Outbreak Investigation Report on Avian Influenza in British Columbia, 2014
5. Summary of Findings and Working Hypotheses on Source and Transmission of NAI

5.1. Source of the Virus

Waterfowl are well established primary reservoirs for a wide variety of strains of avian influenza. A recent report from the CFIA entitled "Qualitative Risk Assessment of a Highly Pathogenic Avian Influenza Virus in Canada" examined multiple potential sources of the H5N2 virus. The epidemiological data currently available suggest that migratory wild birds within the Pacific Flyway may be the source of introduction of the new HPAI H5N2 virus into the Fraser Valley. This hypothesis is supported by the mid-December 2014 detection of H5N2 in a non-commercial operation (IPNC-01) where birds were known to have shared a pond with wild waterfowl. Also, H5N2 was detected in Whatcom County in Washington State, (close to the border with BC) in a Northern pintail duck in December, 2014. The H5N2 virus, similar to the one involved in the outbreak in BC, was detected in non-commercial operations in Washington State in 2014 and 2015, and Oregon in 2015. The virus has also been reported in wild migratory birds and captive wild birds in several other Western states in the USA in 2014 and 2015.

In January 2015, an HPAI H5N1 of Eurasian/North American origin was identified again in Washington State. The same virus was then detected on February 7th, 2015 in a non-commercial operation in BC. In the Fraser Valley during the fall, tens of thousands of migrating waterfowl use the farm lands of the region as a resting stop before continuing their journey southward. This fall migration period usually coincides with high rainfall and this was the situation in 2014. In addition, December recorded higher-than-normal average temperatures and rainfall, which is associated with a higher-than-normal number of wild migratory birds still present in the region. On many of the infected farms, standing water and large numbers of waterfowl were observed in the fields.

5.2. Secondary Spread

Waterfowl contaminate the environment by shedding the virus in their feces. The likelihood of exposure to indoor-reared poultry by migrating waterfowl is subject to several factors: the density of the poultry population in the affected area, the strength of biosecurity measures, the level and frequency of contact between farm workers and birds, and the proximity of the poultry to the migration routes of wild birds. While poultry production systems in Canada are designed to reduce or prevent contact between wild birds and commercial poultry, the potential does exist for the introduction of the virus from the environment if a breach in biosecurity occurs.

Once a poultry barn has become infected by HPAI, the viral load produced by thousands of infected birds makes the spread within and between farms much more difficult to control. Yet, in this outbreak, very little localized spread of the virus was observed. Most new infected farms were many kilometers away from each other, even in high density areas. The introduction of mandatory biosecurity programs in the poultry industry seems to account for some of this. As well, the CFIA detects, depopulates and disposes of infected poultry in barns in less than half the time required in 2004, and surveillance practices limited on-farm traffic. These factors likely contributed to reducing local disease transmission around IPs in the Fraser Valley during this outbreak.

5.3. Field Epidemiology – Summary of Findings

Seven of the 11 commercial farms infected over the first three weeks in December were broiler breeder operations. It was determined that two farms likely became infected via the transfer of breeding stock (spiker males) from IP1 to IP3 and IP4 during the infectious period of IP1. Service provider information indicates that infection of IP6 by IP5 was probably due to sharing a catching crew that did not apply biosecurity measures before going to IP6. All other contacts examined did not help identify how the virus spread; there were no links among IPs that suggested spread via direct or indirect contact.

A review of the preliminary field epidemiological investigation identified three hypotheses related to the means of introduction and spread of HPAI virus during this outbreak:

1. There were at least four independent primary introductions of HPAI virus into commercial farms during this outbreak (IP1, IP2, IP5 and IP9). This hypothesis was supported by the distance and absence of any identified epidemiological links between the farms (they were located at least 10 km apart).
2. Each of these point source infections led to limited local spread, with possible breaches in biosecurity, particularly the presence of rodents and contact with infected premises or wild birds as major risk factors.
3. Secondary spread of virus between farms may have occurred due to the high density of farms in this particular geographic area as well as the proximity to waterfowl. Local contact events with people, vehicles, machinery, vermin, dust particles could be responsible for this.

Figure 4 - Hypothesized spread of HPAI among commercial farms in the 2014 outbreak in BC resulting from the analysis of field epidemiological data. IPs in bold represent potential point source introductions.

Description for photo - Hypothesized spread of HPAI among commercial farms in the 2014 outbreak in BC resulting from the analysis of field epidemiological data. IPs in bold represent potential point source introductions

This diagram shows the hypothesized links among the different infected farms (represented by IP) during the outbreak that result from the epidemiological investigation.

The diagram shows IP1 in bold, a broiler breeder producer, with two solid arrows to IP3 and IP4 representing confirmed links to these two broiler breeder producers. The bolded character for IP1 means that it is considered to have been infected as a point source introduction. IP3 and IP4 are considered to have been infected as a result of the movement of live birds.

To the right, IP2 is a meat turkey producer, is also shown in bold as it is also considered a point source introduction. Two dashed arrows point from IP2 to IP7 and IP8 who are broiler breeder producers. Dashed arrows represent suspected links and the image refers to local spread or environmental contamination as a means of spread from IP2.

Below IP1 we find IP5, a meat turkey producer, in bold, representing a point source introduction. A solid arrow points from IP5 to IP6 (also a meat turkey producer) representing spread through a shared catching crew and close proximity (they shared a driveway).

Dashed arrows leave IP5 and IP6 and point to IP10 representing the suspected link through airborne spread or local spread. IP10 is a table egg producer. Another dashed arrow leaves IP10 to IP11, a broiler breeder producer representing local or environmental spread.

Finally, at the bottom right of the diagram we find IP9 in bold representing a point source introduction. It has no links to other infected farms.

Field epidemiology investigations after December 16th did not identify any links between the IPs. Farm workers on commercial IPs did not own birds or did not work on other commercial poultry farms. Indirect contacts account for no links between the two sectors and there were no movements of birds from the commercial IPs to the non-commercial IPs.

5.4. Airborne Spread Investigation – Summary of Findings

Airborne spread was suspected as an important mode of transmission in the 2004 HPAI outbreak in BC (Power, 2005). An epidemiological investigation into this pattern was conducted with risk factors of wind speed and direction, level of precipitation and location of farms in relation to an IP (at levels of 1.5 and 3 km) being examined. Following this investigation, there was no significant evidence of this being a source of spread in this outbreak.

The likelihood of airborne spread during specific depopulation events was also assessed using the same methodology as described above but for the specific time period when depopulation events took place. It was estimated that the likelihood of specific depopulation events being responsible for contamination was very low to low.

5.5. Viral Sequencing and Phylogenetic Analysis – Summary of Findings

Full genome sequencing was performed on the original specimens and on the viruses isolated from those specimens for each of the infected farm. The sequencing indicated that no significant changes had occurred as a result of passaging in chicken embryos.

The viral sequencing demonstrated a very high degree of homology between the IP's. Additionally, there was evidence that this strain was circulating as a well-adapted highly pathogenic strain in the wild bird population rather than a low pathogenic strain which mutated into a highly pathogenic strain.

5.6. Wild Bird Surveillance – Summary of Findings

In partnership with the Canadian Wildlife Service (CWS) and the BCMAGRI-AHC, the BC Wild Bird Mortality Investigation Plan provided an established infrastructure for the public reporting, triage and collection of dead wild birds. The program was immediately enhanced to define target species (waterfowl, raptors, herons) and to provide a field collection and carcass retrieval service provided by a private contractor. In addition, Fraser Valley poultry producers were encouraged to submit dead wild birds found in close proximity to their barns. Between December 1st and February 18th, 2015, a total of 125 wild birds in the Fraser Valley were tested for Avian Influenza by RT-PCR and matrix positive swabs were further tested for H5 and H7. All positive detections were forwarded to NCFAD for further characterization. Of these samples, one wild bird (duck) was found to be positive for H5N8.