Swine influenza test results from animal health laboratories in Canada

Abstract

Due to its infrastructure and partnerships the Canadian Animal Health Surveillance Network was able to rapidly collect test results from 9 Canadian laboratories that were conducting primary testing for influenza on swine-origin samples, in response to the threat posed by the pandemic H1N1 influenza virus in 2009.

The Canadian Animal Health Surveillance Network (CAHSN), a partnership between provincial, academic, and federal animal health laboratories, was formed in 2005 to improve animal disease detection and response in Canada (1). One of the main objectives was to develop a national surveillance system for animal disease threats. The CAHSN has successfully established the infrastructure to collect, collate, and analyze selected diagnostic test results from laboratories within the network for diseases such as bovine spongiform encephalopathy (BSE) and notifiable avian influenza. This article is written to highlight how the existing infrastructure and partnerships of CAHSN were used to rapidly begin the collection of swine influenza test results, and to describe in basic terms the data that were collected.

In spring of 2009, a new strain of influenza virus was identified in the human population. This strain, designated as pandemic H1N1 (pH1N1), was found to affect swine, although the severity of the disease in swine and the potential for inter-species transmission was unknown (2). The detection of the pH1N1 strain on a swine farm in Alberta, Canada was reported to the World Organization for Animal Health (OIE) on May 2, 2009 (2). In response to this event, there was an immediate need to collect and collate all influenza test results originating from Canadian swine from provincial, private, and academic laboratories in Canada that were conducting primary testing. A Web-based database was rapidly developed using the Canadian Network for Public Health Intelligence (CNPHI) platform (3) utilizing its Web Data technology (4), which was developed to serve such a purpose. This solution ensured that the data were stored centrally using Web-based technology to enable seamless secure access as well as providing the ability to interactively interrogate and extract data. The goal was to produce a database of swine influenza test results in Canada, including those for the new pH1N1 strain, which could be used to inform decision-making if the disease in swine proved to be significant to public or animal health. In conjunction with this data collection, the Canadian Food Inspection Agency’s (CFIA’s) National Centres for Animal Disease (NCAD) (laboratories based in Winnipeg, Manitoba and Lethbridge, Alberta), which are also partners in CAHSN, provided the other CAHSN laboratories with testing methodologies for the new strain and provided further differentiation of influenza isolates (5–8).

The CAHSN swine influenza database was designed and completed by May 2, 2009, the same day the Alberta results were reported to the OIE, and the first data entry began on May 7, 2009. Initially all results were requested from January 1, 2009 and as of December 31, 2010; 2 full years of results were collected. During the time of the data collection, 9 laboratories in Canada conducted primary influenza testing on swine-origin samples: British Columbia Ministry of Agriculture and Lands, Alberta Agriculture and Rural Development, Prairie Diagnostic Services (SK), Manitoba Agriculture, Food and Rural Initiatives, Animal Health Laboratory (ON), Ministère de l’Agriculture, des Pêcheries et de l’Alimentation du Québec, Biovet (QC), Atlantic Veterinary College (PE), and Université de Montréal (QC). The laboratories voluntarily supplied their test results to the database.

The data requested were based on the minimum data set for laboratory results as developed by the Surveillance/Epidemiology Advisory Committee to the Council of Chief Veterinary Officers of Canada (9). Six data fields were available and were completed for all submissions: testing laboratory, submission number, the date the submission was received by the laboratory, species, test performed, and results. Other data fields requested included reason for test, postal code of farm or postal code of the submitter if the farm postal code was unavailable, province of origin of the sample, and sample type; however, these fields were not available for every submission. In particular 2 laboratories did not have the postal code or province of origin information associated with the sample results in their system, 1 laboratory did not record sample type and the reason for test information was inconsistent. No specific identifying information on the farm of origin was requested as this may be considered to be personal information. The data fields requested enabled temporal and basic spatial analysis by utilizing the laboratory’s submission number, which is a unique identifier assigned by the laboratory to each submission. A submission consists of 1 or more samples from 1 or more animals residing on 1 farm premises, which are collected on 1 date, with each sample having 1 or more tests performed on it. Each submission number was assumed to identify a unique occurrence or case for analysis.

For a 2-year period, from the beginning of January 2009 to the end of December 2010, there was a total of 2993 submissions collected from swine for influenza testing. These submissions consisted of 15 937 samples on which a total of 29 500 tests or analyses were performed to detect or differentiate influenza virus. Unique submissions by month are shown in Figure 1. It is interesting to note that the greatest number of submissions for any month in the 2-year period occurred in May 2009, immediately after the new influenza strain had been detected in humans and subsequently found in swine in Canada (2,5). This increase was most pronounced in 1 laboratory; however, the trend was similar if less pronounced in all laboratories. This suggests that passive surveillance based upon clinical submissions did not decrease in the face of this threat, but that the situation may have caused an increase in submissions relating to influenza. The data were also examined for temporal patterns in the proportion of the tests that were positive for influenza in general, and specifically for the proportions positive to H1 and H3; no patterns were observed. It proved to be challenging to directly compare the results from different laboratories since accepted case definitions do not exist and the ability to do detailed analysis on the data is constrained by the complexity of the data and the different reasons for testing in the various laboratories. There are many reasons that swine are tested for influenza: to detect disease, to monitor ongoing infections, to prove absence or presence of disease in a production unit or to prove absence of disease prior to sale of animals. The diagnostics conducted by the laboratories are tailored to their own clientele and the participating laboratories used 35 unique test methodologies over the 2-year period to detect or characterize influenza in swine in Canada. These tests ranged from virus isolation to basic serology.

Figure 1.

Number of submissions by month, originating from swine that were tested for influenza in Canada in 2009 and 2010.

This data collection and the pH1N1 event in general provided valuable insights for the CAHSN partnership in support of disease surveillance, the collection of data, and laboratory diagnostics. The submission of these data was voluntary as influenza in swine is not a federally regulated disease in Canada. The data provided were limited to information that was collected and recorded by the laboratories, and to what the laboratories could provide, given the limitations imposed by client confidentiality and privacy legislation. Even with these considerations, the collection of data from all laboratories was possible and a limited but complete data set consisting of the 6 data elements for each submission was provided by all laboratories. The elements provided proved to be sufficient to allow basic analysis. If this emerging disease event had proven to be significantly more serious it would have been relatively simple to expand the data collection to include further identifying information on premises to allow more detailed analysis to be conducted. The data collection demonstrates that passive surveillance based upon existing submissions to the laboratories and voluntary data submission can quickly provide information in emergency situations such as emerging diseases, provided that the basic infrastructure is present before the emergency occurs. Ultimately, the most important point demonstrated by the data collection is the value of this infrastructure, of the CAHSN partnership and of the existing relationships, and the ongoing communications within the network. This infrastructure enabled data collection as well as the utilization of the NCAD for diagnostic assistance and differentiation of influenza positive submissions. In a large country such as Canada, with diverse livestock industries and multiple laboratories, structures such as CAHSN are necessary to facilitate cooperation on a national basis for disease surveillance and response.