The study was conducted in Institute of Medical Research (IMR) at Goroka, Eastern Highlands Province. Retrospective poultry samples were used and were originally collected by the National Agriculture Quarantine Inspection Authority (NAQIA) and Department of Agriculture and Livestock (DAL), during their annual surveillance study in poultry survey among the four regions: Southern, Momase, New Guinea Island (NGI), and Highlands, from June 2011 to April 2012. About 385 cloacal swabs were collected and sent to the IMR. The study carried out was designed as a cross-sectional study looking at poultry samples collected over a one year period from different poultry farms in the country. However, no longitudinal analysis was performed due to insufficient collection time points available.

DNA extraction from cloacal swabs were performed using DNeasy Blood and Tissue Kit (QIAGEN Germany), according to the manufacturer’s instructions. DNA extraction was done in the biological safety cabinet class II. A total of 200 µl of the sample was transferred into sterile microcentrifuge tube (1.5 ml) (Eppendorf^®, Germany). About 20 µl of Proteinase K was added into the reaction tube followed by adding of 200 µl of Buffer AL. The samples were then vortex and incubated in the hot water bath at 56˚C for 10 min. After incubation, 200 µl of ethanol (96% - 100%) was added and the samples were mixed thoroughly by vortexing. After vortexing, all the samples in the reaction tubes were pipette into a new DNeasy mini spin column placed in a 2 ml collection tube. The tubes were then centrifuged at 8000 revolutions per minutes (rpm) (≥6000 g) for 1min. The flow was discarded and the DNeasy spin column tubes were placed into a new 2 ml collection tube. 500 ul of Wash Buffer AW1 was added and the suspension was centrifuged at 8000 rpm (≥6000 g) for 1min. The flow was discarded and the spin column was placed into a new collection tube. Then 500 µl of Wash Buffer AW2 was added and the samples were centrifuged again for 3 min at 14,000 rpm (20,000 g). The flow was discarded the spin column was placed into a new 1.5 ml labelled microcentrifuge tubes. Finally, 200 µl of elution Buffer AE was directly added into the DNeasy membrane and centrifuged for 1 min at 8000 rpm. The DNeasy spin column was discarded and the DNA extract was ready for real-time PCR.

Salmonella and Campylobacter primers used in real-time PCR are listed below (Table 1). Primers for Salmonella are specific for outer membrane protein “ompF” gene and primers for Campylobacter are mainly for 16S rRNA gene.

Nucleic acid extracted from cloacal swabs was tested for Salmonella and Campylobacter spp. using real-time PCR analysis. PCR preparation was done by creating a plate map followed with master-mix reaction preparation (Table 2).

All the reagents were thawed down in the bio-safety hood cabinet and pipette into the reaction Eppendorf tube (1.5 ml) (Interpath Australia). The reaction tube was then centrifuge using the micro-centrifuge. 18 µl of the reaction solution was aliquot into each Hard-Shell^® PCR Plates 96-well white (Bio-Rad Australia)

on the green bio-freezer rack and capped with Optical Flat 8-Cap Strips (Bio-Rad Australia). The PCR plates were then covered with aluminium foil and transferred to Dangerous Pathogen Laboratory (DPL) for addition of templates and the controls. Extracted DNA (templates), positive control, and non-template control were added to the PCR well strips respectively at a volume of 2 µl each. The samples were then transferred to the wet lab for molecular analysis using real-time PCR.

The PCR strips were placed in the max-centrifuge (Thermo IEC, Centra Gp8R). The rpm was set to 520 for 1 min. After centrifuging the strips were wiped with tissue and placed in the CFX 96 Real-time PCR system C1000 Thermol Cycler. Salmonella and Campylobacter have the same PCR cycling parameters: Initial heating occurred at 50˚C for 2 min followed by denaturation at 95˚C at 15 sec and 40 cycles of annealing temperature at 94˚C for 1 min and extension temperature at 60˚C for 1 min (Table 3). The amplification read at (FAM) Fluorophores and the final results amplified after 2 hrs 6 min.

All statistical analysis was performed using Real-Stats in Microsoft Excel. Comparison of Salmonella and Campylobacter spp. in different farming methods and in different regions was subjected to chi-square test. Further analysis was done using Fisher’s Exact Test to determine the significance difference of Salmonella and Campylobacter among each of the farming practices, and regions. A p-value of 0.05 or less was considered significant.

Prevalence of Salmonella and Campylobacter spp. were compared among different poultry farming methods: commercial, semi-commercial and non-commercial farming methods. A total of 8 cloacal swabs were collected from commercial farms, 6 from semi-commercial and 33 from non-commercial farms. Salmonella was not observed among the farming methods. However, Campylobacter was observed at 83.3% (5/6) in semi-commercial farms and 6.1% (2/33) in non-commercial farms but absent in commercial farms (Table 4).

Significant differences was observed only in commercial and semi-commercial farms (p-value = 0.0030) and semi-commercial and non-commercial farms (p-value = 0.0002).

A total of 57 cloacal swabs were collected from different poultry farming methods in the NGI region. Salmonella was positive at 5.9% (1/17) in semi-commercial and 2.6% (1/38) in non-commercial farms. Campylobacter was observed at 50.0% (1/2), 29.4% (5/17) and 21.1% (8/38) in commercial, semi-commercial and non-commercial farms respectively (Table 5).

A total of 150 cloacal swabs samples were collected and tested among different poultry farming methods in the Momase region. Salmonella spp. was not observed in the different farming methods. Semi-commercial farms had the highest prevalence of Campylobacter spp. at 57.7% (45/78), with commercial farms at 52.0% (13/25) and non-commercial farms at 12.7% (6/47) (Table 6).

Significant differences were observed between Campylobacter in commercial and non-commercial (p-value = 0.0006) and semi-commercial and non-commercial farm (p-value < 0.0001).

A total of 131 cloacal swabs were collected from poultry farms in the Highlands region. Twenty four swabs were collected from commercial farms, 62 from semi-commercial and 45 from non-commercial farms. Salmonella spp. was observed in both semi-commercial and non-commercial farms at 1.6% (1/62) and

2.2% (1/45) respectively. Campylobacter spp. was found to have a higher prevalence rate in semi-commercial farms at 56.5% (35/62) compared to commercial farms at 8.3% (2/24) and least prevalent in non-commercial farms at 17.8% (8/45) (Table 7).

Significant differences were observed for Campylobacter spp. between commercial and semi-commercial (p-value < 0.0001) and commercial and non-commercial (p-value < 0.0001) farming practices in Highlands region.

All data from different farming methods and data from all the regions were pooled, and subjected to chi-square test. Semi-commercial farms were observed to have higher carriage rates of Campylobacter spp. at 52.1% (85/163) compared to commercial farms at 27.1% (16/59) and non-commercial farms at 9.8% (16/163). Salmonella spp. was present only in semi-commercial farm and non-commercial farms at 1.2% (2/163) and 1.2% (2/163) respectively, but absent in commercial farms (Table 8).

Significant differences were observed for Campylobacter spp. between commercial and semi-commercial farms (p-value = 0.0012) and semi-commercial and non-commercial farm (p-value < 0.0001).

Data from the three different farming practices was pooled, and the presence of Salmonella and Campylobacter spp. was analysed between the different regions: Southern, Momase, NGI, and Highlands by chi-square analysis followed by Fisher’s exact test. Campylobacter was observed in all four regions while Salmonella was largely absent. The Momase and Highlands regions had the higher prevalence rate of Campylobacter at 42.7% (64/150) and 34.4% (45/131) respectively compared to the NGI and Southern regions at 24.6% (14/57) and 14.9% (7/47) respectively (Table 9). Salmonella was observed but at a very low prevalence rate only in NGI region and Highlands region at 3.5% (2/57) and 1.5% (2/131) respectively. Significant differences in Salmonella between regions were not observed.

Significant differences were observed for Campylobacter between the Southern and Momase regions (p-value = 0.0008), Southern and Highlands regions (p-value = 0.0132), and Momase and NGI regions (p-value = 0.0313). Significant differences in Salmonella between regions were not detected.

Campylobacter and Salmonella were both detected from cloacal samples collected among the three different poultry farming methods in the country. There is a high prevalence of Campylobacter in semi-commercial (52.5%) and commercial (27.1%) farms and least in non-commercial or free range system (9.8%) in the present study. The differences in the prevalence of Campylobacter among the farming methods could be as a result of different agricultural activities such as farming methods, vaccines, deep litter management and biosecurity practices in the farm.

Vaccination of chickens is regarded as an additional measure to increase resistance of birds against Salmonella infection and decrease shedding of viable organisms (Dewaele et al., 2012 [3] ; Hassan and Curtiss, 1996 [4] ). Studies have proven that vaccination works by reducing the prevalence of Salmonella in breeder hens and the progeny or by increasing the passive immunity of the birds (broiler and layer) and blocking the horizontal transmission of Salmonella to the flocks (Dórea et al., 2010) [5] . However, no suitable vaccine is readily available for Campylobacter infection (Ahmed et al., 2013 [6] ; Schroeder et al., 2014 [7] ). Hence, chicken are highly susceptible to Campylobacter transmission and colonization than Salmonella.

Under normal poultry farming condition, chicks usually become infected with Campylobacter at about 2 - 3 weeks of age (Hodgins et al., 2015 [8] ; Lin, 2009 [9] ). Maternal antibodies providing passive immunity are transferred to the avian embryos from serum to egg yolks (Hamal et al., 2006) [10] . The level of antibodies is very high at 3 to 4 days after hatching and thereafter gradually decreases to undetectable levels at 2 to 3 week (Sahin et al., 2001) [11] . Once the chicken is infected with Campylobacter, the organisms usually spread so rapidly that close to 100% birds are reported to be colonized for a shorter period of time (Hamal et al., 2006) [10] . The pathogen survives in the chicken gut throughout the grow-out period (Hodgins et al., 2015) [8] .

Poultry litter provides an ideal environment for Campylobacter growth however other researchers have observed lower prevalence of Salmonella in litter. Environmental factors in the poultry house play an important role on the growth of Campylobacter in the litter. Since Campylobacter is sensitive to drying, increased temperature, lower ventilation rate and high humidity with an increasing pH in the litter encourages higher growth of Campylobacter (Chinivasagam et al., 2009) [12] . This could be the reason for the non-detection of Campylobacter in non-commercial poultry farms. Litter management in the farm is a very important agricultural practice to reduce prevalence of Campylobacter (Line and Bailey, 2006) [13] .

Urea and uric acids make up 70% of the nitrogen content of poultry litter and are readily degraded to ammonia (Newell et al., 2011 [14] ; Rothrock et al., 2008 [15] ). Ammonia emissions beyond the threshold level of 25 ppm in the poultry house causes complication in chicken’s health. Birds starting to develop stressful conditions, infection occurs in respiratory system and eventually decreases immunity. Thus leaving the birds susceptible to bacterial colonization and infection (Line and Bailey, 2006) [13] .

Prevalence of Salmonella and Campylobacter in poultry is determined by the on-farm biosecurity practices (Colles et al., 2015 [16] ; Hald et al., 2000 [17] ). Biosecurity is a set of preventative measures to reduce the risk of transmission in poultry farms in three main areas: isolation, mobility, and cleanliness and hygiene. Effective biosecurity measures to control Salmonella and Campylobacter in poultry farms include all-in-all-out system followed by disinfection of sheds, restricted movement of birds, workers and equipment, lack of contact with migratory birds, use of feed pellets, chlorination of drinking water, proper litter management, washing and sanitizing of hatching eggs, and purification of air in hatching cabinets (Colles et al., 2015 [16] ; Conan et al., 2012 [18] ). The prevalence of Salmonella and Campylobacter in the farm is not guaranteed to be eliminated completely, but lowering the prevalence is an important strategy for reducing contamination of birds entering the processing plant and minimising risk of contaminated meat and egg products entering the food chain (McCrea et al., 2006 [19] ; Trampel et al., 2014 [20] ).

Poultry housing system in PNG provides free access for movement of insects such as Synanthropic flies, in particular the house fly Musca domestica and flies belonging to Diptera: Brachycera spp. which have been associated with animal and human activities (Hald et al., 2008 [17] ; Trampel etal., 2014 [20] ). Darkling beetles (Hazeleger et al., 2008) [21] and rodents are ubiquitous and thrives wherever food stocks and habitats are plentiful and can live in close proximity to humans and livestock (Meerburg et al., 2006) [22] .

Salmonella and Campylobacter positivity were identified among the four regions in the country, Southern, Momase, NGI and the Highlands regions. Campylobacter was found to be more prevalent than Salmonella. It was observed that Campylobacter had a higher prevalence in the Momase and Highland regions at 42.7% and 34.4% respectively compared to the NGI (24.6%) and Southern (14.9%) regions. Salmonella was observed only in NGI and Highlands regions at (3.5%) and (1.5%). Currently there is no information available about the poultry farming practices in each of the regions. However, it is assumed that higher prevalence of Campylobacter in Momase and Highlands regions is believed to be associated with poultry farms having poor biosecurity practices.