A cross-sectional survey of influenza A infection, and management practices in small rural backyard poultry flocks in two regions of New Zealand

Abstract

AIMS: To obtain baseline data on the management of small non-commercial backyard poultry flocks, in two rural regions of New Zealand, to investigate potential transmission pathways for avian influenza (AI), and to investigate the presence of AI in these flocks.

METHODS: During August–October 2006 a questionnaire was sent to 105 farms in the Bay of Plenty and Wairarapa with poultry flocks comprising fewer than 50 chickens, located near wetlands where AI virus had been detected previously in wild ducks. Information was collected on the number and species of poultry reared, opportunities for interaction between wild birds and poultry, farm biosecurity measures, and health status of poultry. Between September and November 2006, blood and tracheal/cloacal swabs were collected from poultry on a subset of 12 high-risk farms in each location. Influenza A-specific antibodies in sera were assayed using ELISA, and positive sera were further tested for the presence of H5 and H7 subtype-specific antibodies, using haemagglutination inhibition (HI) assay. The presence of influenza A virus in swabs was detected using realtime reverse transcriptase-PCR (RRT-PCR).

RESULTS: Returned questionnaires were received from 54 farms. Overall, 80% had only chickens, 13% chickens and ducks, and 7% had chickens and other galliform species. Nearly all (96%) kept backyard chickens for personal consumption of eggs, with a small proportion (19%) preparing birds for the table. On surveyed farms wild waterfowl were seen on pastures (70%) and/or farm waterways (46%). Waterfowl were recorded as visiting areas where domestic birds were kept on 31% of farms. Bird litter and manure were composted (94%) or buried (6%) on-farm, as were most (82%) dead birds. During the targeted cross-sectional survey of 24 farms clinical disease was not recorded in any poultry flock. Of 309 chicken sera tested, 11 (3.6%) from five farms across both regions tested positive for influenza A antibodies. In contrast, 16/54 (30%) duck sera from three farms in the Wairarapa were positive. Avian influenza H5 and H7 subtype-specific antibodies were excluded in ELISApositive sera using HI testing, and influenza A virus was not detected using RRT-PCR.

CONCLUSIONS: The study confirmed that small backyard poultry flocks located near waterfowl habitats were exposed to non-notifiable low-pathogenic AI viruses. Findings indicate a number of potential risk pathways for the transmission of AI viruses between wild birds and non-commercial poultry, and hence the need for continued surveillance for AI in backyard flocks and wild birds in New Zealand.

KEY WORDS:

• Influenza
• avian
• risk factors
• backyard poultry
• antibodies
• seroconversion
• ecology

AI	=	Avian influenza
EID50	=	Median egg infectious doses
HI	=	Haemagglutination inhibition
RRT-PCR	=	Real-time reverse transcriptase-PCR
S/P	=	Sample/positive

Introduction

Outbreaks from highly pathogenic avian influenza (AI) subtype H5N1 viruses in Southeast Asia have had a devastating impact on poultry industries and local communities. Due to the widespread and persistent nature of H5N1 in poultry and its more recent limited emergence in mammals and humans, it constitutes a serious public health concern (Sims et al. 2003). Transmission of AI viruses through trade of infected poultry and contaminated equipment is considered one of the main means of spread, but migratory birds have also been implicated as vectors for transmitting highly pathogenic H5N1 AI viruses (Gilbert et al. 2006; Kilpatrick et al. 2006; Keawcharoen et al. 2008). While Anseriformes, especially asymptomatic wild ducks, are considered the main reservoir for AI viruses (Kim et al. 2009), shorebirds might also play an important role in the ecology of AI (Krauss et al. 2004). New Zealand is part of the East Asian-Australasian Flyway for trans-hemispheric migratory shorebirds. The flyway passes through 22 countries, including some countries where outbreaks of highly pathogenic AI H5N1 infection have occurred. There is a theoretical risk of migratory shorebirds exchanging viruses with local waterfowl at stopovers, and carrying highly pathogenic AI strains onto Australia and New Zealand. The isolation of AI viruses from migratory shorebirds in Australia (Hurt et al. 2006) highlights the potential risk of introducing novel strains into New Zealand, but this remains unproven (Langstaff et al. 2009). New Zealand has never recorded a case of highly pathogenic AI in captive or wild birds, however non-pathogenic AI viruses have been detected in healthy wild mallard ducks (Anas platyrhynchos), including low-pathogenic notifiable AI virus H5 and H7 subtypes (Stanislawek et al. 2002; Langstaff et al. 2009).

In the study presented here, information was collected about poultry farming practices on small non-commercial backyard poultry flocks from two regions of New Zealand, where AI viruses have been detected previously in wild ducks. The aim was to identify potential pathways and risk-mitigation options for the introduction of AI to poultry, and investigate the presence of AI viruses, including the notifiable subtypes.

Materials and methods

Study farms

Study farms near major water bodies in the Bay of Plenty and Wairarapa regions of the North Island of New Zealand were identified from the AgriBase database (AgriBase 2006; AsureQuality Limited, Auckland, New Zealand; www.asurequality.com/ geospatial-services/agribase.cfm). AI viruses of several different subtypes had previously been isolated from wild mallard ducks in the two regions (Stanislawek et al. 2002). In the spring of 2006, there were 1,723 properties recorded in AgriBase located within a 15-km radius of either Maketu or Waihi Estuaries in the Bay of Plenty, and Lake Wairarapa in the Wairarapa, respectively (Figure 1). Amongst these two regions, 145 (8.4%) farms were registered as having more than five chickens; 78 in the Bay of Plenty and 67 in the Wairarapa, respectively. The owners were contacted by telephone or visited in person, to update farm information and to inform them of the study. In addition, public meetings were held in both local communities, and relevant local authorities and organisations were also informed. As a result of the public consultations, some owners volunteered to partake in the study.

Questionnaire

Following the community consultations, information sheets and a self-administered farming-practices questionnaire (see Supplementary Table 1 ¹) were mailed to farmers and property owners in the Bay of Plenty (n=52) and Wairarapa (n=53). Each of these farms had a small, non-commercial flock of between five and 50 chickens, and was within the area of interest. The questionnaire explored poultry farming practice, and potential risk factors relevant to the transmission of AI to poultry. It included questions on the number and species of domestic birds and other animals on the farm, farm biosecurity measures, animal husbandry, and animal health.

Cross-sectional survey for AI infection

A subset of the respondent farms, 12 in each location, was selected for further study following examination of the completed questionnaires. Selection of farms was based upon the farmers' availability to participate, and risks for exposure to AI viruses, including sighting waterfowl on the farm, and managing multiple poultry species.

Collection of samples

Collection of samples for the cross-sectional survey occurred between September and November 2006. A blood sample (∼1 ml) was collected from the wing vein of each bird in a 4-ml BD Vacutainer Plus plastic tube (with clot activator) (Beckon Dickson, Franklin Lakes NJ, USA). Serum was separated from the clot after centrifugation at 3,500 rpm for 10 minutes at room temperature, and stored at -20°C. Tracheal and cloacal swabs (PR McArthur Ltd, Auckland, NZ) were also collected from each bird, as described elsewhere (Anonymous 2005), and preserved in separate vials containing 1.5 ml of virus transport medium (Eagle's minimum essential medium; Invitrogen Corp, Carlsbad CA, USA), containing 0.5% bovine serum albumin, 0.3% gentamycin, 0.01% kanamycin, 0.01% streptomycin, and 0.06% penicillin. The vials were kept chilled with ice packs in the field, and stored at -80°C until processed. All procedures involving birds were approved by the Wallaceville Animal Ethics Committee, Upper Hutt, New Zealand.

Serological tests

ELISA

Antibodies to influenza A viruses in sera from chickens and turkeys (Meleagris gallopavo) were assayed using an AI virus antibody test kit (FlockChek; IDEXX Laboratories, Westbrook ME, USA), according to the manufacturer's instructions. Serum samples were tested in duplicate at a dilution of 1:500. The relative levels of antibodies in the samples were determined by calculating the ratio of sample-to-kit-positive control (S/P); samples with an S/P ratio >0.5 were considered positive.

Antibodies to influenza A viruses in sera from ducks, including domesticated mallards, Peking ducks (Anas peking), and Muscovy ducks (Cairina moschata); geese (Anser cygnoides), and guineafowl (Numida meleagris), were assayed using an AI blocking ELISA (Institute Pourquier, Montpellier, France), according to the manufacturer's instructions. The serum samples were tested in duplicate, at a dilution of 1:4. The relative levels of antibody in the test sera were determined by calculating the percentage of inhibition observed in the test sample compared with that of the negative control; samples with <35% inhibition were considered positive, using the formula: Inhibition %=(OD₄₅₀ of sample/mean OD₄₅₀ of negative control) x 100, where OD₄₅₀ is the optical density at 450 nm.

Haemagglutination inhibition (HI) test

ELISA-positive sera were tested using HI, using H5N2 and H7N3 antigens to identify H5 and H7 virus antibodies, respectively, following the procedure described elsewhere (Alexander 2009).

Virological testing

RNA purification

Vials containing frozen swabs were thawed at room temperature, and the swab, in virus transport medium, was removed after pulse-vortexing for 25 seconds. After centrifugation at 1,800g for 5 minutes at 4°C the supernatants from three tracheal or cloacal swabs were pooled in equal amounts. RNA from the pooled supernatants were purified using QIAamp viral RNA mini kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. Allantoic fluid harvested from chicken embryonated eggs inoculated with influenza virus A/duck/NZ/6/97 (H4N6) was used as a positive control. The titre of the H4N6 was 10^8.2 median egg infectious doses (EID₅₀)/ml. The control virus was serially diluted 10-fold in virus transport medium, and the RNA was purified in a manner similar to that for the swab samples.

Figure 1. Location of 54 farms (

) in (a) the Bay of Plenty, and (b) Wairarapa, including 24 involved in a cross-sectional study (

), for the presence of avian influenza virus, in relation to lakes and estuary. The radius circles indicate 5, 10 and 15 km away from the edge of a lake or a river mouth.

Real-time reverse transcriptase-PCR (RRT-PCR)

Influenza viral RNA was assayed in duplicate using the RRT-PCR specific for the matrix gene of influenza A virus, as described previously (Spackman et al. 2002), with minor modifications. Briefly, a SuperScript III platinum one-step quantitative RRT-PCR system (Invitrogen Corp) was used in a 25-µl reaction mixture, which included primers M +25 (5′-AGA TGA GTC TTC TAA CCG AGG TCG-3′) and M -124 (5′-TGC AAA AAC ATC TTC AAG TCT CTG-3′), and hydrolysis probe M +64 (5′-FAM-TCA GGC CCC CTC AAA GCC GA-BHQ1-3′), all at 0.8 µM, and 0.5 µl kit-supplied enzyme mixture and 5 µl RNA extract. All RRT-PCR were performed using either an Mx3000P machine (Stratagene, La Jolla CA, USA) and cycle threshold data were collected using MxPro-Mx3000P software v3.00, or a Rotor-Gene 6000 5-Plex HRM machine (Corbett Life Science, Mortlake NSW, Australia) and cycle threshold data were collected using Rotor-Gene 6000 software v1.7.65. The reverse transcription condition was incubation at 50°C for 15 minutes and 95°C for 2 minutes, followed by 40 cycles of a two-step cycle of 95°C for 15 seconds and 60°C for 30 seconds.

The sensitivity of detection of the RRT-PCR was evaluated using serial dilutions of the H4N6 isolate, and was shown to be 1–10 EID₅₀/ml of H4H6 at a cycle threshold of approximately 35. A positive (H4N6) and negative sample control, in addition to no template and no enzyme controls, were included for each run. Samples with a cycle threshold value <35 were recorded as positive, those with repeated cycle threshold values between 35 and 40 were regarded as weak positives, and samples with cycle threshold values >40 were deemed to be negative.

Statistical analysis

The association of the independent variables elucidated from the questionnaire, such as wild ducks recorded as visiting areas where poultry were kept, and poultry being allowed to free-range over the property (see Supplementary Table 1 ¹), and the dependent variable (AI virus seropositivity) was analysed using the χ² test (or Fisher's exact as appropriate) for the categorical variables, and student t-tests for the continuous variables. The continuous variables followed a normal distribution, and subgroups had equal variance (Barlett's tests of variance). Data were analysed using Minitab v15.1.0.0. (LEAD Technologies Inc, State College PA, USA). A p-value of p<0.05 was considered significant.

Results

Questionnaire

Of the 105 mailed questionnaires, 54 (51%) were returned completed. Most (43/54; 80%) farms had chickens as the only species of domestic poultry. The remaining farms had a mix of chickens and other domestic bird species (Table 1). Almost one-third (17/54; 31%) of respondents also had pigs (median n=4, range 2–20), kept in an enclosure. Chickens were usually home-bred (21/54; 39%) or obtained privately (20/54; 37%), with a smaller proportion bought from commercial suppliers (6/54; 11%). Other species of domestic birds, including ducks, were almost always obtained privately (12/25; 48%) or domesticated from the wild (8/25; 32%). Two (8%) further farms obtained domestic birds from a commercial supplier; the source of the birds was not indicated for the remainder (3/25; 12%). The primary reason for keeping chickens was for collecting eggs for personal consumption (52/54; 96%) and for sale (8/54; 15%), with a small number of farms also preparing poultry for the table (10/54; 19%). None of the farms sold live poultry or meat from poultry.

Two-thirds (36/53; 68%) of the farms allowed chickens free access to pastures, which, on four farms, all in the Wairarapa, included free access to waterways, i.e. ponds, drains, streams, rivers. On the seven farms that also kept ducks, the ducks were either free-ranging on the property (n=2), had free access to waterways (n=4), or were confined to a coup (n=1). Overall, 80% of the farms had waterways on or bordering the farm, that attracted waterfowl. Wild ducks were often seen on pastures (38/54; 70%), farm waterways (25/54; 46%), or in areas where poultry were kept (17/54; 31%), more often in the spring. The types of birds found in areas where domestic poultry were kept are detailed in Table 2.

Drinking water provided for poultry was mainly from bores (22/54; 40%) or rivers and streams (24%). Other water sources included town supply (14%), rainwater collected off the roof (9%), water from springs (9%), and ponds or dams on the property (4%). Of the 25 farms that collected water from a source liable to contamination by faeces from wild birds (springs, rivers, streams, dams, ponds, or off the roof ), 22 (88%) supplied the collected water unprocessed to poultry.

For waste management, on the majority of farms (51/54; 94%) bird litter was composted prior to being used as a fertiliser, while on the remainder the waste was buried on-farm. Dead birds were usually buried on-site (42/51; 82%), incinerated (6/51; 12%), disposed of with household refuge (2/51; 4%), or composted (1/51; 2%).

Ill health had been experienced in poultry on 11 farms in the previous 3 years. The majority (6/11) had just a single sick bird; four reported two, and one three, sick birds. Birds were generally found dead (7/11), or prior to death exhibited diarrhoea (3/11), lameness (2/11) or lethargy (1/11). In only two (18%) cases was advice sought from others; one from a veterinarian, and the other a friend. The majority of respondents reported that they would seek advice in the event of illness affecting their poultry, with 31/54 (57%) indicating that they would ask a veterinarian, and 4/54 (7%) the Ministry for Agriculture and Forestry. Around one-third (19/54; 36%) of respondents said that they would not seek any advice.

Cross-sectional survey for AI infection

The 24 farms participating in the cross-sectional study were located between 0.9 and 12.1 (median 8.4) km from major lakes or estuaries (Figure 1). Wild waterfowl had been sighted on pastures and waterways of 88% and 75% of these farms, respectively. The number of chickens on each farm ranged from four (median 16) to 35. In addition to chickens, 10 farms, three from the Wairarapa and seven from the Bay of Plenty, managed other species of birds, including ducks, geese, turkeys, guineafowl, pheasants (Phasianus colchicus), peacocks (Pavo cristatus), and pigeons (Columba livia domestica).

Detection of antibodies to influenza A virus in sera

Of 309 serum samples collected from chickens on the 24 farms in the cross-sectional study, 6/136 (4.4%) and 5/173 (2.9%) tested positive by ELISA for AI virus antibodies, respectively. HI tests did not detect H5 or H7 subtype-specific antibodies for notifiable AI viruses in any ELISA-positive sera from chickens and ducks. The seropositive chickens originated from two farms (2/6 and 4/12; number of seropositive samples/number of samples tested) in the Bay of Plenty and three farms (2/8, 1/15, 2/17) in the Wairarapa. Seropositive ducks were detected on three of the six farms with ducks, all from the Wairarapa (Table 3). The proportion of seropositive ducks was significantly greater than that for chickens on one farm (Farm SW06; p<0.05).

No significant associations were identified between farms with seropositive birds and any of the independent variables elucidated from the questionnaire. The proportion of farms where wild waterfowl were present in poultry areas did not differ between those with seropositive compared with seronegative birds (4/5; 80% vs 6/19; 32%; p=0.12), nor was there any difference between farms where poultry were allowed to range over farm pasture (5/5; 100% vs 10/19; 53%; p=0.12). The mean distance from farms to major lakes or estuaries did not differ between farms with seropositive compared with seronegative birds (5.2 (SD 1.6) km vs 6.9 (SD 4.7) km; p=0.43), and there was no difference in the percentage of farms with a waterway on or bordering the farm property (5/5; 100% vs 14/19; 74%; p=0.54). A similar proportion of farms with seropositive and seronegative birds had a risk, i.e untreated, water source for poultry (7/19; 37% vs 2/5; 40%; p=1.00), kept domesticated ducks or geese (2/5; 40% vs 4/19; 21%; p=0.57), and kept pigs (2/5; 40% vs 11/19; 58%; p=0.63).

Table 1. Different species and numbers of domestic birds reported on 54 farms in the Bay of Plenty (BoP) and Wairarapa (W'pa), New Zealand, by respondents to a survey of owners of backyard poultry flocks.

Birds kept	BoP	W'pa	All farms^a	% of Total	Bird populations^b
					Median	Range
Chickens	29	25	54	100	12	3–45
Ducks (Anas platyrhynchos), (Cairina moschata) or (Anas peking)	3	4	7	13	7	1–45
Guineafowl (Numida meleagris)	3	2	5(2)	9	4	4–18
Turkeys (Meleagris gallopavo)	3	1	4(2)	7	2	1–30
Pheasants (Phasianus colchicus)	2	1	3(1)	6	8	2–9
Geese (Anser cygnoides)	2	0	2(1)	4	1.5	1–2
Partridges (Alectoris rufa)	1	0	1	2	4	4
Pigeons (Columba livia domestica)	1	0	1	2	4	4
Peacocks (Pavo cristatus)	0	1	1(1)	2	3	3
Doves (Streptopelia chinensis)	1	0	1	2	7	7
Other	2	1	3(1)	6	10	7–10
^a The number in brackets denotes the number of farms that kept ducks as well as the other domestic bird specified	^b Number of birds per farm

Table 2. Housing of backyard poultry on farms in the Bay of Plenty and Wairarapa, New Zealand, and the species of wild birds reported in areas where domestic poultry were kept.

Parameter	Bay of Plenty	Wairarapa
Farms surveyed	29	25
Housing of chickens^a
Shed or coup	2	1
In a run	8	6
Free range over property	18	14
Free range, having access to waterways	0	4
Wild bird species reported
Duck (Anas platyrhynchos), teal (Anas chlorotis), swan (Cygnus atratus), shag (Phalocrocorax carbo)	9	8
Seagull (Larus novaehollandiae scopulinus), tern (Sterna albostriata), godwit (Limosa lapponica)	4	1
Pheasant (Phasianus colchicus), turkey (Meleagris gallopavo), quail (Coturnix novaezealandiae)	11	2
Sparrow (Passer domesticus), starling (Sturnus vulgaris), thrush (Turdus merula)	23	21
Pukeko (Porphyrio porphyrio)	13	5
Hawk (Circus approximans)	9	3
Other birds^b	3	6
^a This question was not completed for one farm in the Bay of Plenty	^b These included, but were not limited to: magpie (Gymnorhina tibicen), tui (Prosthermadera novaeseelandiae), yellowhammer (Emberiza citronella), cockatoo (Cacatua galerita), rosella (Platycercus eximius), fantail (Rhipidura fulginosa), finch (Fringilla coelebs), and kereru (native wood pigeon) (Hemiphaga novaeseelandiae)

Table 3. Number of samples that tested positive for avian influenza virus antibodies of all samples collected and tested from 6/24 farms, with multiple species of birds, including chickens, ducks, geese (Anser cygnoides), turkeys (Meleagris gallopavo) and guineafowl (Numida meleagris), in the Bay of Plenty (BP) and Wairarapa (SW), New Zealand, September–November 2006.

Farm	Chickens	Ducks	Geese	Turkeys	Guineafowl
BP01	0/11	0/5^a	0	0	0
BP08	2/6	0/2^b	0	0	0
BP11	0/12	0	0/2	0/2	0
SW06	1/15	14/36^c	0	0	0/3
SW08	0/22	1/2^b	0	0	0/1
SW11	0/9	1/9^d	0	0	0
^a Muscovy ducks (Cairina moschata)	^b Domesticated mallards (Anas platyrhynchos)	^c Thirty-five domesticated mallards and one Peking duck (Anas peking)	^d Five domesticated mallards, two Peking ducks, and two Muscovy ducks

Discussion

To our knowledge this is the first report outlining management practices and potential exposure pathways to AI virus for noncommercial backyard poultry in New Zealand. The study farms chosen were considered at higher risk for exposure to AI viruses as they were in close proximity to wetlands where AI viruses had previously been identified in wild waterfowl, and opportunities for interaction between domestic and wild birds on these farms had been identified prior to sampling. The study confirmed that there is a potential for small backyard flocks located near waterfowl habitats in New Zealand to be exposed to AI viruses.

In this study, direct exposure to wild birds was identified as a common potential pathway for the introduction of AI viruses to domestic backyard poultry. On more than two-thirds of the farms surveyed wild waterfowl had been seen on pasture, and chickens from two-thirds of the farms had free access to pasture. Furthermore, on 31% of farms waterfowl were reported in areas where domestic poultry were kept. Numerically, a higher proportion of farms with seropositive birds reported the presence of wild ducks in poultry enclosures, and allowed poultry free-range access to outside areas, although these differences were not statistically significant. The potential risk associated with contact with wild birds is consistent with findings in other studies (Alexander and Capua 2008; Ward et al. 2009). A further interesting finding in the current survey was that small birds, including sparrows (Passer domesticus) and European starlings (Sturnus vulgaris), were the wild birds most frequently reported (44/54; 81%) visiting poultry roosting areas. Both sparrows and starlings are susceptible to experimental highly pathogenic H5N1 infection, and excrete high titres of virus (Boon et al. 2007). Therefore, excluding wild birds from entering poultry keeping areas would reduce the exposure of poultry to AI viruses in wild birds, especially during the spring and summer when the prevalence of AI virus infection in wild ducks is high (Stanislawek et al. 2002). Keeping free-range birds in bird-proof runs has been employed in some European countries to mitigate the threat of H5N1, and could potentially be implemented in New Zealand under similar conditions.

In the study presented here, drinking water for poultry was identified as a common potential indirect pathway for the transmission of AI viruses between wild birds and domestic poultry. Twenty-five farms used water sources liable to contamination by faeces from wild birds, and most (22/25) applied no risk-mitigating measures such as chlorination, filtration or ultraviolet-light treatment. AI viruses can persist for extended periods in water (Webster et al. 1978; Brown et al. 2007), and water was implicated as the likely source of AI virus in an outbreak of highly pathogenic AI affecting commercial poultry in Australia (Morgan and Kelly 1990). Chlorination or ultraviolet-light treatments of water sources for poultry are practical farm-level biosecurity measures that would reduce the risk of introduction of AI viruses.

Live poultry bought from other farmers is a further potential route for the introduction of virus, and took place on 36% of the farms surveyed. Studies in Southeast Asia identified buying poultry from live-bird markets as a risk factor for outbreaks of H5N1 (Cristalli and Capua 2007). In the survey reported here, poultry waste was composted prior to use as fertiliser on pastures or gardens, and dead birds were disposed of by burial on most. In Southeast Asia, the use of untreated poultry manure as a fertiliser, and methods of disposal of sick and dead birds were associated with outbreaks of H5N1. In New Zealand, the use of good farming practice that would help minimise the dissemination of pathogens has been confirmed by the findings of this study.

Findings from the study presented here also confirmed the importance of targeted communication to enhance awareness, and clearly identify lines and mechanisms of reporting. Thirty-six percent of respondents said they would not inform or seek advice from anyone regarding a mortality event affecting their poultry. The low reporting rate appears to be verified by the finding that only 2/11 farmers experiencing mortalities in the last 3 years notified or sought advice from anyone. In a study in Cambodia, despite half of the respondents believing it was important to report deaths of poultry to the authorities this occurred in only 18/162 (11%) cases. In Cambodia, the main reason cited for non-reporting was that the survey respondents did not know how this was done (Ly et al. 2007). Timely reporting by farmers and the public is an essential element of surveillance systems for exotic diseases in New Zealand and worldwide.

The seroprevalence of influenza A determined in the study presented here in backyard flocks was generally very low. No influenza A virus was identified in any swabs collected, and there was a moderate seroprevalence in only a single duck flock. Findings suggest that infection with AI virus in high-risk backyard flocks, during the spring months in New Zealand, is low. Aquatic avian species, primarily waterfowl, are believed to be the main reservoir hosts for AI viruses, and the exposure to wild birds is inherent in backyard management systems, so those which were seropositive were considered to have occurred as a result of local exposure to waterfowl at nearby wetlands (Stanislawek et al. 2002; Alexander and Capua 2008). The finding of a low seroprevalence of AI in backyard chickens is similar to studies from other middle- and high-income countries in temperate regions. In Argentina, during 1998 and 2005, both serological and virological methods failed to detect evidence of AI virus infection in backyard chickens (Buscaglia et al. 2007), while in Italy during 2004 to 2006, AI viruses were isolated from 12% of 164 backyard ducks but not from chickens (Terregino et al. 2007). In contrast, seroprevalence ranged between 70% and 100% in commercial poultry, using the FlockChek ELISA, from low-income countries such as Jordan and Pakistan (Naeem et al. 2003; Al-Natour and Abo-Shehada 2005; Terregino et al. 2007). In one of the farms surveyed here (SW06), where both chickens and ducks were farmed together, the seroprevalence was significantly lower in chickens than in ducks. This may reflect differences in susceptibility to AI virus infection, as ducks are the natural hosts of influenza viruses (Alexander et al. 1978), or could result from the use of different serological tests for the two species. The single serum dilution in the two assays was 1:100 and 1:4 for the FlockChek ELISA and the AI blocking ELISA, respectively. Although both FlockChek ELISA and AI blocking ELISA were able to detect antibodies to AI virus in sera from chickens, the FlockChek ELISA had been validated using a wider variety of AI virus subtypes compared with the blocking ELISA, as indicated in the manuals of the kits. The FlockChek ELISA kit was accredited by the United States Department of Agriculture for the detection of antibody to AI virus in serum from chickens, and has been used by other researchers conducting serological surveys for AI virus (Naeem et al. 2003; Al-Natour and Abo-Shehada 2005; Terregino et al. 2007). For these reasons, the FlockChek ELISA was chosen for as the most suitable test for the detection of antibodies to influenza A virus in the sera collected from chickens in this study.

The sampling regime in the current study targeted theoretical high-risk farms from regions close to wetlands. It is unknown if the rate of infection of AI virus in chickens remains low throughout the year, and how it varies across New Zealand. A longitudinal study of AI virus infection in backyard flocks will help to determine seasonal trends of infection, and provide useful information for formulating an effective AI virus surveillance strategy for New Zealand. Although there have never been any outbreaks of AI in poultry in New Zealand, serological evidence of AI virus identified here in backyard poultry flocks illustrates the need for continuing active surveillance in domestic and wild birds in New Zealand.

Acknowledgements

The authors thank G Flanagan, F Khan, H Barton and F Cox of AgResearch, for excellent technical assistance; Dr RL Sanson of AsureQuality Ltd, for assistance with AgriBase; and Dr J Williman of the Institute of Environmental Science and Research Ltd, for setting up a database of the study. The generous support from all participating farmers, and the financial support from MAF Biosecurity New Zealand, Environmental Science and Research Ltd, and AgResearch Ltd, are greatly appreciated.