Dept title


Current state of knowledge on Highly Pathogenic Avian Influenza

2.3 Mode of Entry of H5N1 HPAI to Countries and Places

When infections occur in new locations, the sources to consider include live poultry, poultry products, vehicles, objects and materials (including feed and water) used on farms or in markets contaminated with virus, people (e.g. failure to wash hands after handling infected poultry, contaminated clothing or footwear), wild birds and trade in other types of birds. These various routes are examined in detail in Chapters 3 and 4.

There has been considerable speculation about the mode of entry of H5N1 HPAI virus into unaffected countries, especially concerning the relative role of trade in poultry and movement of free-flying wild birds. It is highly likely that both have played a role, although the role of wild birds has been questioned because surveillance studies on clinically normal wild birds have resulted in limited detection of H5N1 HPAI viruses.

Few detailed studies on the mode of introduction of H5N1 viruses have been conducted.  Evidence from HKSAR demonstrated how trade in live poultry could be a potential source of infection. Repeatedly, geese and ducks sent to a central slaughtering facility from mainland China to HKSAR between 1999 and 2001 were found to be infected on arrival (Agriculture, Fisheries and Conservation Department, 2001; Cauthen et al, 2000; Guan et al, 2002a). However, H5N1 viruses have also been found in free-flying non-domesticated birds in HKSAR, and these birds could potentially act as a means of entry to non-biosecure farms or zoological collections.

The large volume of trade in live chickens and other terrestrial poultry between Guangdong province in China and HKSAR was also implicated (but never proved) as a source of infection from 2001 to 2003. During this period, the H5N1 virus genotypes found on farms in HKSAR were less varied than those detected in HKSAR’s live-poultry markets (Guan et al, 2002b; Li et al, 2004a). These markets were supplied with poultry from farms in HKSAR and from Guangdong province, suggesting that local Hong Kong farms were not the only source of infection for these poultry markets.

The outbreak in the United Kingdom in February 2007 implicated trade in poultry meat as the most likely source of infection.

One study (Kilpatrick et al, 2006) assigned relative probabilities to trade in poultry 1 and in wild birds and also to natural movements of wild birds as the source(s) of infection in various countries. The study suggested that most of the introduction of H5N1 HPAI virus to Europe was probably via natural movements of wild birds. However, not all of the conclusions in this paper concur with those from field investigations – for example, the study predicted trade as the most likely source for Kazakhstan, but this was not fully supported by field observations (Gilbert et al, 2006b, citing a World Organisation for Animal Health [OIE] investigation). The Kilpatrick paper also suggested spread of virus from Thailand to Indonesia but, genetically, these viruses belonged to different clades. Furthermore, from the known or suspected time of introductions, infection probably occurred in Indonesia before it occurred in Thailand.

Regarding the relative contributions of anthropogenic versus wild bird factors in transborder spread of these viruses, much is still unknown. Debate on the role of wild birds intensified when viruses closely related to those found in wild birds at Lake Qinghai, China, in 2005 (belonging to subclade 2.2), spread across Russia to Europe and into Africa (see for example Feare, 2007).

Available information on wild birds and their potential to spread H5N1 viruses was reviewed in an European Food Safety Agency risk assessment, and this demonstrated the many gaps in the available data – both virological and ornithological (European Food Safety Agency, 2006). This risk assessment concluded that migratory birds pose a potential threat for the introduction of virus to Europe, but considerable uncertainty was associated with the conclusions.

Many of the original Russian outbreaks in domestic poultry in 2005 occurred in remote communities adjacent to wetlands having limited contact with the outside world, and where poultry and wild waterbirds mixed. Russian veterinary authorities have since concluded that wild birds were responsible for much of the spread within their country and have based their preventive programmes on this, providing vaccination coverage for non-biosecure flocks in the vicinity of sites where wild aquatic birds congregate (Irza, 2006). This does not mean that spread through trade did not occur once virus was established in a particular area.

Spread from Russia to other countries bordering the Black Sea is also considered to be due to wild bird movements. This followed extremely harsh winter conditions that led to dispersal of wild birds into western Europe in 2006 (Gilbert et al, 2006b).

Other cases in wild birds occurred in places with very few poultry or no evidence of infection in poultry. Wild bird cases in 2005 in Mongolia provided the best circumstantial evidence for spread of H5N1 HPAI by wild birds over long distances, given the very low domestic poultry population in that country. However, it is not possible to rule out other extremely remote possibilities such as use of live decoy ducks for hunting, which has been reported but is unlikely to have occurred (Williams, 2005; Feare, 2006). Infection was detected in Mongolia again in 2006, coinciding with the return of migratory birds in the northern spring (OIE, 2006).

In a number of countries, including China, cases of infection in poultry have occurred on relatively isolated farms. Subsequent investigations have revealed no apparent contacts between these farms and other potential sources of infection, apart from wild birds (F. Guo, personal communication with Les Sims). Wild birds have also been proposed as a possible source of infection for three out of four outbreaks in Japan in 2003-04 (Nishiguchi et al, 2005). This conclusion was based on the distance between outbreaks and lack of connection between the first three cases, the unprotected water supplies to affected farms, and the location of the farms in mountainous areas attractive to wild birds. Similar conclusions were made for infections in the Republic of Korea in 2003 based on circumstantial evidence (Wee et al, 2006). The return of H5N1 viruses to the country in 2006 coincided with movements of migratory birds and, although wild birds were not proved to be the source, they were considered to be the likely route of introduction (Lee et al, 2008).

The evidence for wild bird involvement in all the above cases, while compelling in some instances, remains circumstantial. It is also evident that wild birds have sometimes provided a convenient scapegoat for those wanting to deflect attention away from inherent problems in poultry production and marketing systems and their contribution to the spread of H5N1 HPAI.

The situation in Africa is less clear. Viruses isolated in Nigeria also belong to subclade 2.2. This particular subclade does not appear to be widely established in poultry or in places with high poultry density (see for example Smith et al, 2006a). If these introductions were due to illegal trade in poultry products, it is difficult to understand why such outbreaks did not occur in earlier years and involve genotypes more commonly found in poultry such as clade 1 or subclade 2.3, given that these strains of H5N1 HPAI viruses have been circulating widely in Asia for some time.

Recently it has been suggested that there were multiple incursions of H5N1 virus into Nigeria (although all strains in Africa belonged to subclade 2.2, there are minor differences between these viruses). This means it is not possible to rule out wild birds as a source of infection (Ducatez et al, 2006).

The spread of virus to other countries in West Africa may well have occurred through the movement of poultry and poultry products or fomite spread, but this, too, remains speculative. The sources of virus for Egypt, the Sudan and Djibouti remain unknown, but genetic studies have revealed that those in the Sudan and Egypt are not identical (Ducatez et al, 2006), possibly suggesting different origins. One case of infection in a wild duck was detected in Egypt before outbreaks were reported in poultry (Earhart et al, 2007).

2.3.1 Factors complicating analysis of modes of introduction

Analysis of routes of entry to uninfected places has been complicated by the limited capacity in some countries to investigate diseases, imprecise information on illegal movements of poultry or poultry products and delays in reporting of outbreaks when they first occurred. Even when the disease has been recognized early and full investigations undertaken, it has not always been possible to determine how the virus gained entry to the country and then to flocks of poultry.

Analysis has also been hampered because the first case of disease or infection recognized (the index case) in many countries was unlikely to be the first case of infection. HPAI can occur in smallholder or scavenging poultry without being diagnosed because mortality in village flocks from other causes occurs regularly (see for example FAO, 1998; Johnston, 1990). These deaths are not always reported and even if local authorities are advised, there is no guarantee that all cases will be investigated or reported further up the chain to central or provincial veterinary authorities. This under-reporting is most clearly evident in places where human cases have occurred in the absence of reported avian infections (‘human sentinels’) (Sims, 2007). In addition, targeted village surveillance in Indonesia and studies in Cambodia have also demonstrated significant under-reporting of mortalities in poultry (Elly, 2007; Desvaux et al, 2006).

Outbreaks have also gone unreported in large commercial chicken farms (Nishiguchi et al, 2005) although it is usually harder to hide such cases for extended periods because the number of infected birds increases rapidly. This leads to high mortalities, a large virus load, and a high probability of spread once infection is established in a large flock in an area with high concentrations of interlinked poultry farms. The possibility of failure to report disease needs to be taken into account when analysing outbreak data. Indeed, the sale of sick poultry is recognized as one of the so-called ‘coping strategies’ for farmers and villagers when their poultry or those of their neighbours are showing signs of disease.

Conclusions on the source of infection are often based on a process of elimination or on associations and probabilities rather than definitive evidence. Disease investigations are time consuming and require appropriately resourced and trained investigators, preferably focussing on investigation rather than control when the disease outbreak occurs. Unfortunately, investigations in countries with limited veterinary resources have often been relegated to a secondary role, or even neglected entirely, due to the competing demands of control programmes. This situation is improving in many countries as a result of increased funding of veterinary services, and it is expected that the quality of data on disease outbreaks and prevalence of infection will improve in the future (Sims, 2007). Even in developed countries, control and eradication of the disease have taken priority over investigations when outbreaks first occur (Balicer et al, 2007).

2.3.2 Molecular epidemiology – establishing linkages between viruses

Sequencing of the genes of H5N1 HPAI and other avian influenza viruses has assisted greatly in establishing links (and differences) between H5N1 virus isolates from different countries. In some cases, these data demonstrate independent introductions of virus to different countries. For example, molecular investigations suggest initial introductions of separate clades of H5N1 HPAI virus to Thailand (clade 1) and Indonesia (subclade 2.1) (i.e. viruses of different origin), followed by spread within each country (Chen et al, 2006c; Smith et al,2006b). Subsequently, additional incursions have been detected in Thailand (Smith et al, 2006b) but not in Indonesia, where only the initial strains have persisted and evolved. The precise origin or mode of entry of these viruses is not known. The closest known related viruses are those isolated in China in 2003 (WHO, 2007).

Strains of H5N1 HPAI virus isolated in Japan and the Republic of Korea in 2003-04 are very closely related, suggesting a common (unknown) origin (Mase et al,2005). The closest related virus was detected in poultry in southern China.  Preliminary data also suggest that viruses introduced to these countries in 2006-07 were closely related to each other, but not to the 2003-04 viruses (i.e. the 2007 viruses apparently belong to subclade 2.2 whereas those in 2003-04 belong to subclade 2.5 [WHO, 2007]). This indicates new incursions of virus rather than persistence of the earlier strain.

Genetic data also suggests that multiple incursions of H5N1 HPAI occurred in Viet Nam (Smith et al,2006a) and HKSAR, with the latter exposed to many different genotypes since 1997. This reflects the broad genetic diversity of H5N1 viruses in the region over this period, starting with the original goose viruses (Li et al, 2004a). Based on sequence data, it has been suggested that some strains of H5N1 HPAI virus in Viet Nam may have originated in China (Chen et al, 2006c). Movement in the opposite direction cannot be excluded, but does not match the predominant trade pattern for poultry, much of which is illegal and driven by price differences for poultry between China and Viet Nam (Sipress, 2006).

Molecular studies have also demonstrated that evolution of H5N1 viruses is continuing, as shown by the emergence of new sublineages of virus. For example, viruses belonging to a new subclade (subclade 2.2)  of H5N1 viruses were detected in wild birds in Qinghai province in China in 2005 (Chen et al, 2005; Chen et al, 2006a; Zhou et al, 2006) and these have now been found in Europe and Africa. Other new lineages that have been detected in the last few years include clade 7 viruses in Shanxi province in China in 2006, clade 4 viruses in Guiyang province and subclade 2.3 viruses that have been detected in southeast China and Southeast Asia since 2005 (WHO, 2007).

The subclade 2.2 viruses from Qinghai in 2005 appeared to be reassortants, with four different variants detected during the outbreak (Chen et al, 2006a).  Many of these viruses have lysine (K) instead of glutamate (E) at position 627 in the PB2 protein (the E627K mutation), generally regarded as a marker of passage through a mammalian host and, until 2005, rarely seen in avian isolates (Taubenberger et al, 2005). The presence of this mammalian signature in viruses in subclade 2.2 – first found in wild birds – is yet another demonstration of our limited knowledge and understanding of how these viruses evolve. The closest related viruses were detected in wild ducks at Poyang Lake in Jiangsu province in China in March 2005 (Chen et al, 2006c).

For infection and disease in wild birds with subclade 2.2 H5N1 HPAI viruses, it is not clear whether the outbreak in wild birds in May 2005 in Qinghai was the first such outbreak associated with this subclade, or merely the first one identified (due to its size). There are thousands of lakes in remote parts of the Qinghai plateau and across northeastern Asia where mortalities in small numbers of wild birds could have occurred unnoticed. Unfortunately, when reviewing the genesis of this H5N1 HPAI epizootic, absence of data has repeatedly been mistaken for absence of infection by the popular media, and even by some scientists. Unless the quality and range of surveillance programmes in apparently uninfected locations are considered, incorrect conclusions can be drawn about the infection status of countries or parts of countries (Sims, 2007).

It has been suggested that the failure to detect infected live wild birds in West Africa indicates that they probably played no role in the introduction of virus to Nigeria (Brown, 2006). Instead, the considerable movement of poultry and poultry products into Nigeria has been proposed as a reason to support the role of trade as the more likely source of introduction. As with many countries, it is still not known whether the first cases of infection detected in Nigeria were the first cases to occur and molecular data also suggests that wild birds cannot be ruled out.

1 The main indicator used for risk from trade in poultry was legal trade in live poultry from infected countries, much of which involves day-old chicks.
>>> Back to top