intercept of 2 distinct regression lines MR1 = 0.6997 + 0.01389 × THI below the break point (29 to 73.6 THI range) and MR2 = –42.76398 + 0.6048 × THI above the break point (73.6 to 84 THI range). Figure 4. View largeDownload slide In-transit (a) and lairage (b) mortality rates (MR; × 10,000) in relation to mean values of the temperature–humidity index (THI). (a) A break point was detected at 78.5 THI for dead on arrival (DOA) and the model provided slope and intercept of 2 distinct regression lines: MR1 = 1.47827 + 0.02095 × THI below the break point (31 to 78.5 THI range) and MR2 = –109.20 + 1.43013 × THI above the break point (78.5 to 83 THI range). (b) A break point was detected at 73.6 THI for dead in pen (DIP) and the model provided slope and intercept of 2 distinct regression lines MR1 = 0.6997 + 0.01389 × THI below the break point (29 to 73.6 THI range) and MR2 = –42.76398 + 0.6048 × THI above the break point (73.6 to 84 THI range). Dead in Pen. When the database associated with climatic information was considered, with 2,155,658 transported pigs and 399 of DIP recorded (Table 2), the 2-phase regression computed between MR (calculated for DIP) as the dependent variable and THI as the independent variable indicated a break point at 73.6 THI (Fig. 4b). The model provided the slope and intercept of 2 distinct regression lines: MR1 = 0.6997 + 0.01389 × THI (R2 = 0.1735; F1,43 = 9.02; P = 0.004) below the break point (29 to 73.6 THI range) and MR2 = –42.76398 + 0.6048 × THI (R2 = 0.9317; F1,9 = 122.7; P = 0.0001) above the break point (73.6 to 84 THI range). DISCUSSION First of all we wish to notice that the mortality recorded in the present study was lower than the threshold of 0.1% indicated by the excellent U.S. reference for in-transit pig losses (http://www.grandin.com/pig.transport.audit.form.html). The reasons for such result are not known. Even if differences in the age and weight of animals transported may have a role, further studies comparing U.S. and European procedures and legislation with respect to protection of animals during transport and at slaughtering would be of interest. Month The risk of pigs to die was significantly greater in July, which was the hottest month among those considered in the study. Additionally, pigs transported in summer months were more likely to die than those transported in non-summer months. Analyzing the pig slaughter procedures in Germany, Werner et al. (2007) indicated a significant increase of mortality rates in pigs during transport to the slaughter house and lairage at the abattoirs in the months of June, July, and August. Analogously, in a study conducted in the Czech Republic, Vecerek et al. (2006) reported an increase in mortality rates during transport in the summer months of June, July, and August. In a Canadian study, Haley et al. (2008a) showed that August had the greatest monthly in-transit pig losses and that August losses accounted for 20% of the total annual losses. On the other hand, Sutherland et al. (2009) reported that the percentages of pigs DOA increased at temperature above 20°C, but the percentages of pigs DOA were greater from July to December, a period during which the average monthly temperature decreased. Similarly, it was pointed out that the increase in the average temperature during the journey significantly increased the mortality risk, although the effect of the season was not significant (Averos et al., 2008). Therefore, in the light of present and previous findings, we believe that it would be important to implement market procedures aimed at improving animal welfare (i.e., pigs loading during night time, travel in the early hours of the day, and cooling devices on trucks) during summer months. Slaughter House In the present study, the risk of death during lairage was associated with the abattoir of destination. In detail, the greatest risk for DIP was recorded for the slaughter house A, which was managed with poor cooling devices (drop), high brightness (60 lx), and greatest stocking density (0.46 m2/100 kg LW), which is above the limit of 0.42 m2/100 kg LW indicated in European regulation (European Council, 2005). Conversely, the lowest risk of DIP was recorded for the slaughter house C, which presented the lowest stocking density (0.64 m2/100 kg LW), roof and side walls equipped with large open windows, low brightness (40 lx), and high-pressure sprinklers as cooling devices. Several studies reported that high stocking density reduces animal welfare and is associated with greater mortality during transport and lairage (Ritter et al., 2007; Fitzgerald et al., 2009). However, these studies did not distinguish the effect relative to each of the 2 phases, transport and lairage. In particular, little information is available on pig losses at lairage. High-pressure sprinkler systems are effective in reducing air temperatures and preserve animal welfare much better compared with low-pressure systems (Nienaber and Hahn, 2007; Collier et al., 2006). Weeks (2008) has recently reviewed pig welfare in lairage and suggested that high stocking rates and mixing of animals should be avoided and well-ventilated and thermally comfortable conditions should be ensured to maximize pigs' welfare. When temperature increases, a well-ventilated pen and greater space between animals allow better air circulation and low humidity during lairage. This condition, associated with body wetting by drops or sprinklers, may ensure a good functionality of the thermoregulatory processes favoring heat loss. The pig welfare at the abattoir is affected by several factors and a multifactorial approach should be considered to assess pig losses during lairage. Time Traveled The relationships between the length of the journey and mortality were studied. Although it is generally accepted that longer journeys can negatively affect animal welfare, different studies (Colleu and Chevillon, 1999; Werner et al., 2007; Vecerek et al., 2006) reported evidence of greater mortality in slaughter pigs for transports <8 h, which is the upper critical value reported in the European directive (European Council, 2005) above which journeys are categorized as long. These European studies are characterized by short distances traveled and documented greater pig losses in very short journeys (1–2 h). These findings may reflect the failure of pigs to recover from the stress experienced during loading procedures and in the early stage of journey. Analogously, Canadian and U.S. studies, where distance traveled were greater, reported that pig losses increase in short journeys and decrease for greater distances. Longer journeys may give pigs that have been stressed when being loaded on to the trailer the time to rest and recover (Haley et al., 2008b; Sutherland et al., 2009). Our findings indicated that for heavy weight slaughter pigs the increase in the duration of the journey may represent a critical factor even in the case of transports lasting less than 8 h. The analysis of the interactions between time traveled and the months of the year indicated a significant increase of the risk to die starting from transports greater than 2 h performed during the hottest months. To the best of our knowledge, the present study provided the first evidence that the combination of these 2 factors (time traveled and month of the year) may be critical for pigs' welfare during market procedures and also indicated a threshold of transport duration (2 h) above which summer mortality increases significantly. Temperature–Humidity Index The 2-phase regression analysis indicated a positive association between in-transit and lairage mortality and the increase of THI. The same analysis provided 2 break points 78.5 and 73.6 THI for DOA and DIP, respectively. The first break point corresponds to combinations of temperature and RH of about 35°C and 20% or 30°C and 50% or 27°C and 80%. The second break point corresponds to combinations of temperature and RH of about 30°C and 20% or 26°C and 50% or 24°C and 80%. We may hypothesize that the greater threshold of THI for DOA compared with DIP may be related to ventilation generated by the moving truck and/or by the fans that may have been present in some of the trucks included in the study. Moreover, the procedures (loading, travel, and unloading) experienced by pigs before lairage may cause the animal stress and make them more susceptible to the heat at the slaughter house. In a previous study, Fitzgerald et al. (2009) reported that slaughter pig losses were associated with the increase of the THI. On the basis of THI values calculated according to the National Oceanic and Atmospheric Administration (1976), they indicated that between the range of –19 and 1 THI, the mortality rate was constant and that above the value of 1 THI, the increase to 30 THI was associated with a 2.19-fold increase (0.0025%) in total losses. Moreover, they reported that a decrease in total losses by 0.00025% was associated with an increase in wind speed. A direct comparison between our results and those reported by Fitzgerald et al. (2009) is not possible because of the use of different formulas for THI calculation and different approaches for thresholds detection. In our case, the formula used was the one commonly used for the Mediterranean area (Segnalini et al., 2013). However, in general terms, both studies agree in indicating that a climatic threshold exists above which pig market losses increase. The thermal comfort zone in livestock is defined by the functionality of thermoregulatory processes in relation to weather parameters such as temperature, humidity, wind speed, and solar radiation. In view of this, the development of new bioclimatic indices, which may also include other climatic variables in addition to temperature and humidity, could be useful to better define the range of thermal comfort in marked pigs. These new indices may improve adaptation to climate scenarios that were recently described for the Mediterranean area (Segnalini et al., 2013). Conclusions Pig losses during transport and lairage are a multifactorial problem. In the present study, month, slaughter house of destination, and length of the journey were identified as sources of significant variations for in-transit and lairage heavy pig losses. The results indicated that for heavy weight slaughter pigs, the risk to die increases in the hottest months and with the time traveled even in the case of transport times less than 8 h. Lower stocking density, appropriate ventilation, and cooling devices should be functional to ensure the animal welfare at the abattoirs when temperature increases. The THI may represent a useful tool to predict the risk of heat-related mortality even if the development of new bioclimatic indices that take into account other climate variables is encouraged. These findings may be of help for the pig industry to further improve in-transit and lairage management in slaughter pigs. Provision of emergency interventions as well as implementation of adaptation measures for reducing the adverse consequences of heat stress could represent useful strategies to improve animal welfare and pig industry profit. LITERATURE CITED Averos X. Knowles G. Brown S. N. Warriss P. D. Gonsalvez L. F. 2008. Factors affecting the mortality of pigs being transported to slaughter. Vet. Rec.  163: 386– 390. Google Scholar CrossRef Search ADS PubMed  Colleu T. Chevillon P. 1999. Incidence des paramètres climatiques et des distances sur la mortalité des porcs en cours de transport. (In French.) Techni-porc  22: 31– 36. Collier R. J. Dahl G. E. Van Baale M. J. 2006. Major advances associated with environmental effects on dairy cattle. J. Dairy Sci.  89: 1244– 1253. Google Scholar CrossRef Search ADS PubMed  European Council 2005. Council Regulation No. 1/2005 of 22 December 2004 concerning the protection of animals during transport and related operations and amending Directives 64/432/EEC and 93/119/EC and Regulation, 97/1255/EC. In: Off. J. Eur. Communities: Legis.  3: 1– 44. Fitzgerald R. F. Stalder K. J. Matthews J. O. Schultz Kaster C. M. Johnson A. K. 2009. Factors associated with fatigued, injured, and dead pig frequency during transport and lairage at a commercial abattoir. J. Anim. Sci.  87: 1156– 1166. Google Scholar CrossRef Search ADS PubMed  Google Earth 2012. Software (Version 7). Mountain View, CA: Google Inc. (2012). Available on line at http://www.google.com/earth/download/ge/agree.html. Accessed 15th December 2012 Google Maps 2013. Mountain View, CA: Google Inc. (2013). Available on line at https://maps.google.it/maps?hl=en&tab=wl. Accessed 25th January 2013. Haley C. Dewey C. E. Widowski T. Poljak Z. Friendship R. 2008a. Factors associated with in-transit losses of market hogs in Ontario in 2001. Can. J. Vet. Res.  72: 377– 384. Haley C. Dewey C. E. Widowski T. Poljak Z. Friendship R. 2008b. Association between in-transit loss, internal trailer temperature, and distance travelled by Ontario market hogs. Can. J. Vet. Res.  72: 385– 389. Hintze J. ( 2007). NCSS 2007. NCSS, LLC. Kaysville, Utah, USA. ISTAT 2012. Istituto Nazionale di Statistica. Codici dei comuni, delle province e delle regioni. (In Italian). Available on line at www.istat.it/it/archivio/6789. Accessed 10 October 2012. Kelly C. F. Bond T. E. 1971. Bioclimatic factors and their measurements. In: A guide to environmental research in animals.  Natl. Acad. Sci., Washington, DC. p. 7– 92. Nanni Costa L. 2009. Short-term stress: The case of transport and slaughter. Ital. J. Anim. Sci.  8: 241– 252. Google Scholar CrossRef Search ADS   National Oceanic and Atmospheric Administration 1976. Livestock hot weather stress. Regional Operations Manual Letter C-31–76.  National Oceanic and Atmospheric Administration, Kansas City, MO. Nickerson D. M. Facey D. E. Grossman G. D. 1989. Estimating physiological thresholds with continuous two-phase regression. Physiol. Zool.  62: 866– 877. Google Scholar CrossRef Search ADS   Nienaber J. A. Hahn G. L. 2007. Livestock production system management responses to thermal challenges. Int. J. Biometeorol.  52: 149– 157. Google Scholar CrossRef Search ADS PubMed  Ritter M. J. Ellis M. Bertelsen C. R. Bowman R. Brinkmann J. DeDecker J. M. Keffaber K. K. Murphy C. M. Peterson B. A. Schlipf J. M. Wolter B. F. 2007. Effects of distance moved during loading and floor space on the trailer during transport on losses of market weight pigs on arrival at the packing plant. J. Anim. Sci.  85: 3454– 3461. Google Scholar CrossRef Search ADS PubMed  Segnalini M. Bernabucci U. Vitali A. Nardone A. Lacetera N. 2013. Temperature humidity index scenarios in the Mediterranean basin. Int. J. Biometeorol.  57: 451– 458. Google Scholar CrossRef Search ADS PubMed  StatSoft, Inc 2004. STATISTICA (data analysis software system), version 7. StatSoft Inc., Tulsa, OK, USA. Sutherland M. A. McDonald A. McGlone J. J. 2009. Effects of variations in the environment, length of journey and type of trailer on the mortality and morbidity of pigs being transported to slaughter. Vet. Rec.  165: 13– 18. Google Scholar CrossRef Search ADS PubMed  Vecerek V. Malena M. Malena J. R. Voslarova E. Choupek P. 2006. The impact of the transport distance and season on losses of fattened pigs during transport to the slaughterhouse in the Czech Republic in the period from 1997 to 2004. Vet. Med. (Praha)  51: 21– 28. Warriss P. D. 1998. The welfare of slaughter pigs during transport. Anim. Welf.  7: 365– 381. Weeks C. A. 2008. A review of welfare in cattle, sheep and pig lairages, with emphasis on stocking rates, ventilation and noise. Anim. Welf.  17: 275– 284. Werner C. Reiners K. Wicke M. 2007. Short as well as long transportation duration can affect the welfare of slaughter pigs. Anim. Welf.  16: 385– 389. Footnotes 1 This study was financially co-supported by Italian Ministry of Health (IZSLER 03/07, PRC 2007003), Ministry of Agriculture (FACCE-MACSUR project D.M. 2660-20.06.2012), and European community (GI-Neu EU project n° 324476) Università degli Studi della Tuscia. The authors gratefully acknowledge F. Guizzardi, S. Bignami, C. Morini, and S. Ambrosini (ASL Mantova, distretto di Viadana) for provision of the mortality data and Lombardia, Veneto, and Emilia Romagna Environmental Protection Agency and the Research Unit for Agricultural Climatology and Meteorology (CRA-CMA) for providing meteorological data. American Society of Animal Science
TI - Analysis of factors associated with mortality of heavy slaughter pigs during transport and lairage
JF - Journal of Animal Science
DO - 10.2527/jas.2014-7670
DA - 2014-11-01
UR - https://www.deepdyve.com/lp/oxford-university-press/analysis-of-factors-associated-with-mortality-of-heavy-slaughter-pigs-AxDNhqRGz3
SP - 5134
EP - 5141
VL - 92
IS - 11
DP - DeepDyve
ER -