**4. Discussion**

This study documented transport-related mortality and prevalence of animals considered unfit for transport for farm animals (i.e., cattle, sheep/goats, pigs) on a stopover at a control post in Southern Italy using reports of OVs. This study reports on the number of animals which died and were judged unfit to continue travel by OVs in compliance with Section 5 of Annex 1 of Council Regulation (EC) No. 1/2005, thus providing the first report containing statistics on farm animal movements from Northern Europe to the Balkan Peninsula. Considering that only 0.8% of the shipments was associated with one of the considered welfare issue (i.e., DOA, UFT), and that the Eurobarometer [21] reports that the perception of animal welfare has increased among European consumers, our data may help consumers gain knowledge of live animal transport within Member States.

The animals on the stopover at our CP were mainly travelling to Greece, a country which relies heavily on meat and dairy imports [22]. Animals were being transported for different purposes: for the final fattening period before slaughtering in the destination countries, for breeding, for direct sale or for slaughter. The number of vehicles stopping at the CP dropped significantly after 2012, probably due to the economic crisis in Greece [22]. The reasons for the grea<sup>t</sup> demand for live animal transport have been analyzed by the European Union [23] and include a high level of demand for fresh meat; use of indigenous slaughterhouse facilities; use of by-products (such as skin and offal) from the slaughter process; poor meat transport, refrigeration and storage capacity; use of religious slaughter rituals in some EU-countries. The EUR-Lex (1998) states that even though animal welfare suggests slaughtering an animal close to its origin and then transporting the carcass to its final destination, individual Member States have such conflicting interests that no consensus can be reached on the issue. Safeguarding animal welfare during transport is therefore crucial.

In our study, the majority of the trucks came from France, transporting mainly bovines. This was expected, since France exports millions of stocker calves to other EU countries each year for fattening [7]. Spain was the second most common provenance, but these vehicles were transporting sheep/goats; this was also expected, since Spain has the second-highest number of sheep (a total of 18,136,050) amongs<sup>t</sup> European Union Member States [24]. The transport companies were mainly Greek, probably because the final destination was Greece. Indeed, Greece has the highest per capita consumption of sheep meat in Europe and is a major importer of sheep [25]. However, such data lead us to reflect on the importance of live animal transport for the Greek economy and of the training of Greek livestock hauliers. Europe is currently the only part of the world where hauliers are required to undergo training, but according to a 2007 survey, only 33.3% of the hauliers had attended training courses provided either by the Local Health Authority or trade organizations [26]. More recently, a survey proved the need of additional education and training for livestock drivers in Denmark [27]. Consequently, courses for transporters should be implemented and promoted worldwide, since training for all those involved in live animal transport has been identified as a key factor for improving welfare outcomes among transported animals [28].

Mortality is often used as an animal welfare indicator, since it is beyond doubt that such deaths are preceded by a period of suffering and poor welfare [14]. In relation to transport, mortality has often been reported as DOA, down on trailer or before weighing and total dead [29]. In our case, DOA, DCP and total dead showed an overall mortality rate within the range reported in the literature [25]. In our study, the mortality rate for bovines was similar to the rates reported in Canada (0.011%) [16] and in Czechia (0.012%) [30]. Contrastingly, the mortality for sheep/goats was outside the range (0.006–0.018%) reported in literature [31]. The highest mortality rate was found in lambs, probably because they are more likely to suffer from protracted transport stress than adult sheep and because they are more likely to be transported in large loads [31].

The mortality in pigs was 0%, lower than reported by Averós et al. in 2010 [17]. However, the latter authors focused on weaned piglets, while in our study fattened or breeding pigs were being transported, so they were older and often being transported in single compartments. All these factors may have helped the animals cope better with the long journey, thus reducing the overall mortality rate [14]. However, the number of pig transport records in this study is low, so this should also be taken into account when interpreting our data.

The morbidity rate was lower than reported in literature [2,16,29] and the reasons for this discrepancy may be due to a variety of factors. Firstly, the criteria for assessing morbidity in the above mentioned studies varied from one author to the next, while our reports were filed by OVs based on the criteria set out in Council Regulation (EC) No. 1/2005 for the fitness for transport. The OVs stopped only those animals showing severe injuries and diseases; consequently, no minor injuries or pathologies were counted in our dataset, thus reducing the morbidity rate. Secondly, the enforcement of a correct assessment of fitness for transport pre-departure may have played a role in reducing morbidity. While Marlin et al. (2011) reported that many transport related injuries assessed at unloading were due to animals' pre-existing poor health and welfare status, the number of animals stopped at departure was not considered in our dataset. Finally, the majority of publications assess morbidity at the slaughterhouse, while our data are the first to record it at a CP; our animals could have had time to recover from the long journey at the CP, being milked and humanely handled by the technicians working there. It has recently been shown that sheep rested for at least 16 h at a CP recovered from the stress, dehydration and fatigue induced by a 29-h journey [11]. The goal of CPs is indeed to give the animals the time to recover from the effects of a long journey before being loaded and transported again. Unfortunately, there are very few scientific studies on the effects of different managemen<sup>t</sup> systems and resting period at CPs [10,11]. Our data could not therefore compared with the literature. Further studies should be conducted comparing transport related morbidity at different CPs and in more countries.

In this study, sheep/goats were found to be more likely affected by transport than the other types of animals. It has been reported that sheep usually cope better with transport than other species [31], though there have been occasional reports of high mortality among single loads of animals during journeys that are non-compliant with regulations, such as an infamous case of 65 deaths out of 400 sheep being transported from Poland to France [31]. Similarly, our rate may have been affected by the fact that 14 out of 214 lambs were found dead on arrival in a single vehicle, which proved to be not compliant with EU regulation on space allowance and fined for overcrowding. It has been assumed that transport is less tiring for sheep/goats because as opposed to cattle and horses, both lambs and adult sheep lie down during transport [14]. However, studies have documented that transport reduced resting and rumination behaviours of lambs which showed signs of stress soon after loading and high levels of dehydration and weight loss after journeys lasting 12 h [31,32]. The number of inspections should be increased on trucks transporting lambs in Europe.

Overcrowding has been identified as a risk factor for transport-related health and welfare issues and it was the most frequent infringement observed during vehicle inspections carried out between 2001 and 2010 at the border between France and Italy [7]. Council Regulation (EC) No. 1/2005 gives a stocking density range for bovines and sheep/goats expressed in m2/animal, which vary according to the category of animal transported. The range for lambs goes from 0.20 to 0.30 m2/animal. However, it has been shown that stocking density should not be expressed in square meters per animal but in square meters per 100 kg to allow lambs to lie down and cope better with transport stress, and that lamb health and welfare may be affected by high stocking density [32]. Our data may be useful to implement the existing European Transport Guidelines and further studies should be performed to determine the optimal stocking density of lambs transported over long journeys within Europe.

Journey duration was confirmed as the most important risk factor in the development of transport-related diseases [33]. Surprisingly, in our data set there was no association between welfare outcomes and provenance. One reason for this finding may be related to the fact that the journeys had to follow a fixed route which was checked and approved by OVs before departure. However, it may also be due to the fact that we could not ascertain the exact journey duration in our dataset and that we used provenance as a predictive variable. Further studies should be conducted on a larger dataset using journey duration as a predictive variable to ascertain our findings.

Season and month have been identified as risk factors for transport-related mortality [29,30,33]. However, in our study, neither season nor month were significant. Our results are similar to those reported for pig transport by Gosálvez et al. 2006 [34], who explained their results by Spanish pig hauliers taking precautions to protect animals from extreme conditions, such as undertaking journeys at night, reducing loading densities and showering animals. Apart from those precautions taken by the hauliers, our results might also be due to the development of new vehicles. Those vehicle must have appropriated ventilation systems, temperature monitoring and recording systems [35], to comply with the new standards required by Council Regulation (EC) No. 1/2005 for journeys exceeding 8 h. These new technologies may have enhanced the welfare of the transported animals, assuring thermal comfort throughout the year. Overall, our data should be interpreted with caution and confirmed using a larger dataset.

The NATT was affected by many of the factors studied, in particular by species and category. There was a high average number of transported animals among vehicles transporting sheep/goats or younger—and consequently less heavy—animals. Since species and category were associated with other factors such as provenance, nationality of the transport company and destination, the effects of these latter factors on the number of transported animals per truck should be interpreted as a consequence. For instance, since sheep/goats came mainly from Spain and Hungary, all trucks coming from those countries contained higher than average numbers of transported animals. The association between species and category with season or month could, however, reflect consumer trends

(e.g., eating more meat in winter; traditional consumption of lamb during religious events) and the typical commercial life of each animal. For instance, lambs are usually born in spring, weaned after three months and are then shipped to feed lot or slaughter; this could explain why we found that the transport of sheep/goats was positively associated with summer months.

Our data should be considered preliminary because this study was limited by a number of factors. Firstly, due to the small number of events (welfare problem occurred) the logistic model was likely to suffer from small-sample bias. Secondly, the surveillance reports were cumulative by truck, which made it impossible to perform a proper risk analysis based on single transported animals. Thirdly, as the assessment of welfare outcomes was related only to death and severe pathology, in compliance with Section 5 of Annex 1 to Council Regulation (EC) No. 1/2005, many minor injuries and pathologies were not recorded and consequently were not analyzed in this manuscript. Finally, as previously discussed, our dataset was unable to include the health condition of the animals before the journey, the exact duration of the journey and stocking density expressed in m2/100 kg. Notwithstanding these limitations, this is the first paper reporting statistics based on surveillance reports filed by official veterinarians for livestock transported from Northern Europe to the Balkan Peninsula and transiting through a control post in Southern Italy.
