Eb

MLa

b

c

a

ARRA

KAAEN

VCm

a

0

Preventive Veterinary Medicine 118 (2015) 169–181

Contents lists available at ScienceDirect

Preventive Veterinary Medicine

j ourna l h om epa ge: www.elsev ier .com/ locate /prevetmed

pidemiology of avian influenza in wild aquatic birds in aiosecurity hotspot, North Queensland, Australia

d. Ahasanul Hoquea,b,c,∗, Graham William Burgessa, Ai Lee Cheama,ee Francis Skerrattb

School of Veterinary and Biomedical Science, James Cook University (JCU), Townsville QLD 4811, AustraliaSchool of Public Health, Tropical Medicine and Rehabilitation Sciences, JCU, Townsville QLD 4811, AustraliaFaculty of Veterinary Medicine, Chittagong Veterinary and Animal Sciences University, Khulshi, Chittagong 4225, Bangladesh

r t i c l e i n f o

rticle history:eceived 1 February 2014eceived in revised form 6 November 2014ccepted 11 November 2014

eywords:ustralian nomadic wild aquatic birdsvian influenzapidemiologyorth Queensland

a b s t r a c t

Migratory birds may introduce highly pathogenic H5N1 avian influenza from SoutheastAsia into Australia via North Queensland, a key stopover along the East Asian-AustralasianFlyway, with severe consequences for trade and human health. A 3-year repeated cross sec-tional study on the epidemiology of avian influenza in Australian nomadic wild aquatic birdswas conducted in this potential biosecurity hotspot using molecular and serological tech-niques. Avian influenza virus subtypes H6 and H9 were commonly present in the studiedpopulation. It is likely that one of the H6 viruses was newly introduced through migra-tory birds confirming the perceived biosecurity risk. The matrix gene of another H6 viruswas similar to the Australian H7 subtypes, which suggests the reassortment of a previouslyintroduced H6 and local viruses. Similarly, a H9 subtype had a matrix gene similar to thatfound in Asian H9 viruses suggesting reassortment of viruses originated from Australia andAsia. Whilst H5N1 was not found, the serological study demonstrated a constant circulationof the H5 subtype in the sampled birds. The odds of being reactive for avian influenza viralantibodies were 13.1(95% CI: 5.9–28.9) for Pacific Black Ducks over Plumed Whistling Ducks,highlighting that some species of waterfowl pose a greater biosecurity risk. Antibody titreswere slightly higher during warm wet compared with warm dry weather. Routine surveil-lance programmes should be established to monitor the introduction of avian influenzaviruses from Asia and the interactions of the introduced viruses with resident viruses in

order to better detect emerging pathogens in aquatic birds of North Queensland. Surveil-lance should be targeted towards highly susceptible species such as the Pacific Black Duckand carried out during favourable environmental conditions for viral transmission such asthe wet season in northern Australia.

© 2014 Elsevier B.V. All rights reserved.

∗ Corresponding author at: Faculty of Veterinary Medicine, Chittagongeterinary and Animal Sciences University, Zakir Hossain Road, Khulshi,hittagong 4225, Bangladesh. Tel.: +88 031 659093x105;obile: +88 01757796866; fax: +88 031 659620.

E-mail addresses: md.hoque@my.jcu.edu.au,hasanul2010@yahoo.com (Md.A. Hoque).

http://dx.doi.org/10.1016/j.prevetmed.2014.11.009167-5877/© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Wild aquatic birds are considered a natural reservoirof avian influenza viruses (AIV) and harbour all known

subtypes of the influenza A viruses including the highlypathogenic avian influenza (HPAI) strains (Brown et al.,2007; Munster et al., 2007). Australia has recorded fiveoutbreaks of HPAI caused by H7 subtypes in commercial

eterinar

170 Md.A. Hoque et al. / Preventive V

chickens (Selleck et al., 2003: Westbury, 2003). In at leasttwo of the five Australian HPAI outbreaks, surface drinkingwater contaminated with Australian nomadic wild aquaticbird faeces was suspected to be the source of the spill-over infection. Therefore, understanding the epidemiologyof endemic viruses in reservoir hosts such as the Australiannomadic wild aquatic birds is important in order to betterevaluate and mitigate the risk of spill over.

Unlike all other human-inhabited continents, Australiahas remained free of HPAI H5N1, but there are increasedconcerns that the virus could be introduced via migratorybirds which travel between Australia and Southeast Asiawhere HPAI H5N1 is endemic (East et al., 2008). An H5N1outbreak in wild migratory birds at Lake Qinghai, Chinain May 2005 (Lei et al., 2007; Wang et al., 2008) posedserious concerns because the lake is a major breeding sitefor migratory birds whose flyways extend to SoutheastAsia, India, Siberia, Australia and New Zealand. Moreover,the HPAI H5N1 strain has been confirmed in Australia’sclose neighbour Indonesia (Capua and Alexander, 2004)and notably in West Papua on the island of New Guinea(McCallum et al., 2008). Migratory birds could introduceHPAI into Australia via North Queensland, the first pointof entry into Australia for many migratory birds. Australiacould provide another substantial reservoir for the virusresulting in impacts globally on trade and biosecurity.Therefore, conducting surveillance for these exotic virusesand understanding their risk of introduction into Australiais important. This can be achieved through avian influenza(AI) surveillance in high risk biosecurity areas like NorthQueensland and the genetic characterisation and sub-sequent molecular epidemiological understanding of AIsubtypes in Australian nomadic wild aquatic birds.

Although cross sectional studies have been conductedon the epidemiology of AI in Australian nomadic wildaquatic birds (Downie et al., 1977; Senne, 2003; Hayneset al., 2009), systematic repeated studies with the aim ofestimating seroprevalence and determining risk factors forincreased AIV antibody prevalence and hence the risk ofspill over have not been done. The determination of the fullset of hemagglutinin (H) serotypes and patterns of occur-rence over time and their influence on risk of spill over havealso rarely been investigated in nomadic wild aquatic birdsin Australia. Systematic long-term studies with the aim ofestimating AIV ribonucleic acid (RNA) prevalence and thedistribution and reassortment (gene evolution) of AIV sub-types, including exotic viruses, in Australian nomadic wildaquatic birds and how these may contribute to biosecu-rity risks are limited. Therefore, we performed a 3-yearrepeated cross sectional study in North Queensland fromApril 2007 up to March 2010 to understand the molecular-and sero-epidemiology of AI in Australian nomadic wildaquatic birds given it is a biosecurity hotspot with perhapsthe greatest risk of introduction of HPAI into Australia bymigratory birds (Murray et al., 2012). Specific objectivesof the study were to estimate the prevalence of AIV RNAand AIV antibodies, establish the distribution of AIV sub-

types (based on molecular and serological testing), identifyrisk factors associated with the prevalence of AIV anti-bodies and determine the molecular epidemiology of AIVsin nomadic wild aquatic birds in North Queensland. This

y Medicine 118 (2015) 169–181

knowledge will improve our understanding of the biosecu-rity risks posed by AIV in Australian nomadic wild aquaticbirds.

2. Materials and methods

2.1. Sites, sample size and sampling

The epidemiological study of AI was performed onnomadic wild aquatic birds in North Queensland. Birdswhich roosted at different wetlands located within theselected sites were considered as the sampling frame.Birds were sampled from the wetlands of four differentsites in North Queensland (Fig. 1) using the most conve-nient sampling technique. Sites were selected based ontheir proximity to migratory routes, ease of access, thepresence of nomadic wild aquatic birds and generally alarge bird population. A repeated cross sectional studywas performed on nomadic wild aquatic birds at the Bil-labong Sanctuary (19◦22′ S and 146◦54′ E) between April2007 and March 2010 and at Green Acres Lagoon (Cro-marty) (19◦34′ S and 1476◦90′ E) between December 2007and 2009. Opportunistic studies were also conducted atCape York Peninsula (15◦59′ S and 141◦65′ E) and the Ather-ton Tablelands (16◦58′ S and 145◦24′ E) between 2007 and2009.

The study required 138 birds, regardless of species,to be sampled per quarter of each year. Accordingly,1656 samples were needed for the 3-year study atBillabong Sanctuary and 1104 samples for the 2-yearstudy at Green Acres Lagoon. The following formula wasapplied to calculate sample size assuming a 100% sen-sitivity and specificity (Noordhuizen et al., 1997): N =Z2

˛ × SD2/L2(37). N = sample size; Z = 1.96 at 95% confi-dence level; SD2 (variance) = 0.001 for the molecular study(MS) and 0.09 for the serological study (SS); L (absoluteprecision) = 0.005 (MS) and 0.05 (SS). The variance wascalculated as follows: P × (1 − P). P is the expected pointprevalence based on the preliminary data analysis and lit-erature (Haynes et al., 2009) (0.01, MS and 0.10, SS). Aninfinite bird population was assumed.

Birds were sampled quarterly at Billabong Sanctuaryand Green Acres Lagoon (Cromarty) in order to study thetemporal pattern of AI. Funnel traps were mostly used tocapture birds (Ethics approval no. A1175, James Cook Uni-versity: JCU). Birds on Cape York were opportunisticallycaptured for sampling using mist nets and a net launcher(licence no. WISP04524607) between 2008 and 2009. Birdson the Atherton Tableland were sporadically sampled dur-ing 2008.

2.2. Sample collection and epidemiological datarecording

Cloacal and oropharyngeal swabs along with blood sam-ples were collected from each bird. Cloacal swabs aresensitive for identifying low pathogenic AIVs, whereas

oropharyngeal swabs are sensitive for identifying highlypathogenic AIVs (such as H5); serum samples were usedfor serological investigation of AI in cELISA and HI assays.Swabs were taken from birds by inserting swab sticks

Md.A. Hoque et al. / Preventive Veterinary Medicine 118 (2015) 169–181 171

sampli

dbiD0wWdt

firbtg

pde

papwg

Fig. 1. Map showing different

eeply into the vent or oropharyngeal airway and swab-ing the wall. Each of the swab samples was then

nstantly placed into a vial containing 1 mL of sterile chilledulbecco’s modified eagle’s transport medium (Cat. No. 50-20-PB, TropBio, JCU) and stored in an insulated containerith ice packs until transferring to −80 ◦C at the laboratory.hole blood samples (0.5–3 mL, <1% body weight) were

rawn aseptically from wing veins and then immediatelyransferred to individual sterile 15 mL plastic tubes.

Birds were marked by using appropriately sized identi-cation leg bands, each with a unique numbering, beforeeleasing to be able to identify recapture. Handling andanding protocols were followed as described by the Aus-ralian Bird and Bat Banding scheme (http://www.deh.ov.au/biodiversity/science/abbbs).

Capture date, site, species, age, gender, weight and sam-le type were recorded for individual birds. Gender wasetermined by examination of the cloaca with cloacal pli-rs.

Environmental samples consisting of fresh faecal sam-les found in the birds roosting sites were collected in

ddition to swabs from captured birds to increase the sam-le size. Fresh moist faecal samples in the environmentere collected using sterile cotton sticks. The species ori-

in of the faecal samples was determined by observation of

ng sites in north Queensland.

bird species in the area prior to sample collection and alsobased on the size and shape of the faeces. Individual sam-ples were given unique identity numbers. Collection date,site and species were recorded for each sample. Samplestorage details were identical to methods used for swabsof cloaca or oropharynx.

2.3. Molecular evaluation of swab and faecal samples forAvian Influenza

RNA extraction was performed as described in Hoque(2011). Molecular detection of AIV RNA was performedusing one-step real time reverse transcriptase-polymerasechain reaction (rRT-PCR) directed at the Matrix (M) genein a RotorGene 3000 or 6000 PCR cycler (Corbett Robotic,Brisbane, Australia). Two parallel assays designated as Aus-tralian Animal Health Laboratory (AAHL) assay and Wardassay (Heine et al., 2007 and Ward et al., 2004) were per-formed as described (Hoque, 2011). Any test sample thatwas reactive to either of the assays with a threshold valueof ≤40 was considered as reactive for AIV RNA in this study

(Hoque, 2011).

Two-step RT-PCR was performed on reactive samplesfrom the screening assays in an Eppendorf MastercyclerGradient thermocycler to amplify the M, H, non-structural

eterinar

172 Md.A. Hoque et al. / Preventive V

protein and nucleoprotein genes as previously described(Hoque, 2011).

PCR products on the 1% agarose gel were visualisedthrough ethidium bromide staining. Reactions were puri-fied and sequenced by Macrogen (Seoul, South Korea) usingABI 3700 (Applied Biosystems) automated sequencingmachines. Sequence analysing, comparison and phyloge-netic analysis were performed as previously described(Hoque, 2011).

2.4. Serological detection of avian influenza viralantibodies

Serum samples were separated and then evaluatedby a Competitive Enzyme Linked Immuno-sorbent assay(cELISA) (Hoque et al., 2012a,b). Sera that produced ≥40%inhibition in cELISA testing were considered reactive(Hoque et al., 2012a,b). The cELISA reactors were furtherassessed by the haemagglutination inhibition (HI) testsusing a panel of AIV antigens to determine the distributionof AIV serotypes as described (Hoque, 2011). Sera show-ing inhibition at dilutions of 1:16 or greater against fourhaemagglutination units of AIV antigen were consideredreactive for the antibody (Hoque, 2011).

2.5. Statistical evaluation

Statistical analyses were performed using STATA/SETM

11.0 (StataCorp, USA). Descriptive statistics were used toexpress individual results of each category as a frequencypercentage and 95% confidence interval (CI).

2.5.1. Risk factor analysis with Avian Influenzaserological cELISA data

All targeted risk factors were examined. Due to vari-ability in the numbers of birds captured over time it wasnecessary to test three subsets of data in order to ensureadequate sample numbers in the classes of the risk factorcategories under investigation.

For the first data subset (A) we tested data from the pre-dominantly caught species, the Plumed Whistling Duck, atBillabong Sanctuary from June 2007 to May 2009 (N = 394).This data set was suitable for examining the effect ofdifferent seasons on AIV antibody levels. Seasons weredefined as warm dry (September–December), warm wet(January–April) and cool dry (May–August) in order toexamine seasonal effects. Years were divided into June2007–May 2008 and June 2008–May 2009. Body weightwas not normally distributed even after natural log trans-formation; if it was organised into two categories separatedby the 50th percentile (0.20–0.66 and 0.67–1.16 kg). Ageand sex were excluded from the analysis as a significantnumber of birds in this subset were not assessed for ageand sex.

For the second data subset (B) we tested data collectedfrom Plumed Whistling Ducks at Billabong Sanctuary andGreen Acres Lagoon between January 2008 and December

2009 (N = 958). This was the only dataset for assessing theeffect of different locations on AIV antibody levels. Wehad a complete 2-year dataset; therefore year were cate-gorised as 2008 and 2009. Age was grouped into two classes

y Medicine 118 (2015) 169–181

(≤sub-adult and adult) in order to have sufficient samplesin each class. Weight was used as either a continuous vari-able in a linear model or in two classes (0.20–0.64 and0.65–0.91 kg) in a logistic model. Sex was also included asa category in the analysis.

The third data subset (C) consisted of data from PlumedWhistling Ducks and Pacific Black Ducks at BillabongSanctuary and Green Acres Lagoon (Cromarty) collectedbetween September and December 2009 (N = 329). The twocommon species were frequently sampled at both sitesduring this period. This data set allowed us to test theeffect of species on AIV antibody levels along with locationand age. Age was divided into two categories (≤sub-adultand adult) to ensure sufficient sample size. Pacific BlackDucks are slightly larger in size than Plumed WhistlingDucks; therefore, we excluded body weight from this anal-ysis. Sex was not fully determined for sampled Pacific BlackDucks, therefore this factor was excluded from the analy-sis.

In order to identify potential risk factors, both linear andlogistic models were performed as follows.

Linear model: mean percentage inhibition data wereanalysed using one-way analysis of variance (ANOVA) forcategorical factors, while scatter plot and Spearman’s ranktests were used for continuous factors, to identify fac-tors for inclusion into linear models. Factors were selectedfor inclusion in the model according to the identificationof significant univariate analyses (p ≤ 0.20) or accordingto a subjective decision to include biologically interestingfactors. The model was manually constructed by forward-selection applying the maximum likelihood estimationprocedure (Dohoo et al., 2003) and the statistical signifi-cance of the contribution of individual predictors (or groupof predictors) was determined. Wald’s test and the likeli-hood ratio test (LRT) were used to accomplish this (Dohooet al., 2003). We assessed for interaction between factorsby constructing two-interaction product terms for the sig-nificant main effect factors in the model, forcing them intothe model and examining changes in the coefficients and pvalues of the main effects. The presence of confounding fac-tors was investigated by removing one of the variables andassessing changes in the coefficient. A coefficient changeof more than 10% was considered to indicate the presenceof confounding variables. Variance inflation factors (VIF)for the factors were examined to diagnose collinearity andto identify highly correlated factors to avoid a duplicationof effects. A VIF value of more than 10 indicates seriouscollinearity (Dohoo et al., 2003). The Cook–Weisberg testwas used to examine the homogeneity of variance andwhether the overall data fitted the model. The results werepresented for each adjusted factor as a coefficient, p valueand 95% CI.

Logistic model: categories of reactive and non-reactivefor AIV antibody were used as a binary response variable.Sera that produced <40% inhibition in cELISA testing werenon-reactive and sera that produced ≥40% inhibition werereactive (Hoque et al., 2012a,b).

Chi square tests were performed to assess uncondi-tional relationships between the same set of factors thatwere used for the one-way ANOVA testing and betweenthe binary categories of mean percentage inhibition.

Md.A. Hoque et al. / Preventive Veterinary Medicine 118 (2015) 169–181 173

Table 1Numbers of sampled Australian nomadic wild aquatic birds (mostly) and sample types, cloacal and oropharyngeal swabs and faeces and sera along withthe number of AIV RNA and AIV antibody reactors samples (+) in north Queensland (from April 2007 to March 2010).

Time No. birds for moleculartesting (AIV RNA +)a

Cloacal swab(AIV RNA +)

Oropharyngealswab (AIV RNA +)

Faecal sampleb

(AIV RNA +)No. birds for cELISAtesting (AIV Antibody +)

Apr 07 6 (0) 6 (0) 6 (0) 6 (2)May 07 4 (0) 4 (0) 4 (0) 4 (1)Jun 07 19 (0) 19 (0) 19 (0) 48 (2)Jul 07 27 (0) 27 (0) 27 (0) 27 (5)Aug 07 11 (0) 11 (0) 11 (0) 11 (1)Sep 07 19 (0) 19 (0) 19 (0) 19 (2)Oct 07 68 (0) 67 (0) 66 (0) 73 (3)Nov 07 8 (0) 8 (0) 8 (0) 8 (1)Dec 07 31 (0) 31 (0) 31 (0) 31 (3)Jan 08 13 (0) 13 (0) 13 (0) 11 (2)Feb 08 67 (1) 66 (0) 66 (1) 67 (5)Mar 08 79 (0) 78 (0) 78 (0) 79 (9)Apr 08 47 (0) 47 (0) 47 (0) 47 (2)May 08 11 (1) 11 (0) 11 (1) 11 (1)Jun 08 38 (2) 38 (1) 38 (1) 38 (1)Jul 08 67 (0) 67 (0) 67 (0) 79 (9)Aug 08 42 (0) 42 (0) 42 (0) 42 (12)Sep 08 48 (1) 48 (1) 26 (0) 62 (10)Oct 08 47 (0) 47 (0) 47 (0) 107 (0) 47 (9)Nov 08 18 (0) 18 (0) 18 (0) 158 (0) 32 (1)Dec 08 0 0 0 0 0Jan 09 5 (0) 5 (0) 5 (0) 0 5 (1)Feb 09 1 (0) 1 (0) 1 (0) 35 (0) 1 (0)Mar 09 106 (0) 106 (0) 106 (0) 111 (0) 109 (8)Apr 09 114 (0) 114 (0) 114 (0) 14 (0) 118 (5)May 09 36 (2) 36 (0) 36 (2) 51 (0) 36 (4)Jun 09 141 (4) 141 (1) 77 (3) 0 138 (6)Jul 09 34 (2) 34 (2) 34 (0) 127 (0) 16 (0)Aug 09 12 (1) 12 (0) 12 (1) 114 (3) 12 (2)Sep 09 41 (4) 41 (0) 40 (4) 92 (0) 41 (9)Oct 09 126 (0) 126 (0) 126 (0) 141 (3) 126 (23)Nov 09 83 (1) 83 (0) 83 (1) 123 (3) 82 (8)Dec 09 90 (2) 90 (0) 88 (2) 84 (2) 92 (12)Jan 10 0 0 0 0 2 (0)Feb 10 0 0 0 0 0Mar 10 2 (0) 2 (0) 2 (0) 0 2 (0)

Total 1461 (21) 1458 (5) 1368 (16) 1158 (11) 1522 (159)

l or oroa

10 both

Ff

fliacttataa

3

3

t

a Birds were reactive for AIVs if AIV RNA was detected in either cloacabove table.b Faecal samples were collected between October 2008 and January 20

actors with significant �2 results (p ≤ 0.20) and those usedor the linear models were selected for the logistic analysis.

Selection of the most parsimonious model was per-ormed by the same procedure as described for theinear model. We assessed for interaction and confound-ng between factors using the LRT in a similar procedures discussed for the linear model analysis. We tested forollinearity between categorical factors using the two-ailed p value by the �2 test. Two factors were consideredo be collinear if the p value was ≤0.05. The model was thenssessed for goodness-of-fit using the Hosmer–Lemeshowest Dohoo et al. (2003). The results were presented for eachdjusted predictor variable as an odd ratio (OR), p-valuend 95% CI.

. Results

.1. Avian influenza virus RNA and antibody prevalence

A total of 1555 live birds from 19 species were cap-ured from four sites in North Queensland between April

pharyngeal swabs. Results of recaptured birds were not included in the

at Billabong Sanctuary and Green Acres Lagoon (Cromarty).

2007 and March 2010. The results of AIV RNA and antibodyprevalence are presented in Tables 1 and 2.

3.2. Risk factor analysis with cELISA results

No interaction was detected regardless of the classesof data subsets and types of statistical models constructedexcept linear model B where a significant interactionwas detected between year and sex (p = 0.03). Interactionfactors were then adjusted in the model. No collinear vari-ables were identified in the data set. The Cook–Weisbergtest indicated that all linear models (A–C) fitted thedata reasonably well (p ≥ 0.144). The Hosmer–Lemeshowgoodness-of-fit test showed that all logistic models fittedthe data (p ≥ 0.541).

According to the final model based on data subset A(N = 394) year-2 (June 2008–May 2009), warm wet weather

(January–April) and heavier ducks were significantly asso-ciated with higher mean percentage inhibition values ascompared with their respective referent groups (p < 0.05)(Table S1). The odds ratio (OR) of mean percentage

174 Md.A. Hoque et al. / Preventive VeterinarTa

ble

2N

um

bers

of

sam

ple

d

Au

stra

lian

nom

adic

wil

d

aqu

atic

bird

s

(mos

tly)

and

sam

ple

typ

es, c

loac

al

and

orop

har

ynge

al

swab

s

and

faec

es

alon

g

wit

h

the

nu

mbe

r

of

AIV

RN

A

reac

tors

sam

ple

s

(+)

and

AIV

anti

bod

yre

acto

r

sam

ple

s

(+)

in

nor

th

Qu

een

slan

d

(fro

m

Ap

ril 2

007

to

Mar

ch

2010

).

Spec

ies

Scie

nti

fic

nam

e

(Ord

er)

No.

bird

s

for

mol

ecu

lar

test

ing

(AIV

RN

A

+)a

Clo

acal

(AIV

RN

A

+)O

rop

har

ynge

al(A

IV

RN

A

+)Fa

ecal

sam

ple

s(A

IV

RN

A

+)N

o.

bird

s

for

cELI

SA

test

ing

(AIV

An

tibo

dy

+)

Plu

med

Wh

istl

ing

Du

ck

Den

droc

ygna

eyto

ni

(Ans

erif

orm

e-A

)

1180

(17)

1178

(3)

1136

(14)

712

(3)

1209

(90)

Paci

fic

Bla

ck

Du

ckA

nas

supe

rcili

osa

(A)

112

(4)

111

(2)

92

(2)

66

(3)

114

(53)

Mag

pie

Goo

se

Ans

eran

as

sem

ipal

mat

a

(A)

94

94

94

128

119

(7)

Du

sky

Moo

rhen

Gal

linul

a

tene

bros

a

(Gru

ifor

me)

20

20

20

3

20

(0)

Wan

der

ing

Wh

istl

ing

Du

ckD

endr

ocyg

na

arcu

ata

(A)

10

10

0

0

8 (1

)M

usc

ovy

Du

ck

Cair

ina

mos

chat

e

(A)

10

10

10

0

10

(0)

Rad

jah

Shel

ldu

ck

Tado

ra

radj

ah

(A)

8

8

2

0

13

(3)

Gre

en

Pygm

y

Goo

seN

etta

pus

pulc

hellu

s

A)

7

7

0

0

4

(0)

Au

stra

lian

Wh

ite

Ibis

Thre

skio

rnis

mol

ucca

(A)

4

4

4

246

(4)

7

(1)

Bu

sh

Ston

e

Cu

rlew

Burh

inus

gral

lari

us

(Cha

radr

iifor

me)

3

3

3

0

3

(1)

Cro

w

Corv

us

orru

((Pa

sser

ifor

me)

3

3

3

0

3

(0)

Har

dh

ead

Ayt

ha

aust

ralis

(A)

3

3

0

0

2

(0)

Au

stra

lian

Shel

ldu

ck

Tado

rna

tado

rnoi

des

(A)

2

2

2

1 2

(2)

Bla

ck

Swan

Cygn

us

atra

tus

(A)

1

1

1

1 (1

) 1

(1)

Dom

esti

c

Goo

se

Ans

er

anse

r

subs

p. D

omes

ticu

s

(A)

1

1

0

0 1

(0)

Gre

y

Teal

Ana

s

grac

ilis

(A)

1

1

0

0

4

(0)

Wh

ite

Face

d

Her

onEg

rett

a

nova

ehol

land

iae

(Cic

oniif

orm

e)1

1

0

0

0N

anke

en

Nig

ht

Her

onN

ycti

cora

x

cale

doni

cus

(Cic

oniif

orm

e)

0

0

0

0

1

(0)

Pige

on

Duc

ula

bico

lour

(Gal

licif

orm

e)

1

1

1

0

1

(0)

Tota

l

1461

(21)

1458

(5)

1368

(16)

1157

(11)

1522

(159

)

aB

ird

s

wer

e

reac

tive

for

AIV

s

if

AIV

RN

A

was

det

ecte

d

in

eith

er

cloa

cal o

r

orop

har

ynge

al

swab

s.

Res

ult

s

of

reca

ptu

red

bird

s

wer

e

not

incl

ud

ed

in

the

abov

e

tabl

e.

y Medicine 118 (2015) 169–181

inhibition in heavier ducks was 3.2 compared with lighterducks (p = 0.01). Results were non-significant for year andseason (Table S2).

According to the final model based on data subset B(N = 958) year 2009, adult ducks, female and heavier duckswere significantly associated with higher percentage inhi-bition values. Results were non-significant for site (TableS3) (Hoque, 2011). The OR of being reactive for AIV anti-body was 2.9 for adult over ≤sub-adult (p = 0.01); 1.6 forfemale over male (p = 0.187); 2.3 for heavier ducks overlighter ducks (p = 0.005); and 0.90 for Billabong Sanctuaryover Green Acres Lagoon (Cromarty) (p = 0.94) (Table S4).

According to the final model based on data subset C(N = 329), Pacific Black Duck was significantly associatedwith higher mean percentage inhibition values as com-pared with Plumed Whistling Duck (p < 0.001) (Table S5).The OR of being reactive for AIV antibody was 13.1 forPacific Black Duck over Plumed Whistling Duck (p < 0.001)(Table S6).

3.3. Haemagglutination inhibition testing on cELISAreactors

A total of 147 cELISA reactor samples were further eval-uated for H serotyping using the HI test. Thirty five samplesreacted with only one serotype and 95 samples reactedwith multiple serotypes.

Further classification of samples returning multiple Hserotypes (N = 95) was performed by identifying the high-est HI titre (HT) produced. This classification produced54 samples with one dominant serotype and 26 withtwo serotypes, nine with three serotypes, four with fourserotypes and two with more than four serotypes (TableS7). The subtype H9 was the most frequent type for singlereactors (15/35) and H5 and H9 were the most frequentlyoccurring types for the combined data for single and mul-tiple reactors with a single dominate type (42/89). Overallresults suggested that birds were commonly exposed to H3,H5, H6, H9, H11 and H12, and HI titres were generally low.

3.3.1. Longitudinal pattern of haemagglutinin serotypesPlumed Whistling Ducks appeared to be exposed

throughout the study period to H5, H6, H9, H11 and H12subtypes. In contrast Pacific Black Duck was exposed to theH1 subtype over the time period (Table 3). The subtypeH1 was also detected in Magpie Geese. The H4 subtypewas more frequently detected in the samples of PlumedWhistling Ducks obtained during December 2007–April2008 (Table 4).

3.4. Results of screening assays for thecaptured-marked-recaptured Australian nomadic wildaquatic birds

Of 596 swabs (298 each for cloacal and oropharyngealswabs) obtained from 204 recaptured birds (one to fiverecaptures), AIV RNA was detected in one cloacal swab of

an adult recaptured Plumed Whistling Duck at Green AcresLagoon (Cromarty).

We evaluated 295 serum samples that were obtainedfrom 204 recaptured birds (1–5 recaptures). Of 204 birds,

Md.A

. H

oque et

al. /

Preventive V

eterinary M

edicine 118

(2015) 169–181

175

Table 3Matrix of time versus haemagglutinin serotype for Plumed Whistling Ducks at Billabong Sanctuary and Green Acres Lagoon (Cromarty) (from April 2007 to March 2010).

2007 2008 2009 2010

H Ap M J Ju A S O N D Ja F Ma Ap M J Ju A S O N D Ja F Ma Ap M J Ju A S O N D Ja F Ma

13 • • ••••••• •4 ••• • •• •5 • • • • • • • •• ••• ••••6 • • • •• ••••• •7 ••9 • • • • • • ••••• •••••• •11 • • ••• • • • •12 •• • •••••••• • • •••• • • •• •13 • • • •15 • • • • • •n 0 0 2 0 1 2 1 3 2 1 9 2 0 0 0 8 1 4 0 1 7 3 1 1 0 1 4 11 7 6 0N 1 0 17 22 5 17 64 7 26 11 38 71 45 9 25 17 31 17 32 10 0 4 0 101 108 28 74 32 11 31 99 77 82 0 0 3

n, number of cELISA reactors tested for H serotype (78); N = number of serum samples tested using cELISA (115).•: frequency number of serotypes (multiple serotypes in individual samples were counted individually).

Table 4Matrix of time versus haemagglutinin serotype for Pacific Black Ducks at Billabong Sanctuary and Green Acres Lagoon (Cromarty) (from April 2007 to March 2010).

2007 2008 2009 2010

H Ap M J Ju A S O N D Ja F Ma Ap M J Ju A S O N D Ja F Ma Ap M J Ju A S O N D Ja F Ma

1 • • • • • • • •3 • • •• •• ••4 • •5 • • • • •• •••679 • • •• ••••• •••••••• ••11 • • •• ••12 • • • • • •13 • • ••15 • • • • • • •n 0 1 1 1 0 0 1 3 4 3 4 0 1 3 0 1 1 5 12 0 5N 1 0 0 0 2 2 3 0 1 0 0 0 0 1 1 3 6 9 9 1 0 0 0 0 4 3 1 2 1 10 20 1 10 0 0 0

n, number of cELISA reactors tested for HA serotype (46); N = number of serum samples tested using cELISA (91).•: frequency number of serotypes (multiple serotypes in individual samples were counted individually).

eterinar

176 Md.A. Hoque et al. / Preventive V

13 previously non-reactive birds were newly diagnosed asreactors in cELISA testing; 22 birds were reactive at theirfirst capture and 169 birds were consistently non-reactive.Of 22 reactor birds, which were reactive at first capture,only 12 birds maintained reactivity at subsequent recap-tures.

Samples from 35 recaptured birds (N = 204) demon-strated increasing or decreasing or alternate trends in AIVantibody titres at different recaptures (Table S8). Impor-tantly, the results also indicated that AIV antibody lastedmore than 2 months while maintaining a high titre of66–85% mean inhibition in samples from three captured-marked-recaptured Pacific Black Ducks.

3.5. Avian influenza viral sub-typing and sequencingresults

Of 33 reactor samples (including one from a recapturedbird) (N = 1461 live birds and 1157 faecal samples) from theAIV RNA screening assays, we obtained sequencing resultsfor nine samples (Table S9).

3.5.1. Phylogram of the matrix geneWe analysed six of our M gene sequences along with

58 sequences that were obtained from GenBank. Fig. 2provides the results of the phylogenetic analysis andpresents a subset of the isolates having greatest similar-ity to the sequences of reactor samples from our study. Oursequences were most closely related to Australian isolates(for Pacific Black Duck: PBD-1 and Plumed Whistling Duck:PWD-2) or Australian or Eurasian isolates (for PWD-14,Australian White Ibis-1 (AWI-1), PBD-5 and PBD-2) withPWD-14 and AWI-1 being most closely related to eachother (also found in the H sequence of the H6 subtype,Fig. 3). Sequence of PBD-2 was more closely related to NorthAmerican isolates than many other Australian and Eurasianisolates.

3.5.2. Phylogram of the H gene (H6)For phylogenetic analysis we used a 563 base pair

(bp) fragment (124–686) that was obtained for four H6sequenced reactor samples. We used both Australian andoverseas H6 reference sequences that showed the great-est similarity to our H6 reactor sequences for phylogeneticcomparisons and results are presented in Fig. 3. The cleav-age site of reactor sequences in this study indicated lowpathogenic (LP) H6 subtypes. Reactor sequences of PWD-14 and AWI-1 were most similar to each other but also aclose congener with Eurasian isolates and this relationshipwas also consistent with the results from the analysis ofour M gene sequences. Reactor sequences of PWD-2 andPWD-4 were very similar and moderately related to a NewZealand isolate.

3.5.3. Phylogram of the H gene (H9)We sequenced 1350 bp (243–1592) of the H gene for the

single reactor sequence from this study that had the great-

est similarity to the H9 subtype (CY024016, AB303077).We combined this sequence with sequences obtained fromGenBank representing the H9 subtype viruses with closestsimilarity to our sequence for phylogenetic analysis. The

y Medicine 118 (2015) 169–181

H gene sequence of H9 subtype in this study was similarto the sequence of a H9 subtype obtained from ducks inChina (CY024016) and Japan (AB455035), as well as Ireland(AB303077) (Fig. 4). The cleavage site of the isolate in thisstudy indicated a LP H9 subtype.

4. Discussion

4.1. Ingress of Asiatic forms of H6 and H9 subtypes intoAustralia

A globally significant finding is our phylogenetic anal-ysis of the H6 and H9 isolates providing further evidencefor the transmission of AI viruses from Asia and Europe toAustralia.

The phylogenetic analysis suggested that the H6 iso-late (2009) was similar to the isolates from Sharp-TailedSandpipers (Calidris acuminata) for the M gene (CY025198)and Eurasian ducks for the H gene (GQ414861, HQ244430)and the H9 (2009) was a close relative of Asian duck iso-lates for both genes (EF597281, AB455035 and CY024016).These patterns are therefore indicative of recent introduc-tion of H6 and H9 subtypes to Australian nomadic wildaquatic birds from other countries through migratory birdssuch as Sharp-Tailed Sandpipers. These birds and otherspecies of Charadriiformes regularly travel between Asiaand Australia and they can carry LPAI viruses (Tracey et al.,2004; Haynes et al., 2009; Hansbro et al., 2010). Therefore,migratory birds likely have the capacity to introduce exoticHPAI viruses to Australian nomadic wild aquatic birds. Thispossibility is strengthened by evidence that wild migratorybirds infected with HPAI H5N1 are capable of disseminat-ing the virus over long distances (Chen et al., 2005). Ourstudy also identified reassorted H6 subtypes in PlumedWhistling Ducks (2008) according to phylogenetic analy-sis. This analysis indicated a reassortment process couldhave occurred in Australian nomadic wild aquatic birds,which were concurrently infected with the Australian H7wild bird and introduced H6 subtypes. The exotic H6 sub-type may be introduced to Australia through migratorybirds. The H6 novel subtype most likely contains the Aus-tralian backbone M and overseas H genes. Interaction ofbirds belonging to Charadriiforme and Australian-resident(nomadic) Anseriformes on coastal lands of Australia cre-ates conditions for a bird to be infected concurrently withforeign and resident AIV subtypes and then yield a hybridsubtype through a reassortment process. Therefore, thelink between Australian AIV and other countries appears tobe in the association between migratory Charadriiformes,which potentially harbour foreign AIV, and resident Aus-tralian Charadriiformes and Anseriformes. In terms ofbiosecurity, Anseriformes such as wild Australian nomadicducks are more likely to spread viruses to domestic free-range poultry or commercial poultry as they disperse fromcoastal areas and are the final step in a risk pathway forintroduced viruses to cause disease outbreaks in Australiandomestic poultry. The H6 and H9 subtypes identified in this

study through molecular and serological evaluation may beof concern if they are introduced to Australian domestic andcommercial poultry because these subtypes have been pre-viously shown to adapt quickly to domestic avian species

Md.A. Hoque et al. / Preventive Veterinary Medicine 118 (2015) 169–181 177

lications

aa

n“piv

4s

HarHAlta

Fig. 2. Bootstrap consensus trees with 5000 rep

nd can acquire moderate pathogenic potential (Forrestnd Webster, 2010).

Therefore, the AI surveillance programme of Australianomadic wild aquatic birds should continue to provideearly warning” for the introduction of exotic viruses androvide access to virus strains for characterisation, mon-

toring of viral evolution and detection of emerging HPAIiruses.

.2. Absence of highly pathogenic avian influenzaubtypes H5 and H7

Molecular investigation confirmed the absence of HPAI5 and H7 subtypes in samples obtained from live, sicknd dead birds (Hoque et al., 2012a,b) in this study. Theesult is consistent with earlier findings of no HPAI H5 or7 detected from samples of live and dead wild birds in

ustralia (Haynes et al., 2009; Anonymous, 2010). The sero-

ogical study, however, identified a constant circulation ofhe H5 subtype in Plumed Whistling Ducks (2007–2009)nd Pacific Black Ducks (2008–2009) while serological

for AIV matrix gene (between 299 and 777 bp).

evidence for the transmission of the H7 subtype was onlyidentified in Plumed Whistling Ducks in October, 2008. Thepatho-type of these sero-subtypes was not determined, butLPAI H5 and H7 subtypes have been previously isolatedfrom samples of wild ducks in Australia (Haynes et al.,2009; Bulach et al., 2010). Therefore, we assume the H5and H7 subtypes confirmed in serology were non-highlypathogenic subtypes. However, these viruses still pose arisk as AIVs may mutate from LPAI to HPAI after introduc-tion of an LPAI virus to poultry from wild birds (Campitelliet al., 2004). Recently identified LPAI H7 isolates in wildducks were found to be antigenically similar to previousHPAI H7 subtypes which caused several disease outbreaksin Australian poultry between 1975 and 1997 (Bulach et al.,2010). Although the probability of spill-over and evolu-tion of virulence appears to be low, the consequences canbe expensive with the current ongoing H7 outbreaks in

large poultry farms in the state of New South Wales requir-ing destocking (http://www.dpi.nsw.gov.au/agriculture/livestock/poultry/health-disease/avian-influenza, AnimalDiseases and Animal Pests (Emergency Outbreaks) Act

178 Md.A. Hoque et al. / Preventive Veterinary Medicine 118 (2015) 169–181

replicati

Fig. 3. Bootstrap consensus trees with 5000

1991, No. 73). Hence, it is important to better understandthe risk factors associated with spill-over and how they canbe better managed.

4.3. Evidence of multiple H serotypes in individual birds

AIV antibodies corresponding to multiple AIV subtypeswere frequently detected in individual birds in this study.This pattern, however, was not observed in the AIV molecu-lar investigation, which may be due to the limited successin the sub typing of AIV RNA reactor samples. However,some earlier studies have found more than one AIV sub-type infecting individual wild ducks (Halvorson et al., 1983;Jackwood and Stallknecht, 2007). These previous resultsalong with the serotyping results in this study indicate thatwild birds can be infected with more than one AIV subtypethroughout their lives.

4.4. Viral prevalence

Overall AIV RNA prevalence was low (∼1.0%) in the

sampled Australian nomadic wild aquatic birds of NorthQueensland, whereas the AIV antibody prevalence was 11times higher. These results remain consistent with simi-lar studies in Australia and other countries (Haynes et al.,

ons for AIV H6 gene (between 124 and 686).

2009; Raleigh et al., 2009; Hansbro et al., 2010; Tracey,2010). The low AIV RNA prevalence in this study may be dueto a number of reasons. Australian nomadic wild aquaticbirds are frequently exposed to more than one AIV sub-type concurrently or at different times, which significantlydecrease the viral shedding period (Fereidouni et al., 2009,2010) and therefore, the opportunity to detect AIV RNA.Another reason is that the duration of antibodies appears tobe much longer than viral shedding. Analysis of serum sam-ples obtained from captured-marked-recaptured birds inthis study indicated long lasting AIV antibodies. For exam-ple, persistence of AIV antibodies for more than 2 months asobserved in Pacific Black Ducks (66–85%) is the first obser-vation of persistence for this length of time in wild birds inAustralia.

4.5. Risk factors for infection

4.5.1. SpeciesThe AIV RNA prevalence was observed to be sig-

nificantly higher in the samples of Pacific Black Ducks

(3.5–4.5%) than Plumed Whistling Ducks (0.4–1.4%) fromNorth Queensland. A similar species pattern was observedin the screening for AIV antibodies. The odds ratio of beingpositive for AIV antibodies was 13.1 for Pacific Black Ducks

Md.A. Hoque et al. / Preventive Veterinary Medicine 118 (2015) 169–181 179

plication

(Tgsp(as

4

pearioowmi

Fig. 4. Bootstrap consensus trees with 5000 re

53.7%) over Plumed Whistling Ducks (10.1%). Similarly,racey (2010) found that AIV prevalence was significantlyreater in dabbling ducks such as Pacific Black Ducks (Anasuperciliosa) and Mallards (Anas platyrhynchos) (3.1%), com-ared to other ducks such as Plumed Whistling Ducks0.7%) and Wandering Whistling Ducks (1.0%) (26). Over-ll these results suggest that Pacific Black Ducks are moreusceptible to AIV.

.5.2. AgeOur serological screening found a higher AIV antibody

revalence in older birds (linear model B and logistic mod-ls B and C) which might be due to longer lasting AIVntibodies via re-exposure and a boosted immunologicalesponse as indicated in our capture-recapture birds. Heav-er birds were significantly associated with increased levelsf AIV antibodies (linear models A and B) or higher odds

f sero-reactivity (logistic model B). This result correlatesith the age of birds because weight can be a surrogateeasure for age. Our results agreed with a previous finding

n coots (Fulica atra) in Italy (De Marco et al., 2003).

s for AIV H9 gene (between 243 and 1592 bp).

4.5.3. SeasonDucks sampled during warm wet weather

(January–April) had higher AIV antibodies comparedwith warm dry weather (September–December). A similarseasonal effect in the tropics was observed in a study ofnative chickens in Bangladesh (Nooruddin et al., 2006).

4.5.4. YearThere have been no comprehensive studies investigat-

ing temporal influences on AIV antibodies in Australiannomadic wild aquatic birds. A previous cross sectionalstudy documented that the AIV antibody prevalence was17% during 2005–2006, and 25% during 2006–2007 in NewSouth Wales, Australia (Haynes et al., 2009). A similartrend of AIV antibody prevalence was observed in Tasma-nia: 21% during 2005–2006 and 28% during 2006–2007(Haynes et al., 2009). Marked differences in viral preva-lence were also found between years in North American

aquatic birds (Krauss et al., 2004). These temporal patternsof viral and AIV antibody prevalence are little understood,but an epidemiologic periodicity in immunology has beenimplicated (Hinshaw et al., 1985). Certain events within the

eterinar

180 Md.A. Hoque et al. / Preventive V

annual life history of birds may also alter their susceptibil-ity to disease. Energetic trade-offs occur between immunedefences and moulting and breeding, which could makebirds more susceptible to diseases (Hasselquist, 2007;Buehler et al., 2008; Martin et al., 2008). Cyclic patternshave been described for measles and whooping cough inhumans, and are determined in part by spatial factors, herdimmunity, and population age-structure (Rohani et al.,1999). Increased rainfall and temperature in the secondyear may explain the associated increased antibody levelsfor Plumed Whistling Ducks (average rainfall: 131 mm in2008 and 161 mm in 2009; average temperature: 23 ◦C in2008 and 24 ◦C in 2009; SILO, Department of Environmentand Resource Management, Queensland). These environ-mental factors could suppress the immune system of birdsthereby rendering them more susceptible to an infectiousdisease like AI. An increased bird population density wasalso observed in the second year as compared with the firstyear (for example, in dataset C, we sampled 312 PlumedWhistling Ducks in 2008 and 647 Plumed Whistling Ducksin 2009).

4.5.5. SiteSites did not differ in prevalence of AIV antibodies. The

Mareeba area in North Queensland has been identified asa risk area for AI emergence in poultry (East et al., 2008;Tracey, 2010) and Cape York Peninsula is a close entry pointfor migratory birds in North Queensland. Therefore, futureAI studies should focus on these two higher risk areas inNorth Queensland, Australia.

In conclusion surveillance should be targeted towardshighly susceptible species such as the Pacific Black Duckand during times of favourable environmental conditionsfor viral transmission such as the wet season in northernAustralia.

The results may not represent the target populationbecause trapped birds may not be a random sample of thepopulation. For example we do not know whether AI posi-tive or AI negative birds are equally likely to be trapped.

Ethical approval

Birds were sampled per season (every 4 months)at Billabong Sanctuary and Cromarty, and AthertonTableland (sporadically). Funnel traps were mostly usedto capture birds at the above study sites (Ethicsapproval no. A1175, JCU, Townsville, QLD and Eco-access permit no. WISp04374507, Queensland Parks andWildlife Service, Northern Region, Australia). Birds onCape York were opportunistically captured for sam-pling using mist nets and a net launcher (licence no.WISP04524607, Queensland Parks and Wildlife Service,Northern Region, Australia) between 2008 and 2009.Handling and banding methodology were applied as spec-ified by the Australian Bird and Bat Banding scheme(http://www.deh.gov.au/bidiversity/science/abbbs).

Acknowledgments

We extend our thanks to scientific and ethics permitauthorities for providing the following project approvals,

y Medicine 118 (2015) 169–181

Eco-access permit no. WISp04374507, Queensland Parksand Wildlife Service, Northern Region and ethics permit no.A1175, JCU. Finally, we extend our heartfelt thanks to theother contributors (Stephen Garland who helped in review-ing the script; David Roshier who helped in catching birdsand providing samples of wild birds from Cape York. RaviDissanayake who helped in producing Fig. 1) to this study.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.prevetmed.2014.11.009.

References

Anonymous, 2010. Wild bird mortality events and avian influenza exclu-sion. Newsl. Aust. Natl. Anim. Health Inf. Syst. 15, 8.

Brown, J.D., Swayne, D.E., Cooper, R.J., Burns, R.E., Stallknecht, D.E., 2007.Persistence of H5 and H7 avian influenza viruses in water. Avian Dis.51 (1), 285–289.

Buehler, D.M., Piersma, T., Matson, K., Tieleman, B.I., 2008. Seasonal redis-tribution of immune function in a migrant shorebird: annual-cycleeffects override adjustments to thermal regime. Am. Nat. 172 (6),783–796.

Bulach, D., Halpin, R., Spiro, D., Pomeroy, L., Janies, D., Boyle, D.B., 2010.Molecular analysis of H7 avian influenza viruses from Australia andNew Zealand: genetic diversity and relationships from 1976 to 2007.J. Virol. 84 (19), 9957–9966.

Campitelli, L., Mogavero, E., De Marco, M.A., Delogu, M., Puzelli, S., Frezza,F., Facchini, M., Chiapponi, C., Foni, E., Cordioli, P., Webby, R., Barigazzi,G., Webster, R.G., Donatelli, I., 2004. Interspecies transmission of anH7N3 influenza virus from wild birds to intensively reared domesticpoultry in Italy. Virology 323 (1), 24–36.

Capua, I., Alexander, D.J., 2004. Avian influenza: recent developments.Avian Pathol. 33 (4), 393–404.

Chen, H., Smith, G.J., Zhang, S.Y., Qin, K., Wang, J., Li, K.S., Webster, R.G.,Peiris, J.S., Guan, Y., 2005. Avian flu: H5N1 virus outbreak in migratorywaterfowl. Nature 436, 191–192.

De Marco, M.A., Foni, G.E., Campitelli, L., Raffini, E., Di Trani, L., Delogu, M.,Guberti, V., Barigazzi, G., Donatelli, I., 2003. Circulation of influenzaviruses in wild waterfowl wintering in Italy during the 1993–99 peri-ods: evidence of virus shedding and sero-conversion in wild ducks.Avian Dis. 47 (3), 861–866.

Dohoo, I., Martin, W., Stryhn, H., 2003. In: McPike, S.M. (Ed.), VeterinaryEpidemiologic Research. AVC Inc., Charlottetown, PE, Canada.

Downie, J.C., Hinshaw, V., Laver, W.G., 1977. The ecology of influenza. Iso-lation of type ‘A’ influenza viruses from Australian pelagic birds. Aust.J. Exp. Biol. Med. Sci. 55 (6), 635–643.

East, I.J., Hamilton, S., Sharp, L.A., Garner, M.G., 2008. Identifying areas ofAustralia at risk for H5N1 avian influenza infection from exposure tonomadic waterfowl moving throughout the Australo-Papuan region.Geospat. Health 3 (1), 17–27.

Fereidouni, S.R., Starick, E., Beer, M., Wilking, H., Kalthoff, D., Grund, C.,Häuslaigner, R., Breithaupt, A., Lange, E., Harder, T.C., 2009. Highlypathogenic avian influenza virus infection of mallards with homo- andheterosubtypic immunity induced by low pathogenic avian influenzaviruses. PLoS ONE 4 (8), e6706.

Fereidouni, S.R., Werner, O., Starick, E., Beer, M., Harder, T.C., Aghakhan,M., Modirrousta, H., Amini, H., Moghaddam, M.K., Bozorghmehrifard,M.H., Akhavizadegan, M.A., Gaidet, N., Newman, S.H., Hammoumi, S.,Cattoli, G., Globig, A., Hoffmann, B., Sehati, M.E., Masoodi, S., Dod-man, T., Hagemeijer, W., Mousakhani, S., Mettenleiter, T.C., 2010.Avian influenza virus monitoring in wintering waterbirds in Iran,2003–2007. Virol. J. 7, 43.

Forrest, H.L., Webster, R.G., 2010. Perspectives on influenza evolution andthe role of research. Anim. Health Res. Rev. 11 (1), 3–18.

Halvorson, D., Karunakaran, D., Senne, D., Kelleher, C., Bailey, C., Abraham,A., Hinshaw, V., Newman, J., 1983. Epizootiology of avian influenza-simultaneous monitoring of sentinel ducks and turkeys in Minnesota.

Avian Dis. 27 (1), 77–85.

Hansbro, P.M., Warner, S., Tracey, J.P., Arzey, K.E., Selleck, P., O’Riley,K., Beckett, E.L., Bunn, C., Kirkland, P.D., Vijaykrishna, D., Olsen, B.,Hurt, A.C., 2010. Surveillance and analysis of avian influenza viruses,Australia. Emerg. Infect. Dis. 16 (12), 1896–1904.

eterinar

H

H

H

H

H

H

H

J

K

L

M

M

Md.A. Hoque et al. / Preventive V

asselquist, D., 2007. Comparative immuno-ecology in birds: hypothesesand tests. J. Ornithol. 148 (2), 571–582.

aynes, L., Arzey, E., Bell, C., Buchanan, N., Burgess, G., Cronan, V., Dickason,C., Field, H., Gibbs, S., Hansbro, P., Hollingsworth, T., Hurt, A., Kirkland,P., McCracken, H., O’Connor, J., Tracey, J., Wallner, J., Warner, S., Woods,R., Bunn, C., 2009. Australian surveillance for avian influenza virusesin wild birds between July 2005 and June 2007. Aust. Vet. J. 87 (7),266–272.

eine, H.G., Trinidad, L., Selleck, P., Lowther, S., 2007. Rapid detectionof highly pathogenic avian influenza H5N1 virus by TaqMan reversetranscriptase-polymerase chain reaction. Avian Dis. 51 (1), 370–372.

inshaw, V.S., Wood, J.M., Webster, R.G., Deibel, R., Turner, B., 1985.Circulation of influenza viruses and paramyxoviruses in waterfowloriginating from two different areas of North America. Bull. WorldHealth Organ. 63 (4), 711–719.

oque, M.A., (PhD Thesis) 2011. Risk of Spill-Over of Diseases (inParticular Avian Influenza) from Wild Aquatic Birds in North Queens-land. James Cook University, Townsville, Queensland, Australia,http://eprints.jcu.edu.au/21582

oque, M.A., Burgess, G.W., Greenhil, A.R., Hedlefs, R., Skerratt, L.F., 2012a.Causes of morbidity and mortality of wild aquatic birds of BillabongSanctuary, Townsville, North Queensland, Australia. Avian Dis. 56 (1),249–256.

oque, M.A., Skerratt, L.F., Garland, S., Burgess, G.W., Selleck, P., 2012b.Adaptations of a competitive enzyme-linked immunosorbent assaysfor the detection of antibodies to influenza A virus in horse sera foruse in wild aquatic birds. Indian J. Virol. 23 (3), 261–269.

ackwood, M.W., Stallknecht, D.E., 2007. Molecular epidemiologic studieson North American H9 avian influenza virus isolates from waterfowland shorebirds. Avian Dis. 51 (1), 448–450.

rauss, S., Walker, D., Pryor, S.P., Niles, L., Chenghong, L., Hinshaw, V.S.,Webster, R.G., 2004. Influenza A viruses of migrating wild aquatic birdsin North America. Vector Borne Zoonotic Dis. 4 (3), 177–189.

ei, F., Tang, S., Zhao, D., Zhang, X., Kou, Z., Li, Y., Zhang, Z., Yin, Z., Chen, S.,Li, S., Zhang, D., Yan, B., Li, T., 2007. Characterization of H5N1 influenzaviruses isolated from migratory birds in Qinghai province of China in2006. Avian Dis. 51 (2), 568–572.

artin, L.B., Weil, Z.M., Nelson, R.J., 2008. Seasonal changes in vertebrate

immune activity: mediation by physiological trade-offs. Philos. Trans.R. Soc. Lond. B: Biol. Sci. 363 (1490), 321–339.

cCallum, H.I., Roshier, D.A., Tracey, J.P., Joseph, L., Heinsohn, R., 2008.Will Wallace’s line save Australia from avian influenza? Ecol. Soc. 13(2), 1–17.

y Medicine 118 (2015) 169–181 181

Munster, V.J., Bass, C., Lexmond, P., Waldenstrom, J., Wallensten, A., Frans-son, T., Rimmelzwaan, G.F., Beyer, W.E.P., Schutten, M., Olsen, B.,Osterhaus, A.D.M.E., Fouchier, R.A.M., 2007. Spatial, temporal andspecies variation in prevalence of influenza A viruses in wild migratorybirds. PLoS Pathog. 3 (5), e61.

Murray, K.A., Skerratt, L.F., Speare, R., Ritchie, S., Smout, F., Hedlefs, R., Lee,J., 2012. Cooling off health security hot spots: getting on top of it downunder. Environ. Int. 48, 56–64.

Nooruddin, G.M., Hossain, M.T., Mohammad, M., Rahman, M.M., 2006.Sero-epidemiology of avian influenza in native chicken in Bangladesh.J. Poul. Sci. 5 (11), 1029–1033.

Noordhuizen, J.P.T.M., Frankena, K., Van der Hoofd, C.M., Graat, E.A.M.,1997. In: Nordhuizen, J.P.T.M. (Ed.), Application of Quantitative Meth-ods in Veterinary Epidemiology. Wageningen Press, Wageningen, TheNetherlands.

Raleigh, P.J., Flynn, O., O’Connor, M., O’Donovan, T., Purcell, B., De Burca,M., Regazzoli, V., Minihan, D., Connell, J., Markey, B.K., Sammin, D.J.,2009. Avian influenza viruses detected by surveillance of waterfowlin Ireland during 2003–2007. Epidemiol. Infect. 137 (4), 464–472.

Rohani, P., Earn, D.J., Grenfell, B.T., 1999. Opposite patterns of synchronyin sympatric disease meta-populations. Science 286 (5441), 968–971.

Selleck, P.W., Arzey, G., Kirkland, P.D., Reece, R.L., Gould, A.R., Daniels,P.W., Westbury, H.A., 2003. An outbreak of highly pathogenic avianinfluenza in Australia in 1997 caused by an H7N4 virus. Avian Dis. 47(3), 806–811.

Senne, D.A., 2003. Avian influenza in the western hemisphere includingthe Pacific Islands and Australia. Avian Dis. 47 (3), 798–805.

Tracey, J.P., Woods, R., Roshier, D., West, P., Saunders, G.R., 2004. The roleof wild birds in the transmission of avian influenza for Australia: anecological perspective. Emu 104 (2), 109–124.

Tracey, J.P., 2010. Risk-based surveillance of avian influenza in Australia’swild birds. Wildl. Res. 37, 134–144.

Wang, G., Zhan, D., Li, L., Lei, F., Liu, B., Liu, D., Xiao, H., Feng, Y., Li, J., Yang,B., Yin, Z., Song, X., Zhu, X., Cong, Y., Pu, J., Wang, J., Liu, J., Gao, G.F., Zhu,Q., 2008. H5N1 avian influenza re-emergence of Lake Qinghai: phylo-genetic and antigenic analyses of the newly isolated viruses and rolesof migratory birds in virus circulation. J. Gen. Virol. 89 (3), 697–702.

Ward, C.L., Dempsey, M.H., Ring, C.J., Kempson, R.E., Zhang, L., Gor, D.,

Snowden, B.W., Tisdale, M., 2004. Design and performance testingof quantitative real time PCR assays for influenza A and B viral loadmeasurement. J. Clin. Virol. 29 (3), 179–188.

Westbury, H.A., 2003. History of highly pathogenic avian influenza inAustralia. Avian Dis. 47 (special issue), 23–30.