Transmission of Influenza A Virus in Pigs

Authors


M. Torremorell. Department of Veterinary Population Medicine, College of Veterinary Medicine, University of Minnesota, 385 Animal Science/Veterinary Medicine Building, 1988 Fitch Avenue, St Paul, MN 55108, USA. Tel.: +1 612-625-1233; Fax: +1 612-625-1210; E-mail: torr0033@umn.edu

Summary

Influenza A virus infections cause respiratory disease in pigs and are a risk to public health. The pig plays an important role in influenza ecology because of its ability to support replication of influenza viruses from avian, swine and human species. Influenza A virus is widespread in pigs worldwide, and influenza A virus interspecies transmission has been documented in many events. Influenza A virus is mostly transmitted through direct pig-to-pig contact and aerosols although other indirect routes of transmission may also exist. Several factors contribute to differences in the transmission dynamics within populations including among others vaccination, pig flow, animal movement and animal introduction which highlights the complexity of influenza A transmission in pigs. In addition, pigs can serve as a reservoir of influenza A viruses for other pigs and other species and understanding mechanisms of transmission within pigs and from pigs to other species and vice versa is crucial. In this paper, we review the current understanding of influenza virus transmission in pigs. We highlight the ubiquity of influenza A virus in the pig population and the widespread distribution of pandemic H1N1 virus worldwide while emphasizing an understanding of the routes of transmission and factors that contribute to virus spread and dissemination within and between pig populations. In addition, we describe transmission events between pigs and other species including people. Understanding transmission is crucial for designing effective control strategies and for making well-informed recommendations for surveillance.

Introduction

Influenza A virus is an enveloped negative-stranded RNA virus with a segmented genome containing eight gene segments. Influenza A viruses are continuously changing through point mutations that result in gradual change (genetic drift) or by gene reassortment that result in dramatic change (genetic shift). The mechanisms by which influenza A viruses cross species barriers and become transmissible within and between species are poorly understood and appear to be complex and polygenic (Parrish et al., 2008; Taubenberger and Morens, 2009).

Transmission of influenza A virus is complex to say the least. Influenza virus was first recognized as a viral agent causing respiratory disease in pigs in 1918 (Koen, 1919; Shope, 1931). The pig plays an important role in influenza ecology because of its ability to support replication of influenza viruses from avian, swine and human species. The pig has both α2-3 and α2-6 linkages on the glycocalyx of epithelial cells lining the pig respiratory tract that serve as receptors for avian and mammalian influenza A viruses, respectively (Ito et al., 1998). These receptors and the possibility of dual infections can result in the generation of new reassortant strains. However, the distribution of those receptors and the limited replication of avian influenza viruses in swine complicate the understanding of the transmission events between pigs and avian or other mammals (Taubenberger and Kash, 2010). The pig has been implicated in many interspecies transmission events. More recently, the pig has regained attention because of the 2009 pandemic events. While the direct precursors of the 2009 pandemic virus have never been isolated in pigs, the closest ancestors to the 2009 viruses – classical swine, triple reassortant and Eurasian ‘avian-like’ swine virus lineages have been detected in swine for several decades suggesting that the virus evolved in pigs for a long time (Garten et al., 2009; Smith et al., 2009). In addition, influenza virus is considered endemic in the swine population worldwide and risk of transmission events within and between species exists.

In this paper, we review the current understanding of influenza virus transmission in pigs. We highlight the ubiquity of influenza A virus in the pig population and the widespread distribution of pandemic H1N1 virus worldwide while emphasizing an understanding of the routes of transmission and factors that contribute to virus spread and dissemination within and between pig populations. In addition, we describe transmission events between pigs and other species including people. Understanding transmission is crucial for designing effective control strategies and for making well-informed recommendations for surveillance.

Influenza A Virus is Ubiquitous in Pigs

There are many published studies directed at quantifying the occurrence of influenza infection in pigs. Most of these studies are research-based studies with the goal to report prevalence of infection in pigs. Detection of influenza virus infection can be accomplished by either detecting a specific immune response to the virus (antibodies) or by directly detecting the virus or its genetic material by either molecular methods (i.e. RRT-PCR) or virus isolation. In addition, the recovery of influenza viruses is a key component of surveillance programs to identify new viruses of zoonotic or pandemic potential. Overall, the information presented in this section illustrates the ubiquitous nature of influenza virus in pigs and highlights the importance of understanding transmission events.

Until the 2009 pandemic, very few countries (i.e. Norway) had documented records and could be considered negative to influenza infections in pigs (Hofshagen et al., 2009; Grontvedt et al., 2011). Influenza virus antibodies in swine have been found in many countries around the world highlighting the presence of the virus in different areas of swine production in most continents. In Europe, a Spanish seroprevalence study concluded that 93 out of the 98 participating farms had antibodies for influenza virus. Specific hemagglutination inhibition (HI) subtype antibodies were detected in 92.9%, 64.3% and 92.9% of the farms for influenza virus subtypes H1N1, H1N2 and H3N2, respectively. Additionally, antibodies to more than one subtype were found in 87.8% of the farms (Simon-Grife et al., 2011) which could suggest that more than one virus was present at these farms or on the other hand, these findings are the result of cross-reacting antibodies. In UK, a seroprevalence study tested samples from a serum bank that included 2000 culled sows. Based on HI, authors concluded that 59.7% of the sows had antibodies to one or more influenza subtypes. Interestingly, antibodies to influenza B and C were found in 8 and 198 sows, respectively (Brown et al., 1995). A recent British seroprevalence study including 2780 serum samples from 146 farrow-to-finish farms concluded that 52% of the participating farms had antibodies to at least one subtype when samples were tested by HI (Mastin et al., 2011). Van Reeth et al. (2008) conducted a seroprevalence study in seven European countries including Belgium, the Czech Republic, Germany, Italy, Ireland, Poland and Spain. A total of 4190 sow serum samples originating from 651 farms with at least 100 sows which had never been vaccinated for influenza were tested by HI. All seven countries tested positive for influenza antibodies; however, the Czech Republic, Ireland and Poland had lower seroprevalence compared to the other participating countries. Overall, the H1N1 subtype was the most widespread followed by H1N2 and H3N2 subtypes.

Asian countries have also reported the existence of influenza antibodies in pigs. A Chinese study using pooled data analysis from 19 publications estimated that the seroprevalence of H1 and H3 antibodies was 31.1% and 28.6%, respectively (Liu et al., 2011). In Korea, 14 finishing pig samples per farm were collected from 53 different farms. Antibodies to H1N1, H3N2 and to both viruses were found in 51.2%, 43.7% and 25.3% of the samples, respectively (Jung et al., 2007). A different Korean study was conducted by Pascua et al. (2008) in which serum samples submitted for diagnosis at a university laboratory were tested by HI. A total of 6418 growing-finishing pig sera were used in this study from which 2959 (46.1%) had antibodies to influenza A virus. It was estimated that 41.5% of the samples had antibodies for H1, whereas only 3.7% had antibodies for H3; however, 0.9% of the samples tested positive for both subtypes. Studies in Malaysia (Suriya et al., 2008) and India (Chatterjee et al., 1995) also detected antibodies to both H1N1 and H3N2 viruses in pig populations.

In North America, several influenza seroprevalence studies in the 1970s, 1980s and 1990s in the Unites States detected antibodies to all three major subtypes (Hinshaw et al., 1978; Chambers et al., 1991; Olsen et al., 2000). The latest seroprevalence study conducted in the United States was conducted by Choi et al. (2002) in which 111 418 samples submitted to the University of Minnesota Veterinary Diagnostic Laboratory were screened for influenza A antibodies by HI. Out of the total number of sera tested, 22.8% had antibodies to H1N1 or H3N2. Sixty-six per cent of the total number of positive samples accounted for H1N1 infections, whereas the remaining positive samples were because of H3N2. However, authors stressed that detected antibodies could be the result of either natural infection or vaccination. In Canada, prevalence studies in the province of Ontario concluded that 83.1% and 40.3% of the sows and finishing pigs tested, respectively, were positive to H1N1 virus (Poljak et al., 2008a). In 2003, sows tested for H3N2 antibodies resulted in a total of 9.2% and 7.9% of the sows testing positive to the Colorado and Texas strains, respectively (Poljak et al., 2008a). In 2004, seroprevalence estimates in finishing pigs were 13.4% and 2.7%, for H1N1 and H3N2, respectively. The following year, the prevalence for H1N1 increased to 14.9% and for H3N2, the increase was more dramatic, 25.9% (Poljak et al., 2008b).

In South America, an Argentinean study reported findings on a small-scale seroprevalence study involving 13 farrow-to-finish farms. Antibodies to H1N1 and H3N2 were detected in 89% and 73% of the samples, respectively. Additionally, 62% of the sera tested positive for both subtypes by HI (Pineyro et al., 2010). A Brazilian study detected antibodies to influenza A H3N2 in 46% of the samples tested by HI from a total of 675 samples (Caron et al., 2010), and antibodies against influenza B were also detected (Mancini et al., 2006).

Besides commercial farms, wild pig populations have serological evidence of influenza A virus infections. A multi-state project was conducted in the United States in which blood samples from feral swine population were collected in six states found 1% of the samples positive in Mississippi, 5.3% in California and 14.4% in Texas. Seropositivity against the H3N2 subtype was the most common with all samples except four yielding positive results against this serotype (Hall et al., 2008). Another study focusing on investigating the prevalence of influenza A in feral swine located in dense commercial pig farm regions, North and South Carolina, resulted in higher prevalence levels. Antibodies to three different antigenic variants of H1N1 (human-like H1N1 in 73% of the samples), reassortant avian/swine-like H1N1 (7% of the samples), classical-like H1N1 (14% of the samples) and in 47% of the samples for H3N2 were detected (Corn et al., 2009). A recent study tested 50 serum samples from wild pigs originating from Arkansas, Louisiana, Oklahoma and Texas generated similar results in that 2% and 40% of the samples were positive for H1 and H3 influenza antibodies, respectively (Baker et al., 2011). In Europe, a study conducted in specific German regions found that 5.2% of wild pigs had antibodies to both H1N1 and H3N2. Two viruses were isolated from samples collected in the same study (Kaden et al., 2008). These studies indicate that influenza A viruses do circulate in the wild pig population but at a lower rate.

On the other hand, detection of the virus in pig samples (i.e. nasal swabs) has been performed as part of surveillance programs in different countries. Results from these studies may not be directly comparable to serologic studies given that the chance of virus detection in the pig is lower because of the short time period of shedding compared to the more likely detection of antibodies which may be detectable for several weeks. Most of the surveillance programs in place are passive relying on the submission of samples to diagnostic laboratories. A few active surveillance programs exist in North America and Asia with samples actively collected in representative farms or at slaughter.

In the United States, there have been two studies in which nasal swabs were collected from pigs at different slaughterhouses with the aim of isolating the virus. In the first study, 478 (5.1%) of a total of 9400 samples were positive for influenza virus (Hinshaw et al., 1978). On the second study, 26 (2.2%) out of 1200 samples yielded live virus in which 24 out 26 corresponded to swabs collected between October and January (Olsen et al., 2000).

Furthermore, a 3-year virological passive surveillance study conducted in Belgium, UK, Italy, France and Spain aimed at characterizing the current circulating influenza viruses in these five countries successfully isolating 169 viruses. Isolation of influenza viruses was possible throughout the year especially during the cold seasons (i.e. winter and spring). The majority (81.9%) of these viruses originated from pigs 6 months old or younger (Kyriakis et al., 2011).

In Hong Kong, 184 (3.7%) viruses were isolated from a total of 4957 tracheal swabs collected at a slaughter plant (Peiris et al., 2001). Another study conducted in Taiwan collected 881 nasal swabs from pigs at farms that previously tested positive for influenza A antibodies and 7 (0.79%) tested positive for the nucleoprotein gene by RRT-PCR (Shieh et al., 2008).

Overall, all these studies provide enough information to consider influenza A virus ubiquitous in the global pig population. A similar picture may exist in humans and avian species around the world which further highlights the importance of focusing our efforts on understanding and preventing the risk of transmission between species.

Worldwide Distribution of the 2009 Pandemic H1N1 Virus in Swine

The 2009 pandemic H1N1 influenza virus contains a combination of gene segments from both North American and Eurasian swine lineages that has never been detected previously (Garten et al., 2009; Smith et al., 2009). Further analyses concerning the 2009 H1N1 pandemic virus and the evolution of influenza viruses in general have provided important information regarding the potential origin of this novel virus. It has been suggested that the virus may have been circulating in swine, albeit undetected, for some time prior to its initial identification (Garten et al., 2009; Smith et al., 2009). Southern China has received much focus because of the presence of both North American and Eurasian swine lineage viruses in pigs in those provinces. Reassortment events between North American and Eurasian influenza viruses have led to the establishment of stable virus lineages in swine and these reassortment events were shown to occur in China before and after the pandemic. However, a direct 2009 pandemic H1N1 precursor virus was not recognized (Lam et al., 2011; Vijaykrishna et al., 2011).

Infection of a commercial swine farm with the 2009 pandemic H1N1 virus was first confirmed in May 2009 in Alberta, Canada (Howden et al., 2009). Following the description in Canada, the 2009 pandemic H1N1 virus was identified in many other herds around the world (Hofshagen et al., 2009; Moreno et al., 2010; Pasma and Joseph, 2010; Pereda et al., 2010; Song et al., 2010; Sreta et al., 2010; Welsh et al., 2010; Forgie et al., 2011). These worldwide reports regarding the incursion of the 2009 pandemic H1N1 virus not only describe some of the initial 2009 pandemic H1N1 infections in swine herds, but they also provide details regarding clinical signs, pathogenesis, source of virus and the course of infection within herds. Gray and Baker (2011) have summarized the reports to the world organization for animal health (OIE) of pandemic H1N1 virus infections in pigs, indicating that the virus is present worldwide.

Human-to-swine transmission of 2009 pandemic H1N1 influenza virus was suspected in a majority of these case descriptions. Forgie et al. (2011) provided detailed epidemiologic data from a Canadian swine research farm in which humans were infected before swine and swine became infected following contact with the human index case. Hofshagen et al. (2009) described the spread of 2009 pandemic H1N1 influenza virus in Norway, previously free of classical swine influenza, in which humans infected with 2009 pandemic H1N1 influenza virus were the most likely source of virus. However, in many of these cases, definitive evidence of human-to-swine transmission was lacking, and human-to-swine transmission was suspected after other potential sources were ruled out. The 2009 pandemic H1N1 virus is also known to transmit efficiently between pigs, making the introduction of infected pigs into a population another likely route of virus entry into populations (Brookes et al., 2009). The detailed description of 2009 pandemic H1N1 virus infections within herds also pointed to the fact that this virus behaved similarly to endemic swine viruses and all susceptible swine populations were likely able to become infected and transmit the virus.

The worldwide presence of the 2009 pandemic H1N1 virus provided the opportunity for reassortment events between this virus and endemic or other viruses. A novel reassortant virus (A/swine/Hong Kong/201/2010) was identified in swine through systematic virological surveillance in Hong Kong in January 2010 (Vijaykrishna et al., 2010). Additional reassortant viruses derived from the 2009 pandemic H1N1 virus were then discovered including novel H1N2 viruses in Italy and Argentina, and novel H1N1 reassortant viruses in Germany, Southern China, Thailand, and Argentina (Kitikoon et al., 2011; Moreno et al., 2011; Pereda et al., 2011; Starick et al., 2011; Zhu et al., 2011). Similarly, nine pandemic/endemic reassortant influenza viruses were recently described and isolated during syndromic and active surveillance from across the United States (Ducatez et al., 2011). These reports further stress the importance of influenza virus surveillance in pig populations to monitor the evolution of the 2009 pandemic H1N1 virus and endemic viruses. Furthermore, the significance of the acquisition of genes from the 2009 pandemic H1N1 into endemic influenza viruses found in swine may impart a change in transmissibility, replication or viral fitness that should be earnestly evaluated through controlled studies.

Influenza Virus Transmission in Pigs

Transmission routes

The general routes of influenza virus transmission include aerosol, large droplet and direct contact with secretions of infected individuals or contaminated fomites (Tellier, 2006). Influenza virus transmission via direct contact with infected pigs has been observed in many experimental studies and is thought to be a major transmission route. Direct contact transmission was evident during the discovery of the causative agent of influenza in pigs and with more recent influenza viruses in pigs including the 2009 pandemic H1N1 virus (Shope, 1931; Brookes et al., 2009; Lange et al., 2009). Both sick and subclinically infected pigs likely play a large role in the transmission of influenza virus within and between swine herds, highlighting the importance of controlled animal movement practices to minimize the transmission of infectious agents, including influenza virus. Influenza virus in pigs is not transmitted through semen.

Influenza virus transmission without direct contact with infected individuals is also assumed to take place in field settings. For example, water contaminated with bird faeces has been implicated as a source of influenza virus in several swine outbreaks involving avian origin viruses (Karasin et al., 2000, 2004; Ma et al., 2007). Routes of infection in wild pigs are not well known but exposure is likely related to contact with either wild bird droppings or contaminated water. Transmission of influenza virus via other indirect routes such as aerosols and fomites has been documented in species other than pigs (Lowen et al., 2006; Mubareka et al., 2009; Yee et al., 2009a,b). These transmission routes have not been studied in detail in pig populations, but field observations suggest they are important and could include aerosol and fomite transmission (Tofts, 1986; Desrosiers et al., 2004; Poljak et al., 2008b). Influenza virus has also been detected in air samples from rooms of experimentally infected pigs (Loeffen et al., 2011), in the exhaust air from infected farms and at 1 mile from an infected farm (Corzo, C. and Torremorell, M., personal communication) highlighting the potential for aerosol transmission in pigs and between farms. In humans, mathematical models have suggested that the airborne route may be the dominant route of influenza transmission (Atkinson and Wein, 2008).

Transmission through insects has been suspected for influenza A virus. Avian influenza H5N1 has been isolated from blood-feeding mosquitoes (Barbazan et al., 2008). In the case of swine, blood-borne transmission via insects is not likely to play a role because influenza virus has been rarely isolated from blood samples (Romijn et al., 1989). Additionally, influenza A virus has been detected in blow flies (Sawabe et al., 2006, 2009), which highlights the potential of insects as mechanical vectors. However, there is no information on the actual importance of this route of viral dissemination in pigs.

Transmission through transport has not been considered an important route for influenza transmission in pigs. However, long distance pig movement has been implicated in the spatial dissemination of influenza viruses of human origin from swine production areas into the Midwest US swine population (Nelson et al., 2011). In addition, global transport of infected pigs may also be implicated in the movement of virus strains across countries and continents. In China, there are North American and Eurasian swine lineage viruses co-circulating suggesting that international trade may have facilitated the introduction of those viruses (Vijaykrishna et al., 2011).

Overall, further research is needed to fully quantify the impact of indirect influenza virus transmission routes in pigs. This would bring resolve to cases in which transmission takes place in the absence of direct contact with infected pigs and allow for the development and establishment of biosecurity measures to minimize influenza virus dissemination.

Quantification of influenza virus transmission

While transmission experiments and models exist for influenza virus in multiple mammalian and avian hosts, few have been conducted in pigs. Transmission experiments allow one to assess the spread of infection within a population via estimation of the reproduction ratio (R) and allow one to assess the effect of control measures in a population (Velthuis et al., 2007). The basic reproduction ratio (R0) is defined as the expected number of secondary cases in a completely susceptible population because of a typical infectious individual during its entire infectious period (Diekmann et al., 1990). In general, when R0 is >1, an infection will spread in a population, and when R0 is <1, the infection will not spread in a population. Therefore, the impact of control measures on R0 will highlight the ability to control the spread of influenza virus within a population.

One tool that may impact the reproduction ratio is vaccination. Several studies have estimated the reproduction ratio for avian influenza virus and have shown that vaccination may reduce transmission (van der Goot et al., 2005, 2007; Bos et al., 2008; Bouma et al., 2009). However, vaccination does not always reduce transmission of avian influenza virus and the impact of vaccination may vary across species (van der Goot et al., 2007; Poetri et al., 2011). Therefore, it is important to assess virus transmission and the impact of control measures in each specific host species.

Influenza virus has been traditionally characterized as a pathogen that moves rapidly in swine herds, resulting in high morbidity (Olsen et al., 2006). However, experimental quantification of virus transmission and the impact of control measures in pig populations are limited. Multiple examples of transmission from infected to susceptible pigs exist, as well as transmission from infected to seropositive contact pigs (Choi et al., 2004a; Brookes et al., 2009; Lange et al., 2009). More recently, an ‘avian-like’ H1N1 influenza virus was shown to transmit through four pairs of vaccinated pigs at antibody levels thought to protect against infection (Lloyd et al., 2011). Influenza viruses isolated from pigs were also able to transmit efficiently via direct contact in a ferret model (Yen et al., 2011). Influenza virus transmission was recently quantified in non-vaccinated and vaccinated pig populations with a reproduction ratio estimate of 10.66 in non-vaccinated pigs and reproduction ratio estimates of 1 and 0 for pigs vaccinated with heterologous and homologous inactivated vaccines (Romagosa et al., 2011). A follow-up transmission study identified a similar reproduction ratio estimate in non-vaccinated pigs and a reduction in transmission parameters in pig populations with homologous maternal immunity (Allerson, M. and Torremorell, M. unpublished results). Further study concerning influenza virus transmission in pigs may elucidate measures that can mitigate transmission and reduce the burden of influenza virus in pig populations.

Herd and regional risk indicators

Various risk indicators for influenza virus infection at the herd level have been identified across several different swine producing regions in North America, Europe, and Asia. The most common risk indicators at the herd level identified in these studies include high pig and/or farm density, large herd size, high replacement rates and import/purchase of pigs (Ewald et al., 1994; Maes et al., 2000; Poljak et al., 2008a; Suriya et al., 2008; Mastin et al., 2011; Simon-Grife et al., 2011). Other farm management and design factors such as farrow-to-finish farms, open partitions between pens, and uncontrolled access to farms have also been associated with herd infection (Poljak et al., 2008a; Simon-Grife et al., 2011). Recently, a British study concluded that the number of pigs per water space and indoor housing was positively associated with seropositivity. Two factors were found as protective, straw bedding and sampling season in which summer months (July–September) had the lowest likelihood of seropositivity compared to other seasons (Mastin et al., 2011).

Transport per se has not been implicated as an important route for influenza transmission in pigs. However, a recent publication by Nelson et al. (2011) indicated that long distance pig movement from Southern US states to the Midwest was responsible for the spatial dissemination of influenza viruses of human origin into the US swine population. This analysis indicated that indirect routes of influenza transmission may play a role in the evolution and spatial dissemination of influenza virus in swine. This information is particularly important as the structure of the swine industry has changed significantly during the last 20 years and continues to evolve in developing countries. The adoption of multi-site production systems with increased pig movement, larger facilities and increased regional density may be facilitating the dissemination of significant swine pathogens including influenza virus.

Influenza virus population dynamics

Estimates at the herd level indicate that influenza virus infections are common with a seroprevalence of 83% in sow herds in Ontario, Canada and over 90% in sow herds in Belgium, Germany, and Spain (Poljak et al., 2008a; Van Reeth et al., 2008). The commonality of influenza virus infections and the seemingly perennial nature of the virus in pig populations have brought forth several suggestions regarding maintenance of endemic influenza virus infections in pig populations. Historically, reoccurring influenza virus infections were thought to be because of long-term carrier pigs and intermediate hosts (Shope, 1941; Blaskovic et al., 1970). While influenza virus replicates almost exclusively in epithelial cells of the respiratory tract of pigs, infectious virus has been detected in brain and this finding could also have implications on virus transmission (De Vleeschauwer et al., 2009; Loeffen et al., 2011). In both studies, the transference of virus from the nasal turbinates, nasopharynx or tonsils into the brain tissue during the dissection process was not ruled out conclusively. Further study concerning the duration of infection and infectiousness in individual pigs has shown that the duration is relatively short and raises serious doubts regarding the potential existence of carrier pigs (Vannier et al., 1985; Clavijo et al., 2002; Romagosa et al., 2011).

The maintenance of influenza viruses at the population level can also be justified via the continued infection of susceptible pigs introduced or produced in herds (Brown, 2000). The maintenance of influenza virus within herds over time has also been suggested following observational studies in pig populations over extended periods (de Jong et al., 2001; Poljak et al., 2008b). These temporal observations are further strengthened by recent work identifying influenza virus positive neonatal pigs in the absence of positive sows in breeding herds, with susceptible neonatal pigs potentially serving as the maintenance host over time (Larsen et al., 2010). Therefore, virus circulation among weaned pigs, which was evident in up to one half of Dutch breeding herds, could serve to maintain infections within herds and as a source of virus dissemination (Loeffen et al., 2003).

Understanding the population dynamics of influenza virus can help lead to effective strategies to eliminate influenza virus from herds. An H3N2 influenza virus was eliminated from a three-site pig herd by altering gilt introductions and the handling of suckling pigs in combination with nursery and finish depopulation (Torremorell et al., 2009). Strict all-in/all-out procedures by nursery room were unable to eliminate influenza virus in the nursery of the three-site herd mentioned above, likely due to indirect exposures within site. This indicates that the structure of a pig herd can impact the population dynamics of infection. This was shown when the incidence of influenza virus infection was found to be higher at the beginning of the finishing period in farrow-to-finish herds, whereas in specialized finishing herds, the incidence of infection was highest at the end of the finishing period (Loeffen et al., 2009). Clearly, the structure of the pig herd must be taken into account when assessing influenza virus infection dynamics and control at the population level.

Environmental survival

Survival of influenza virus outside the host is an important component of transmissibility because transmission does not necessarily need to involve an effective contact with an infectious individual. Research on the ex-vivo survival of influenza virus has been conducted under different conditions, and data on swine-origin viruses are scarce.

Relative humidity and temperature

Information regarding the relationship between survivability of influenza virus and relative humidity together with temperature provides an idea of the boundaries this virus has on survival. Data on survivability will depend on where the virus is suspended (i.e. aerosol, water, solid material) and on the nature of the experiment. For instance, a study looking into transmissibility of the virus in the guinea pig model concluded that aerosol transmission is more efficient at both low RH (20–35%) and temperature of 5°C compared to 20°C (Lowen et al., 2007). Such finding may be related to the stability of the virus in the air because it has been reported that the virus is more stable at low RH (Schaffer et al., 1976) and low temperatures (Harper, 1961). Similar results were reported in a different experimental setting in that mean titre reduction of influenza virus was lower at 20% RH compared to 84% RH when suspensions were kept at room temperature (Buckland and Tyrrell, 1962).

Water

The persistence of viable influenza viruses in water has been studied and information is mainly limited to avian influenza viruses because of the importance that water fowl and other avian species have on viral dissemination. Survivability of avian influenza viruses increased with decreasing water temperature. The virus was still infective for 207 days when the water temperature was 17°C compared to 102 days when water temperature was 28°C with an initial concentration of 106 TCID50/ml (Stallknecht et al., 1990b) indicating that low temperatures may contribute to viral survivability. Water characteristics (pH and salinity) together with temperature were also evaluated, and it was concluded that the virus survived for longer periods at low temperature and salinity and high pH (Stallknecht et al., 1990a; Brown et al., 2007, 2009).

Fomites

Survival of the virus in different surfaces has been assessed using type A and B viruses. The virus survives more efficiently in non-porous (gumboot, latex, plastic, stainless steel, ceramic tiles, tire) surfaces compared to porous (cotton fabric, egg tray, egg shell, handkerchief, pajamas, polyester fabric, magazine, tissues, wood) surfaces (Bean et al., 1982; Tiwari et al., 2006). An influenza survival study on banknotes was conducted using different type A and one type B influenza viruses. The study agreed with the fact that virus survival on porous surfaces is poor compared to survival on non-porous surfaces; however, in this study, the presence of mucus extended in days the survivability in both influenza A and B viruses (Thomas et al., 2008) highlighting the important properties the mucus has for viral persistence on different surfaces.

Slurry

Even though influenza virus is not commonly detected in mammalian faeces (Kawaoka et al., 1987; Pinsky et al., 2010), transmission through the oral–faecal route in pigs may not be as important as it is with avian species because influenza viruses also replicate in the intestinal epithelium and thus is excreted through their faecal matter (Webster et al., 1992). However, it has been reported that influenza virus can survive in slurry for as long as 9 weeks if maintained at 5°C and 2 weeks if maintained at 20°C (Haas et al., 1995). A regular practice in today’s industry is manure spread as means of field fertilization, although it has not been proven, such practice may play a role on the regional transmission of the virus (Desrosiers et al., 2004).

Transmission between pigs and people

Following the first published description of clinical disease in pigs in 1918 and the isolation from pigs in 1930, influenza viruses have continued to evolve and impact animal and human health across the world (Koen, 1919; Shope, 1931). During the initial description of disease in pigs, it was noted that an outbreak among a family would occur at the same time as an outbreak in pigs (Koen, 1919). Serological evidence of influenza virus transmission from people to pigs under field conditions was later described (1938) in the United States and a human seasonal influenza virus isolate was first obtained from pigs in Taiwan (1969) (Shope, 1938; Kundin, 1970). Following the initial descriptions, influenza virus transmission between people and pigs has been examined more extensively and has been shown to occur sporadically over the last century.

The transmission of influenza virus between people and pigs mainly concerns type A influenza virus, as types B and C are thought to primarily infect humans. Transmission of influenza viruses between species occurs, but this is an uncommon event. There are many barriers that limit interspecies transmission including influenza virus receptor specificity; however, all barriers are not completely known and understood. Influenza virus replicates primarily in the respiratory tract of pigs and humans, and therefore, similar modes of transmission apply to both pigs and humans. Recent work with the pandemic H1N1 virus showed that viable virus was only detected in the respiratory tract of infected pigs, other samples including muscle were negative (Vincent et al., 2010). Therefore, flu transmission via consumption of pork is considered negligible.

In the United States, the classical H1N1 virus was the predominant subtype circulating in pigs until 1998. Several serologic studies in pigs during the time period before 1998 revealed that pigs were infected with H3 viruses, but prevalence of infection was quite low (Hinshaw et al., 1978; Chambers et al., 1991). A study in 1997/1998 indicated that pig exposure to human H3 viruses was greater than previously reported (Olsen et al., 2000). The diversity of influenza viruses then changed in the US in 1998 when a genetic reassortment of human H3N2, classical swine H1N1, and avian influenza virus genes was observed (Zhou et al., 1999). This reassortant virus became established in the North American pig population along with other diverse viral lineages including H1 viruses with a human origin HA gene (Zhou et al., 1999; Webby et al., 2004; Karasin et al., 2006; Vincent et al., 2009). Clearly, transmission of influenza viruses between people and pigs shaped the current epidemiology of influenza virus in North America.

In contrast to the observed low level of H3 influenza virus in North America prior to 1998, H3N2 influenza virus related to human seasonal viruses were circulating in Europe and Asia at seemingly higher levels since the 1970s (Kundin, 1970; Shortridge et al., 1977; Ottis et al., 1982; Haesebrouck et al., 1985). This is one example of how influenza virus epidemiology in pigs differs by geographic region. However, similar to North America, the epidemiology of influenza virus is quite complex with multiple subtypes and variants circulating in Europe and Asia. One common observation regarding the transmission of influenza virus between people and pigs is the appearance of the 2009 pandemic H1N1 virus in pig populations around the world.

The pandemic (H1N1) 2009 influenza virus was first detected in humans in April of 2009 (Centers for Disease Control and Prevention (CDC), 2009a). The virus then spread quickly across the world, causing disease in humans and other animal species, including pigs. The ability of the pandemic (H1N1) 2009 virus to establish in pig populations around the world was unknown when it was first detected in pigs in Canada (Howden et al., 2009). Soon after the case in Canada, the pandemic (H1N1) 2009 virus was identified in pig populations around the world (Moreno et al., 2010; Pasma and Joseph, 2010; Pereda et al., 2010; Song et al., 2010; Welsh et al., 2010). Human-to-pig transmission was suspected in most of the initial pig farm cases, even though definitive evidence was often lacking. Forgie et al. (2011) provide compelling evidence of human-to-swine transmission of the 2009 pandemic H1N1 virus on a swine research farm.

Influenza virus transmission between pigs and people has played a large role in the complex epidemiology and evolution of influenza virus in pig populations. In cases where a particular influenza virus can be transmitted efficiently in pig and human populations, it can be difficult to discern the direction of the transmission event, pigs to people versus people to pigs. Transmission of influenza viruses from pigs (swine-origin viruses) to people has been documented on sporadic occasions.

During routine animal husbandry practices, close contact between people and pigs occurs on a frequent basis. However, a swine influenza virus was first isolated from a man in 1974 (Smith et al., 1976). A swine-origin influenza virus of H1N1 subtype (A/New Jersey/76) was then isolated from soldiers at Fort Dix, NJ, USA in 1976 following a respiratory disease outbreak (Gaydos et al., 1977, 2006). However, the definitive source of the virus at Fort Dix is unknown and the virus did not spread outside of Fort Dix (Gaydos et al., 2006). Following the initial description of swine-origin influenza viruses in humans and the Fort Dix outbreak, additional swine-origin influenza virus infections of humans were observed.

One of these events includes probable human-to-human transmission following exposure to pigs at an agricultural fair in Wisconsin, USA (Wells et al., 1991). A recent literature review by Myers et al.(2007) revealed 50 cases of apparent zoonotic swine influenza virus infection and a review by Van Reeth (2007) displayed all documented human infections with swine-origin influenza viruses, including cases in Europe, Asia and North America. A recent summary concerning human infections with triple reassortant swine-origin influenza viruses (H1) in the United States revealed that 11 human cases were reported between 2005 and 2009, most having exposure to pigs (Shinde et al., 2009). The most recent example of this event was the isolation of reassortant H3N2 swine-origin influenza viruses, including the matrix gene segment of the 2009 pandemic H1N1 virus, from children from various states in the United States (CDC, 2011a). Most humans infected with swine-origin influenza viruses have reported previous direct or indirect exposure to pigs, and the course of disease is similar to that of typical influenza virus infections. However, some of these recent cases also suggest limited human-to-human transmission (CDC, 2011b). These cases over time illustrate the sporadic nature of interspecies transmission and the ultimate role of surveillance to identify these events in pigs and humans.

The clinical manifestations following swine influenza virus infection in humans do not seem to differ from those of typical influenza virus infections; however, many report previous contact with pigs (Myers et al., 2007).

In addition to documented cases of swine-origin influenza virus infections in humans, serological evidence has identified exposure to pigs as a risk of swine-origin influenza virus infection (Olsen et al., 2002; Myers et al., 2006; Terebuh et al., 2010; Gerloff et al., 2011). For example, farmers, veterinarians and meat processing workers with swine exposure were shown to be at higher odds of exposure to H1N1 and H1N2 swine influenza viruses versus controls (Myers et al., 2006). Additional studies have also shown that swine workers and their non-exposed spouses are at increased risk of swine influenza virus infection (Gray et al., 2007). More recently, swine workers in Europe were shown to have more frequent and higher titres to swine influenza viruses, including the pandemic H1N1 virus compared to controls (Gerloff et al., 2011).

Even though occupational exposure to swine has been shown to increase the risk of swine influenza virus infection, there is more involved in transmission events than just exposure and contact with infected pigs. This was highlighted by an H2N3 infection of swine in which transmission from ill pigs to employees was not evident (Beaudoin et al., 2010). A prospective cohort study which assessed the transmission of influenza virus between pigs and swine workers found serologic evidence for infection with human and swine-origin influenza viruses in swine workers (Terebuh et al., 2010). Additionally, this study identified a low rate of symptomatic infection and virus isolation among participant swine workers. Certain employee behaviours and attributes may impact the risk of human and swine-origin influenza virus infection, including the use of personal protective equipment, smoking and pre-existing immunity (Ramirez et al., 2006; Terebuh et al., 2010; Beaudoin et al., 2011).

Transmission between pigs and birds

Phylogenetic evidence suggests that all mammalian influenza A strains are ultimately derived from avian influenza A viruses (Wright et al., 2007). Therefore, it should not come as a surprise that transmission between avian species and pigs is possible and given the proper environment and exposure conditions, it can also be common.

Interspecies transmission of influenza A virus between pigs and birds has been documented in many occasions. Among domestic avian species, turkeys represent the species that is the most susceptible to become infected with swine influenza viruses based on literature reports. As an example, a family farm in Ohio, USA, raised both turkeys and pigs. Each species was housed in separate buildings that were approximately 12 m apart from each other. Two outbreaks of respiratory disease were seen in pigs after introduction of boars into the herd and subsequently in both occasions there was an egg drop in the turkey flocks. Diagnostic results concluded that both species had been infected by the same H1 influenza virus (Mohan et al., 1981). There have been numerous reports (Andral et al., 1985; Ficken et al., 1989; Suarez et al., 2002; Choi et al., 2004a,b; Tang et al., 2005) in which influenza viruses infecting turkey flocks have been isolated and characterized to be swine-origin influenza viruses. Interestingly, these reports mention that the probable source of the virus is a pig farm in the vicinity. Recently, a risk factor study conducted in Minnesota found an association between seropositivity to H1 and H3 viruses in turkey premises and proximity to pig farms (Corzo, C., unpublished results).

Infections of pigs with avian influenza viruses have been reported. Most of these infections are thought to originate from wild ducks based on phylogenetic analyses. In Europe, avian H1N1 viruses were transmitted to pigs in 1979 and became established as a stable lineage which is still under circulation (Pensaert et al., 1981; Kyriakis et al., 2011). Infections with different viral subtypes have been reported including infections with duck-origin influenza viruses H1N1 (Pensaert et al., 1981; Guan et al., 1996; Karasin et al., 2004), H3N2 (Kida et al., 1988), H3N3 (Karasin et al., 2004), H4N6 (Karasin et al., 2000), H7N2 (Kwon et al., 2011) and H9N2 (Peiris et al., 2001). There has also been serologic and virologic evidence of pigs infected with H5N1 in Vietnam and Indonesia. However, experimental data indicated that pig-to-pig transmission of this virus was not observed (Choi et al., 2005; Lipatov et al., 2008; Nidom et al., 2010).

Viral structure and molecular characteristics affecting transmission

The molecular viral characteristics of influenza A viruses that determine their ability to transmit within and between species are poorly understood. Hemagglutinin (HA) and neuraminidase (NA) are the main antigenic proteins of influenza A viruses and they are associated with receptor binding and virus release, respectively (Rossman and Lamb, 2011). Influenza A viruses adapted to humans have a higher affinity for α2-6 SA receptors, while influenza A viruses adapted to birds have a higher affinity to α2-3 SA (Medina and Garcia-Sastre, 2011). The pig has both receptors and is considered a prime species able to support viral replication of different origins. However, there is no specific NA–HA combination and/or amino acid mutations that have been identified to increase viral stability during interspecies transmission (Wright et al., 2007). Multiple mutations appear to be needed to produce a distinct HA (Wolf et al., 2006). Additionally, a new variant of HA may not increase viral fitness unless this new HA variant is linked to the appropriate genome segments that allow a good transmission and replication rate; therefore, a new HA may take some time within a population to find the compatible genetic segments that favour transmissibility or even disappear before finding them (Rambaut et al., 2008). In a recent study, the M segment of the pandemic 2009 virus (Cal/09) promoted aerosol transmissibility to recombinant viruses constructed with backgrounds of another human H1N1 strain [A/Puerto Rico/8/34 (H1N1)] and a pig strain [A/swine/Texas/1998 (H3N2)], both considered of low transmissibility, suggesting that the M segment is a critical factor supporting the transmission of the 2009 pandemic virus. This appears to be a remarkable feature of the 2009 pandemic H1N1 influenza virus given its efficient transmissibility in humans compared to precursor strains from the triple reassortant swine influenza virus lineage and may explain the rapid dissemination to other species as well. Nevertheless, more research is needed to fully understand the molecular characteristics suitable for influenza virus transmission in pigs and from pigs to other species.

Prevention of Influenza A Virus Transmission in Pigs

Biosecurity practices and vaccination remain the primary means of preventing or minimizing transmission of influenza A virus in pigs and from pigs to other species. Prior to introducing replacement animals into farms, animals should be kept in isolation/quarantine facilities and tested prior to releasing them into the units. Prevention of influenza dissemination by weaned animals will be difficult if animals are infected at weaning and if pigs are reared in separated sites located in distant geographical locations. As a result, regional dissemination may occur (Nelson et al., 2011), and measures directed at preventing the introduction of airborne infections should be considered. Air filtration has recently proven effective to control airborne infections of swine pathogens such as porcine reproductive and respiratory syndrome virus (PRRSv) and Mycoplasma hyopneumoniae (Dee et al., 2010), and it may also be effective at preventing the introduction of airborne influenza infections in pigs, although the cost effectiveness of this measure for influenza infections still remains to be evaluated. Bird proofing facilities and keeping avian species and their excretions and secretions away from pigs are also necessary.

Pig vaccination using commercial or autogenous vaccines is common. Vaccines decrease lesions, clinical signs and virus shedding (Choi et al., 2004a,b; Lee et al., 2007). Vaccines can also reduce transmission but it is not likely that they will prevent transmission completely (Van Reeth et al., 2001; Romagosa et al., 2011). Effect of vaccination on transmission will depend on strain homology, immunity levels and vaccine presentation. Vaccination of sows prior to farrowing is the most common protocol although pig vaccination may be beneficial in herds where influenza is a problem in growers/finishers. Whether pig vaccination has an effect on decreasing aerosol transmission and regional spread remains to be investigated.

Because people can infect pigs and vice versa, biosecurity practices need to also include measures directed at minimizing the risk of exposure of people to pigs and from pigs to people. A comprehensive prevention plan should restrict the entry of visitors, prevent the entrance of people with flu-like symptoms, implement shower-in/shower-out policies for all people entering the facilities or ensure that all people entering the facilities wash their hands and arms with warm water and soap before entering, and have farm personnel vaccinated with seasonal flu vaccines. Implementation of strict sick leave policies should also be strictly enforced. In addition, the use of personal protection equipment which includes face masks (or preferably respirators), eye protection and gloves should be encouraged (CDC, 2009b).

Sporadic elimination of influenza virus from swine herds has been reported before. Elimination may happen for certain herds as part of the natural course of the infection with the virus dying-out in all-in/all-out, discrete closed immune populations. In addition, virus elimination may also be possible following herd closure and partial depopulation procedures (Torremorell et al., 2009). Herd closure has been used for the elimination of other swine pathogens (Torremorell et al., 2002; Schaefer and Morrison, 2007) and consists in the temporal interruption of gilt introduction to allow the viral infection to die out while protective immunity develops in the herd. Introduction of negative animals after the closure period results in the development of a virus-free population. However, maintenance of influenza-free populations requires a deeper understanding of the relative risk for all routes of influenza transmission, and it is still premature to advocate the establishment of negative swine populations.

Conflict of Interest Statement

All authors declare no conflict of interests.

Ancillary