The study was conducted in two stages: (1) an antemortem inspection was performed during unloading and lairage, and (2) a postmortem evaluation was conducted during the evisceration process and carcass inspection before refrigeration. Data were collected from 1,161 gilts and barrows (one batch per farm) from commercial lineages that were being transported directly from the farms (n = 22) (Caldas and Quindío Departments in Colombia) to the slaughterhouse. These farms are responsible for 6.0% and 1.6% of national pig production, respectively (Martinez et al., 2022). A large proportion of these farms in this region are certified in good farming practices (Martinez et al., 2022) and it has one of the largest pig slaughter operations in the country (Pastrana-Camacho et al., 2023). Pigs were transported in single-deck and double-deck trucks equipped with passive ventilation systems. The pigs were unloaded from the trucks by personnel contracted by the slaughterhouse. The trucks were equipped with an adjustable slope metal ramp with an anti-skid floor, and the pigs were kept with familiar groups during lairage (5 ± 2 h), and then put into lairage pens (4.3 m × 1.8 m) with ad libitum access to water through a nipple drinker. At the end of the lairage period, pigs were driven to an electrical stunner and restrained in a single animal box. Following stunning, pigs were vertically exsanguinated (5 ± 2 s) and carcasses were placed in a dehairing machine at 62°C for 5 minutes. The residual hair was removed using a knife and flame; carcasses were then eviscerated and split before being placed in a chiller set at 4°C for 24 h. The slaughterhouse had a system for identifying the pigs, whose individual number was assigned during weighing and maintained throughout the process. The assigned number was marked using a tattoo on the back of the live animals, as well as an elastic band located on the carcass to guarantee a match between the viscera and the carcass of each pig during antemortem (AM) and postmortem inspection.

Data collection was performed on 30 non-consecutive days between June 2020 and September 2020. A scheme based on visual inspection was used to assess specific lesions and conditions during both antemortem and postmortem evaluation. To control selection and information bias, the inclusion criterion was the origin of the pigs from fattening farms, which were certified in good animal husbandry practices. Pigs that were not followed with their corresponding postmortem evaluation and culling pigs were excluded from the study.

This investigation included 22 livestock drivers who transported fattening pigs to the slaughterhouse. In interviews, truckers were asked about their age, educational status (i.e., elementary, high school, community college, and university), animal transport training (y/n), years of experience, and knowledge of Colombian transportation (y/n) and animal welfare legislation (y/n), as well as their perception of the animals’ capacity to feel pain and stress using two questions with responses measured a Likert-based scale that ranged from 1 to 5 (i.e., strongly disagree to strongly agree).

During unloading at the slaughterhouse, the following truck features were evaluated through direct observation by a trained veterinarian: type of bodywork (i.e., wood, metal, or mixed), number of decks (i.e., single or double), the presence of water sprinkling devices (y/n), and type of ventilation (active or passive). In addition, truck drivers were interviewed to determine the following characteristics: (a) farm name, (b) use of loading ramp (y/n), (c) transport time (h), (d) specialized transport (y/n), (e) stopovers (y/n), and (f) Type of roads is a separate category, type of roads (i.e., rural or highway).

During lairage, all pigs were individually identified by numbers painted on their sides and rump with washable spray used for animal marking, and weighed before slaughter. The identification assigned by the slaughterhouse during weighing was recorded to track the viscera and carcasses of the animals selected for the study. Pigs were evaluated antemortem during lairage by two trained veterinarians on weekdays from 3:00 a.m. to 5:00 p.m. All pigs had the same lairage time (5 ± 2 h) and were evaluated 3 h after unloading. Table 1 shows the description of the antemortem and postmortem indicators evaluated (Welfare Quality, 2009). The shape of the skin lesions was evaluated to establish their possible cause. Comma-shaped skin lesions, located in different areas and ranging in size from 2 to 5 cm, are characteristic of those produced by bites during antagonistic encounters between pigs (these encounters often occur during social regrouping). Linear skin contusions (long and thin), with sizes between 10 and 15 cm, located on the ham and shoulders of the animals, are caused by the hooves of the pigs during mounting rectangular and diffuse contusions may be related to the use of blunt objects during handling (Varón-Álvarez et al., 2014). The evaluation of tail length, tail injury, and lameness was performed according to the guidelines suggested by Valros (Valros et al., 2004).

Postmortem (PM) inspection was performed by a trained veterinarian in the eviscerating area evaluating the lung score according to Goodwin et al. (1969) (Goodwin et al., 1969), for Mycoplasma hyopneumoniae by quantification of the affected lung area by the mean of 0–10 points (cranial and medial lobes) or 0–5 points (caudal and accessory lobes). Points were then summed up to provide an overall compromised lung area score (i.e., 0–55 points) (see Table 1 ).

Data analyses were conducted using Stata^® software, version 13.0 (StataCorp LP, College Station, TX, USA). In this study, the experimental units were individual pigs. A normality test of the evaluated variables was carried out. The variables with non-normal distribution were transformed by the means of a natural logarithm (i.e., weight, the lung score, and number of skin lesions), and these values were later used for regression analyses; results were transformed back to the original units of measures. Categorical variables regarding the driver (i.e., age, educational status, animal transport training, years of experience, level of education, transport training, and perception of animal welfare), the truck (i.e., type of bodywork, number of decks, water sprinkling devices, and type of ventilation) and transportation (i.e., use of loading ramp, transportation time, stopovers, and type of roads) were subjected to descriptive statistical analysis and presented as a proportion of answers means and, when applicable, the variability was expressed in ranges.

For the identification of the relationship between categorical antemortem (i.e., lameness, umbilical hernia, inguinal hernia, tail length, tail lesion, rectal prolapse, vaginal prolapse, cough, and weight) and postmortem (i.e., pleurisy, pericarditis, and liver lesions) variables, a chi-squared analysis was performed. A Spearman correlation matrix was done to preliminarily inspect the correlation among antemortem and postmortem variables. The Wilcoxon–Mann–Whitney U-test was used to measure the differences between the distribution of skin lesions according to body region (i.e., ears, neck, midsection, rump, and legs) and their color (i.e., bright or dark red). The Spearman correlation was used to measure the differences between the distribution of pulmonary lesions scores (cranial, middle, caudal, and accessory lobes). A multilevel, mixed-effects lineal regression model analysis was performed to identify variables associated with the number of skin lesions (dependent variables) with the sociodemographic characteristics of the truck drivers and transport conditions (independent variables), controlling for the effect of the farm of origin (fixed effect). A multilevel, mixed-effects logistic regression analysis was performed to identify variables associated with tail lesions as binary variables (i.e., absence or presence; dependent variables) with the antemortem and postmortem indicators (independent variables) to control the effect of farm of origin. The model included the live weight of animals as a covariate. The confounding effect was considered present when the estimates changed by at least 20%. The number of decks was evaluated as a confounding variable but was removed from the model because the principal variation source was the effect of the farm of origin. A probability level of p < 0.05 was chosen as the limit for statistical significance in all tests, and probability levels of p ≤ 0.10 were considered as a tendency.

All drivers (n = 22) were male with 46.4 (± 6.4 years) and an average of 6.5 (± 8.3 years) of experience transporting animals; all of them had high school studies and only 50% (n = 11) had specialized training in animal transportation. Only four drivers had knowledge of Colombian animal welfare law, and 12 drivers had knowledge about Colombian animal transportation law. All drivers strongly agreed that animals have the capacity to feel pain and can suffer from stress.

Regarding truck features, all trucks were metal and wooden structures, six trucks were double-decked, whereas the remaining 16 were single-deck trucks, and only six trucks had water sprinkling systems. The evaluated trucks used a passive ventilation system. All trucks were used exclusively for livestock transport and loading ramps were used at all farms. The average transportation time from farms to the slaughterhouse was 1.92 h (± 1.01 h), with a minimum time of 0.5 h and a maximum time of 4 h, and only seven drivers stopped during transportation for more than 10 min to take a break. The type of road was 31.8% rural, and 68.2% highway ( Table 2 ).

Pigs presented with an average weight of 105.6 kg (± 14.1 kg), with an average of 10.5 (± 9.8) lesions on their bodies. These lesions were located mainly on the midsection (36.1%, n = 732), neck (27.9%, n = 566), and rump (21.7%, n = 440), followed by the ears (12.9%, n = 261) and legs (1.3%, n = 26). Most lesions presented a dark-red color (74.7%, n = 1,513); the remainder were bright red (25.2%, n = 512) (p ≤ 0.001) ( Figure 1 ). The most common skin lesion shape was linear (55.8%, n = 1,130), followed by comma-shaped (34.7%, n = 703), diffuse (7%, n = 140), and rectangular (2.6%, n = 52).

Only 0.9% (n = 11) of the pigs evaluated came from farms where tail docking was not practiced, whereas the remaining pigs had undergone caudectomy procedures ( Table 3 ). Tail biting lesions were present in 4.7% (n = 55) of the pigs evaluated. Cough was present in 18.9% (n = 222) of the pigs evaluated, while vaginal prolapse 0% (n = 1), rectal prolapse 0.1% (n = 2), and lameness 1.3% (n = 16) were less prevalent ( Table 3 ).

At PM evaluation, pigs presented an average of 34.2 ( +/- 13.5) pulmonary lesions, and, according to the Goodwin scale, these were most prevalent on the medial and cranial lobes (12.2 ± 5.2 and 11.9 ± 4.9, respectively), followed by the caudal and accessory lobes (5.3 ± 2.6 and 1.6 ± 1.7, respectively) (p ≤ 0.001). Pleurisy and pericarditis were present in only 9.9% (n = 115) and 2.7% (n = 32) of the evaluated pigs, respectively, whereas lung abscesses and liver lesions were present and the presence of lung abscesses and liver lesions was negligible (0.3%, n = 4, and 0%, n = 1, respectively) ( Figure 2 ).

The multilevel, mixed-effects lineal regression analysis identified that the number of skin lesions observed during lairage was related to the driver’s age, marital status, and job experience (p ≤ 0.001) ( Table 4 ). In this analysis, transport conditions were not related to the number of skin lesions. The tail lesions (i.e., presence or absence) were significantly associated with the number of skin lesions, the scoring of pulmonary lesions, and animal live weight (p ≤ 0.001) ( Table 5 ). However, the residual analysis of the two models indicated that the variability was significant, indicating the existence of characteristics related to the farms of origin, transport conditions, driver characteristics, and animal welfare indicators identified in the antemortem and postmortem inspection, which were not taken into account in these analyses and could explain, at least in part, the variations in the number of skin and tail lesions.

The characteristics of livestock drivers in this study, namely sex (n = 22), age (46.4 ± 6.4 years), and education were similar to those in other studies conducted under commercial conditions in Mexico (Valadez-Noriega et al., 2018) and Colombia (Romero et al., 2022). The formal training and instruction of drivers in animal transportation at Latin America is not yet a common practice (Romero and Sánchez JA, 2011), usually, the drivers are hired as multipurpose drivers, with responsibilities other than the transportation of animals only 50% (n = 11) of drivers were trained in animal transportation and only 18% (n = 4) stated that they had knowledge of Colombian animal welfare-specific regulations. This lack of training has also been observed in drivers responsible for the commercial transportation of other species in Colombia, such as cattle (Romero and Sánchez JA, 2011). This effect responds to multifactorial conditions such as jobs in the animal transportation sector not being specialized and exclusive, that is, usually, the drivers are hired as multipurpose drivers (Romero et al., 2018), and animal welfare policies and regulations are being implemented more slowly in developing countries than in developed ones (Gallo and Tadich, 2018). This same issue affects the trucks used for animal transportation, where the use of trucks is regularly non-exclusive. In this study, 72.7% of trucks were single-decker, with metal and wood bodywork. However, other studies in Colombia have found that there is more heterogeneity among trucks (Romero et al., 2018). Nevertheless, besides confirming, in line with previous research, that transportation is a known stressor for pigs that affects their welfare and behavior (Schwartzkopf-Genswein et al., 2012), our results did not show that truck type had a significant impact on the prevalence of animal welfare indicators. However, the lack of evidence pertaining to this effect may be owing to the short transportation times (1.92 ± 1.01 h) observed in this study.

Tail biting is considered an indicator of reduced welfare and production performance in pigs, and it is associated with condemnations in ham (Sinisalo et al., 2012). In this study, we found a prevalence of biting lesions of 4.7% (n = 55). This prevalence is high compared with other reports from studies performed in Colombia (0.7%) during on-farm evaluations (Martinez et al., 2022). Likewise, the prevalence found was lower than that reported in Portugal (Franco et al., 2021). The discordant results could be explained by differences in the farm of origin (Horst et al., 2019); tail-docking techniques (Martinez et al., 2022); the presence or absence of caudectomy (vom Brocke et al., 2019; Heinonen et al., 2021); placement on the slaughter line when assessments are conducted (before or after scalding and dehairing) (Carroll et al., 2016); tail length and healing scores (Gomes et al., 2022); and swine system managements, the last of which is dependent on the region (Martinez et al., 2022). Tail docking is prohibited by the European Union (de Briyne et al., 2018), but is legal under Colombian law, which stipulates that the procedure must be performed with anesthesia and analgesia by trained personnel under the supervision of veterinarians (Colombian Agricultural and Rural Development Ministry, 2020). Nevertheless, although the practice is still controversial (Harley et al., 2012), it is reported that tail docking diminishes tail-directed behaviors on-farm (Thodberg et al., 2018) and since tail lesions are associated with on-farm health-related indicators (Franco et al., 2021), examining tail biting lesions during inspections could improve the efficacy of risk-based meat inspection programs and help farmers provide up-to-date feedback regarding on-farm animal welfare interventions that they have implemented. This was implemented in Germany with the use of a control program called Scchwanzbeiβ-INTERVENTIONS-Program (meaning ‘tail biting intervention program’), which includes the monitoring of health and animal welfare indicators in slaughterhouses. The program has resulted in a decrease in tail biting in fattening pig farms (vom Brocke et al., 2019) and increased adherence to animal welfare plans by producers (Uehleke et al., 2021). In future studies, we suggest evaluating tail lesions during postmortem inspection, as they are more visible on the carcass (van Staaveren et al., 2017); likewise, a combination of lesion scoring and tail length measurement is required, especially when animals have intact tails (Valros et al., 2020).

Respiratory disease management represents one of the most challenging problems in the swine industry, because of the multifactorial etiology associated with respiratory diseases such as housing conditions on the farm, population density, and biosecurity (Ciuoderis-Aponte et al., 2022). Therefore, it is expected that pigs reared under commercial conditions have a high probability of developing respiratory compromise before slaughter (Rampelotto et al., 2022). These diseases affect pig growth and feed conversion, increase mortality and condemnation during slaughter, affect meat quality (Permentier et al., 2015) and, because of their high treatment costs, reduce overall profits (Cornelison et al., 2018). Respiratory diseases also represent a risk factor for the presence of non-ambulatory and dead-on-arrival pigs after transport (Romero et al., 2015). For instance, cough is an indicator of pulmonary disease used in clinical trials and on-farm observations (Nathues et al., 2012). The association between a cough and respiratory disease has been found to predict the prevalence and severity of pneumonia in pigs (Pessoa et al., 2021; Silva et al., 2022). In the present study, 18.9% of pigs presented coughing events during antemortem inspection, and a significant association with pulmonary compromise, manifesting mainly on the cranial lobes, and as pericarditis and pleuritic, was found PM. Pulmonary lesions in pigs are indicative of the presence of infectious agents of porcine respiratory disease complex, such as M. hyopneumoniae, Bordetella bronchiseptica, and Actinobacillus pleuropneumoniae, as well as porcine reproductive and respiratory syndrome virus (PRRSV), and the swine influenza virus all of which are transmitted thorugh close contact among pigs (Brockmeier et al., 2002). Although little research has been conducted into other transmission methods besides close contact among pigs, it is reported that these agents can spread via airborne through biosecurity hazards such as infected animals and contaminated vehicles or people (Dee et al., 2009; Maes et al., 2018). Therefore, coughing can be an indicator both of animal health and the management efficiency of farms, particularly as it relates to animal transport (Baraldi et al., 2019). However, despite these benefits, cough measurement is time-consuming, and this is an aspect that diminishes its utility in commercial scenarios. For this reason, the number of studies evaluating the ability of new technologies and computational devices to record measurements of coughing in pigs is increasing (Pessoa et al., 2021), showing that producers have new alternatives which can enhance the utility of their monitoring surveillance systems for animal welfare. The evaluation of cough during transport and lairage at the slaughterhouse could be strategic, as it would allow inspectors to determine the prevalence of respiratory diseases from several farms at once, in turn diminishing costs and providing feedback to the farmers, which could in turn enable them to develop strategies to improve on-farm health conditions and reduce the prevalence of non-ambulatory pigs after transportation (Kongsted and Sørensen, 2017).

Liver lesions, such as those produced by Ascaris spp., and lung abscesses had a low prevalence, meaning that some sanitary practices were effective. In this study, no liver lesions caused by migrating Ascaris spp. were observed. In contrast, a retrospective observational study carried out in Italy, which evaluated the records of animals postmortem inspection, found that the first cause of partial seizure was “liver milk spot” (91.79% of seizures), which suggests a lack or inadequacy of parasite control plans in farms, which is likely due to the low commercial value of liver in Italy (Guardone et al., 2020) (an issue that does not exist in Colombia, where the consumption of viscera is a common practice). Likewise, other indicators such as vaginal prolapse, rectal prolapse, and lameness were less prevalent, and these results could be explained by the joint effort made by the health authorities and the swine producers’ association in the implementation and certification of good farming practices and animal welfare programs (Martinez et al., 2022). Similarly, Australia has a pig health monitoring scheme in which producers participate, which includes training in on-farm antemortem inspection. In this context, a study was conducted to evaluate the effectiveness of antemortem inspection of pigs by producers, which showed that they were more competent than slaughterhouse inspectors in identifying diseased pigs and in detecting lesions affecting the pigs’ fitness for transport, considering that they knew the health history of the animals and performed the observation of pigs prior to the pigs being transported, and being subject to any additional illness or disease that could take place en-route (Jackowiak et al., 2006). These results are interesting and demonstrate that joint work between innovative producers and health authorities is effective in improving the epidemiological surveillance of animal health and welfare problems. However, they are limited to technology driven production systems with good risk management.

Skin lesions are used as indicators of health, animal welfare and aggression in pigs (Driessen et al., 2020a). In this study, skin lesions inspected AM were found mainly on the midsection and were mostly linear or comma shaped; these lesions are produced by agonistic behaviors such as mounting or fights (Teixeira and Boyle, 2014; Liu et al., 2022). Mounting behavior is a normal sexual display in pigs; nevertheless, it is undesired on commercial farms, where pigs are heavy, and because of the pressure applied by the sternum during this activity, mounting can cause lameness and the formation of carcass lesions on the midsections of pigs. On the other hand, the lesions on the head, neck, and front section of the body were the second most prevalent and they are more representative of reciprocal fighting due to agonistic interactions on-farm, during transport, or lairage. (Driessen et al., 2020b) evaluated 4,507 pig carcasses in a Belgian slaughterhouse and suggested that the most prevalent skin lesions were related to social mixing on the farms. All those interactions have been proven to happen more often in farms with low enriched environments or pens with only males or a higher male-to-female ratios (Teixeira and Boyle, 2014). In this study, there were no on-farm records, so it is not possible to assert that there was an increase in these behaviors, but other authors have reported that a high incidence of these behaviors in groups of male pigs is associated with higher skin lesion scores both prior to and post slaughter (Teixeira and Boyle, 2014).

The color of the lesions in this study were predominantly dark red (74.7%, n = 566) which suggests that the wounds were inflicted on the farm or during transportation. A recent on-farm study evaluated the skin lesions of 532 pigs at 7, 9, and 10 weeks of age (early life), and at 15 and 20 weeks of age (later life). The evaluated pigs were followed up post slaughter to determine if lesions that had been acquired during different stages of production remained visible on the carcass (Carroll et al., 2018). This study found that lesions that occurred 11 weeks prior to slaughter remained visible on the carcass in the form of healed (non-red) skin lesions. These findings suggest that carcass-based assessments of these types of lesions reflect lifetime animal welfare status (when pigs are slaughtered at a standard commercial age), rather than simply reflecting welfare in the period prior to immediate slaughter. However, if the objective is to assess these injuries to reflect levels of on-farm aggression, a comprehensive and robust scoring system is required (Driessen et al., 2020b).

In the analysis of the relationships between skin lesions with the indicators evaluated during transport, and the sociodemographic characteristics of the drivers, as well as tail lesions with the welfare indicators in the antemortem and postmortem inspection, the effect of the farm was included using multilevel analysis. As the prevalence observed can differ depending on the type of production systems, geographical location, number of pigs, management conditions, the health status of the herd, shipping and unloading conditions, and sanitary programs, among others (Horst et al., 2019). The two selected models showed significant variability in the analysis of residuals, indicating that other variables not evaluated in the study could explain the variation in the number of skin lesions and the presence of tail lesions. Other authors have also used multivariate statistical analysis to control for the effect of the farm of origin because, as in our case, no measurements were made at this level (Scollo et al., 2017). Likewise, pig weight was included as a covariate in the multilevel, mixed-effects, logistic regression analysis.

Studies that have evaluated human–animal interaction during the preslaughter of cattle and swine have shown that it is a multifaceted work activity and that workers who are in direct contact with the animals modify the animals’ responses to the preslaughter procedures, and that this affects workers’ productivity, job performance, and personal satisfaction, as well as product quality (Pastrana-Camacho et al., 2023). Skin lesions were associated with truck driver characteristics such as age, marital status, and job experience. The drivers evaluated generally had little training in animal welfare and a general lack of knowledge of the regulations related to animal transport, an aspect that could have influenced the results of the present study, taking into account that the level of knowledge about pig behavior and the driving skills of the driver directly influences animal welfare (Fitzgerald et al., 2009). Many factors have been associated with the effect of driver characteristics on animal welfare in preslaughter: these include their attitude and empathy (Leon et al., 2020), years of experience as a truck driver (Valadez-Noriega et al., 2018), driving style, education and training (Rioja-Lang et al., 2019), work pressure, and interpersonal relations background (Hemsworth PH, 2011). These characteristics of the drivers have been identified as risk factors that increased the probability of finding higher total transport losses (dead pigs and non-ambulatory pigs) (Romero et al., 2022) and with the prevalence of skin lesions in Colombia (Varón-Álvarez et al., 2014). Other studies in Colombia outlined four profiles of workers in swine preslaughter: the first was composed of workers who relate to the animals and their work in a mechanical way, the second by professional workers who were emotionally close to the animals, the third by those who were committed to the animals and their work, and the fourth was composed of workers who were apathetic toward both the animals and the work activity (Pastrana-Camacho et al., 2023). Similar results have been reported in sheep transporters in Mexico (Pulido et al., 2018). These profiles provide a basis for designing training and mitigation strategies to address animal handling and welfare problems, a strategy which could be particularly useful to Colombia and other Latin American countries, where animal welfare sanitary legislation is in the process of being updated and implemented (Martinez et al., 2022).

Tail biting is considered a harmful behavior in pigs, which has been correlated with the prevalence of herd health problems, particularly respiratory, enteric, and locomotor diseases (Boyle et al., 2022). In this study, tail lesions were associated with the scoring of pulmonary lesions and skin lesions. However, it is important to note that, while tail lesion records in slaughterhouses provide a large database, their usefulness is restricted because the environmental and animal husbandry factors associated with biting behavior are unknown. This information would only be possible to obtain through the development of prospective on-farm studies (Carroll et al., 2018), or the systematic analysis of records from highly technologized farms (Taylor et al., 2010). Several authors have outlined the pathogenesis of tail lesions, providing a detailed description of the nature of the lesions and routes of infection, in turn indicating that, in addition to physical damage to the tail and hindquarters of the bitten pig, a local infection can spread systemically, leading to the presence of abscesses, especially in the lungs (Schrøder-Petersen and Simonsen, 2001; Harley et al., 2012).

Regarding the relationship between tail lesions and skin lesions found in this study, other authors have described the same results, as in the case of the study conducted by (van Staaveren et al., 2017), who visited 31 Irish farms at different stages of production and evaluated tail lesions, lameness, bursitis, body condition, and skin lesions. These same indicators were assessed during postmortem inspection and the authors concluded that tail lesions and skin lesions may be possible iceberg indicators of animal welfare on farms. Similarly, other authors have suggested that these damaging behaviors can occur simultaneously on farms and that their interpretation is associated with animal welfare problems (Carroll et al., 2018). Further longitudinal studies are required to assess skin and tail lesions at farms, transport, and lairage to establish the most accurate measurement scales and to control for measurement biases.