Wild Bird Influenza Survey, Canada, 2005

Of 4,268 wild ducks sampled in Canada in 2005, real-time reverse transcriptase–PCR detected influenza A matrix protein (M1) gene sequence in 37% and H5 gene sequence in 5%. Mallards accounted for 61% of samples, 73% of M1-positive ducks, and 90% of H5-positive ducks. Ducks hatched in 2005 accounted for 80% of the sample.

T o provide baseline information about the strains and distribution of infl uenza viruses in Canadian wild ducks and to respond to the emergence of highly pathogenic avian infl uenza (HPAI) type H5N1 in Asia, Europe, and Africa, Canada's Interagency Wild Bird Infl uenza Survey was ini-tiated in July 2005. The goals of the survey were to identify avian infl uenza viruses in wild ducks in Canada and to detect HPAI strains early (1). We report the results of realtime reverse transcriptase-PCR (RRT-PCR) analysis.

The Study
Single cloacal swabs were collected from apparently healthy ducks at 56 sites within 6 geographic regions: British Columbia, Alberta, Manitoba, Ontario, Québec, and the Atlantic provinces (New Brunswick, Nova Scotia, Prince Edward Island, and Newfoundland and Labrador) (Figure). The target number of samples for each region was 800: 500 from mallard ducks (Anas platyrhynchos) and 300 from other duck species. Ducks were trapped and handled as part of annual duck banding carried out by the Canadian Wildlife Service and its associates. They were either caught in baited traps or netted from air boats. In all regions, duck banders were asked to preferentially sample birds hatched in 2005 to maximize virus detection within the sample (2)(3)(4).
Swabs were immediately placed in virus transport medium, refrigerated for up to 3 days, and then frozen at <-20°C until tested. The transport medium comprised Hanks balanced salt solution supplemented with 10% glycerol, 200 U/mL penicillin, 200 μg/mL streptomycin, 100 U/mL polymyxin B sulfate, 250 μg/mL gentamicin, and 50 U/mL nystatin.
All samples were tested at regional veterinary diagnostic laboratories within Canada's Infl uenza Virus Laboratory Network by RRT-PCR, which targets a conserved region of the M1 gene within infl uenza A segment 7. If the M1 gene sequence was detected, RRT-PCR for H5 and H7 hemagglutinin gene segments was performed. All laboratories followed uniform procedures (5) with positive and negative controls, and quality assurance was provided by the National Centre for Foreign Animal Disease (Canadian Food Inspection Agency). All fi eld and laboratory data were entered directly into a national database developed and maintained by the Canadian Cooperative Wildlife Health Centre. Duplicate records and any sample records missing RRT-PCR results or fi eld collection data (species, age, sex, location) were removed from this analysis. Data were exported to Microsoft Excel 2003 (Microsoft Corp., Redmond, WA, USA) and prepared for analysis. Statistical analysis was performed with Epi-Info version 3.3.2 (Centers for Disease Control and Prevention, Atlanta, GA, USA) and SAS version 9.1 (SAS Institute Inc., Cary, NC, USA). Maps were generated in ArcGIS 9.0-ArcMap version 9.1 (ESRI, Redlands, CA, USA).
To assess differences in M1 and H5 virus detection by age class and species, main effects logistic regression models were constructed by using a manual stepwise procedure. Age class, species, sampling area, and sex were included in the model. Age class was categorized into 2005 hatch-year birds and birds from another hatch year. Species was categorized into mallard (mallard and mallard-American black duck [A. rubripes] hybrids), other dabbling duck species, diving ducks, and other tribes ( Table 1). All variables were assessed for potential confounding; variables that changed estimates by >20% were left in the main effects models as confounders (6).

Conclusions
A total of 4,268 valid sample records were available for this analysis; 37% (1,572) of ducks were M1-positive and 5% (208) were H5-positive by RRT-PCR. No samples tested positive for the H7 gene sequence. Sampled areas varied considerably in the proportion of M1-positive samples, ranging from 63% (348/556) in southern Quebec to 9% (22/254) in southern Alberta ( Age, area sampled, and species were included in the final logistic regression model for M1 RRT-PCR test results. Age, sex, and species were included in the fi nal model for H5 RRT-PCR results. Mallards were more likely to be M1positive than other dabbling ducks or diving ducks but not more likely to be M1-positive than other duck tribes (Cairinini and Oxyurini). Mallards were more likely to be H5positive than other dabbling ducks, diving ducks, or other duck tribes. Hatch-year ducks were 1.7 times more likely to be M1-positive and 13 times more likely to be H5-positive than older ducks. Male ducks were 1.4 times more likely to test positive by H5 RRT-PCR than female ducks.
These data provided a snapshot of the frequency and distribution of infl uenza A viruses in wild ducks across southern Canada in 2005. This unique nationwide snapshot showed wide variation in detection of M1 and H5 gene sequences among 6 regions that were broadly representative of the northern terminus of duck migration corridors in North America (7). Previous studies have focused on smaller subregions of Canada sampled once or across multiple years (4,(8)(9)(10). Direct comparisons with previous studies should be made cautiously. Considerable variation exists across studies in the age and species of birds sampled and seasonality of sampling. All of these variables may affect reported prevalence of infection with infl uenza A viruses. Also, most previously published results relied on the cultivation of viable virus as the detection method rather than on RRT-PCR applied directly to cloacal swabs. Because samples for this survey were secured from preestablished duck-banding operations, and young healthy ducks, particularly mallards, were targeted to maximize virus recovery (1,11,12), these data may not represent true infection prevalence in the sampled duck populations. Despite these biases, avian infl uenza viruses were common in wild ducks across Canada in the summer and fall of 2005. Infection rates detected by RRT-PCR among the 6 regions of Canada were similar in scale and degree of variation to rates reported from the Alberta region, which were measured over 8 consecutive years from 1976 through 1983 (4).