Seroprevalence of Newcastle Disease Virus in Local Chickens in Njombe and Bahi Districs in Tanzania

Newcastle disease virus (NDV) causes signicant losses of poultry in Tanzania. Like in many African countries, the regular surveillance of NDV is important for the control of disease. The objective of this study was to determine seroprevalence of NDV in backyard poultry in Bahi and Njombe districts of Tanzania over the rainy (May) and dry (November) seasons in 2016. Using hemaglutination inhibition test, the overall seroprevalence was determined to be 26.8%. The signicant differences in seroprevalence was between seasons ((higher (34.9%) in dry season, p < 0.0001)) and age ((higher (30.3%), p < 0.0001 in adult birds)). There were no signicant differences in seroprevalence between the districts or sex. The higher levels of “protective” antibody titers were signicantly associated with location; Njombe (RR 1.15), dry season (RR 1.08), and age; adult birds (RR 1.16), however the prevalence of these titers was not high enough to conclude any herd immunity among these ocks. This study therefore provides a brief insight of NDV disease dynamics in Tanzania. Future studies focusing on transmission, strain type, and monthly dynamics of NDV in backyard ocks will provide greater insight into the disease dynamics and allow new practical strategies to alleviate the effects of NDV for the smallholder farmers


Introduction
Newcastle disease virus (NDV) causes a devastating disease in poultry, especially in developing countries due to the constant threat of infection from lack of proper vaccination strategies and scavenging environments (Alexander, 2000;Ananth et al., 2008). With the greatest impacts of NDV occurring in backyard poultry due to high mortality rates in susceptible ocks, the pathogen poses signi cant threats to small-holder farmers whose livelihoods depend on the productivity of their ocks for both egg and meat production (Alexander, 2000;Ananth et al., 2008;Alexander et al., 2012).
Many countries in both Africa and Asia have reported endemic NDV, including Kenya, Ethiopia, Nigeria, Sudan, Uganda, Tanzania (Awan et al., 1994;Lawal et al., 2016). Although these countries have reported positive serology, there is no indication of whether the backyard chickens were exposed to lentogenic (low pathogenic), mesogenic (moderate pathogenic), or velogenic (highly pathogenic) strains of the virus, which is particularly important due to all three types of strains previously being isolated from Tanzanian backyard poultry (Awan et al., 1994;Yongolo et al., 2011). It is important to understand the overall prevalence of NDV as well as the pathogenicity of these eld isolates in order to understand the impacts of infection in these countries.
In Tanzania, like most other African countries, there is a large population of backyard poultry with over 90% of all poultry being farmed extensively (Ministry of Agruculture et al., 2016). In these conditions, the chickens are regularly exposed to inadequate feeds, poor management and housing, predation, and infectious diseases (Minga et al., 2004;Gilbert et al., 2015;Lawal et al., 2016;Muchadeyi and Dozmba, 2017). There are also risk factors in these farms that contribute to the spread and persistence of infection including the absence of preventative measures or vaccination, frequent contact with wild birds, improper disposal of carcasses and waste products, and the general practice of selling/buying birds which all have the possibility of introducing disease to healthy birds and ocks (Chaka et al., 2013). Other factors also contribute to infection in the chickens, such as age, immune status, season, breed, and other environmental factors (Awan et al., 1994).
This study in Tanzania was implemented to provide a better understanding of the seroprevalence of NDV in backyard poultry and to assess risk factors associated with the disease. The information may be useful in devising new improved strategies for the control and prevention of ND in smallholder farmers.

Study area
The study was conducted in Njombe and Bahi districts in Tanzania ( Figure 1). Njombe district is found inNjombe region in the southern highlands of Tanzania and is roughly 2000 meters above sea level.It is  Table 1.

Ethical issues and consent to participate
The research was approved by the institutional committee for research and innovation at the Nelson Mandela African Institution of Science and Technology (NM-AIST). The sampling and handling of chickens was performed by a licensed veterinarian. Permission to conduct the research was approved by the Government leaders at the district and village levels. Recruitments of farmers to the study were done after informed consents were provided by farmers themselves, either by the household head or alternative responsible person.

Blood collection from chickens and Hemagglutinationinhibition test
Approximately 2ml of blood was collected aseptically from the wing vein in a disposable syringe and stored at room temperature for six hours. Serum was gently separated and stored via eppendorf vials in a cooler box with ice blocks until arrival at the laboratory where they were stored at -20°C.
Hemagglutination inhibition (HI) tests were performed according to protocols described by OIE (2012). Brie y, 1% chicken red blood cells (RBC) were prepared from apparently healthy, unvaccinated 8 weeks old layer chickens. The viral antigen was from an I-2 attenuated vaccine obtained from Tanzania Veterinary Laboratory Agency (TVLA) using embryonated chicken eggs as previously described (OIE), 2012). HI tests were performed by rst dispensing 25µl of phosphate buffered saline (PBS) into wells of a plastic V-bottomed 96-well microtiter plate. Then 25µl of test sera were added into the rst row of wells followed by serial dilutions by transferring 25µl of the content to consecutive wells until the 10th column.
This was followed by adding a 25µl of previously prepared 4HA antigen up to 11th column. The plate was left to stand for about 30 minutes at ambient temperature. The 11 th and 12 th wells were considered control wells and for checking non-speci c agglutinins respectively. In each test, previously prepared positive and negative control sera were included in the test. A 25µl of 1% chicken RBCs was added to each well and, after gentle mixing, the plate was left for about 40 minutes at room temperature to allow settling of RBC. The agglutination inhibition was assessed by tilting the plates to observe RBC stream in reference to controls.

Data analysis
The counts for seropositive chickens were computed to determine the prevalence of seroposivity and for the prevalence "protective" antibodies. Positive antibodies were considered as titers greater than 4 and "protective" were considered as titers greater than 8 (Kemboi et al., 2013). The overall prevalence (positive samples) and "protective" antibodies (titers greater than 8) were compared between region, season, age, and sex. These samples were also strati ed by region and compared between season, age, and sex. The antibody titers were log 2 transformed for descriptive and t-test analysis using R. Risk ratios were calculated using the epitools package in R.

Results
Prevalence of NDV seropositivity Table 1 shows the overall prevalence of NDV antibodies and prevalence strati ed by season, age and sex. The overall prevalence of NDV in the backyard chickens of Tanzania is 26.8%. There were signi cant differences between titers in rainy versus dry season (15.9% and 34.9%, respectively; p-value < 0.0001) and between adult and non-adult birds (30.3% and 12.4%, respectively; p-value < 0.0001). However, there were no signi cant differences between the regions and sex of the birds (28.7% and 24.6%; 26.7% and 26.9%, respectively). Within each region there was a signi cant difference in the titers between seasons, and in Njombe the titers between the adult and non-adult birds were signi cant as well.
Violin plots representing the distribution of HI titers show a much larger count of negative samples than positive samples across all strata (Figure 2). Comparing strata with no signi cant differences between the samples tends to show similar distribution patterns, for example the distribution of antibody titers between male and female chickens is almost equal (Figure 2D and 2G). Comparing those strata with signi cant differences, the distribution patterns have many differences. For example in Figure 2C (p-value < 0.0001), there is a much larger distribution of negative samples in the non-adult and the distribution is skewed more toward high positives in the adults. This is true for the samples strati ed by season as well (p-value < 0.0001), with a much larger distribution of positive titers in dry than during the rainy season, although the overall titers are higher in rainy season. The patterns of higher positive titers become more apparent in Figure 3 when examining only the "protective" antibody titers. The overall prevalence of "protective" levels of antibody is 13 %. All strata seem to show the higher the titer the fewer amounts of chickens, demonstrating although the titer is considered "protective", the titer is still not high. Comparing the amount of "protective" antibody titers present between seasons and age exhibited a signi cance difference; however there was no signi cant difference between region and sex. When stratifying the data by region rst, the sample sizes become small, and although there are signi cant differences between adult and non-adult chickens in both Bahi and Njombe (p-values 0.04 and 0.007, respectively), the sample size of non-adult birds in each group is equal to 2, so larger samples sizes of non-adult birds is needed to con rm these results.

Risk Ratios, "protective antibody" characteristics
Factors contributing to the likelihood of having a "protective" antibody titer among local chickens determined through risk ratios with a calculated 95% con dence interval (Figure 4). Examining regionally, chickens are more likely to have a "protective" level of antibody titers in the Njombe (RR 1.  (Figure 4, red dots). However, there is no greater chance between males and females in having "protective" HI titers ( Figure 4, purple dots). The respective risk ratio were 0.98, 95%CI 0.93-1.04 and 1.02, 95% CI 0.96-1.08).

Discussion
Preceding studies have determined the prevalence of NDV in Tanzania to be 37.2-46.8%, respectively (Yongolo, 1996;Yongolo et al., 2002).This study estimated seroprevalence to be somewhat lower, around 26%. The lower seroprevalence could be accounted to seasonality while previous studies examined their prevalence throughout a 12-month period in 1995 and had peak prevalence occurring in October this speci c study the sampling was done in May and November. Nevertheless, there is consistent with our ndings demonstrating higher seroprevalence at the end of dry season, November, than rainy season, May. This study also demonstrates overall trends of signi cantly higher seroprevalence in adult versus non-adult backyard chickens (Table 1). These results are consistent with studies done in Kenya and Ethiopia that found higher seroprevalence in adult birds and during dry season. Also, the present study showed no signi cance difference between male and female, the ndings which differ from that reported by Nwanta et al. The overall, higher seroprevalence to NDV in dry season could be contributed to increased nutritional stress during dry versus rainy season. When there is less rain and drought conditions, water and food sources, such as insects and greens, become scarce for the scavenging chickens which leads to increased stress in additional to environmental stress such as higher temperatures during these times (Msami, 2002). Susceptibility to disease has been linked to nutritional de ciencies, such as Vitamin A and iron, which are especially critical to the proper development of the immune system in chickens (Klasing, 1998). Studies have also reported that Vitamin A supplements reduce NDV mortality by 36% in some instances (Opke et al., 2015). Due to the poorer nutritional status of chickens during dry season, this could be a major contributing factor to higher infection rates and possibly higher seroprevalence during dry season as compared to rainy season as demonstrated by our data. Interestingly, although results data analysis demonstrates higher seroprevalence and rate of protection in dry season, there are signi cantly higher overall and "protective" antibody titers in the rainy season ( Fig. 2 and Fig. 3). This higher overall titer could be contributed to the few outliers in the rainy season group (Fig. 2) that skew the data and drive the signi cance of higher titers in rainy season than dry season; however, this remains an area of interest for future studies to examine the seasonality and transmission dynamics of NDV infection in Tanzania.
The overall lower seroprevalence in non-adult chickens could be related to the shorter period of exposure of chicks to NDV compared to adults (McFerran and McCracken, 1988). This may also suggest that higher seroprevalence in adult means birds were exposed to NDV, possibly as chicks and then survived to adulthood.
These ndings also demonstrate that there is a greater chance of having a "protective" level of antibody titer if the chicken is from Njombe, during dry season, or is an adult, however, the chance of having "protective" antibody titers overall is low (Fig. 5). Although these "protective" levels of titers are present in the backyard chicken population, the prevalence is not high enough to infer any type of herd immunity to NDV. One de nition of herd immunity relates high numbers of immune individuals reducing the transmission of disease to others and ultimately eradicating the disease (John and Samuel, 2000). The low prevalence of "protection" among the backyard chickens in Tanzania demonstrates the need for future strategies to help smallholder farmers prevent disease in their ocks.
This data begins to provide insight on the disease dynamics of NDV in Tanzania, although many unanswered questions still remain. It is apparent that these chickens are exposed to NDV in the eld, and even produce high antibody titers, yet remain alive and healthy. Another area to be investigated and understood in greater detail is the underlying mechanisms contributing to the survival of the chickens in the face of NDV exposure. Future studies focusing on transmission, strain type, and monthly dynamics of NDV in backyard ocks will also provide greater insight into the disease dynamics and allow future implementation of strategies to alleviate the effects of NDV for the smallholder farmers of Sub-Saharan Africa to maximize productivity of their ocks as a source of food and contribute to the overall health of their families and communities.

Declarations
Not applicable.

Data sharing
All data are included in this article and Supplemental Table S1 Con ict of interest statement There is no con ict of interest.
Author's contribution EM and EB designed the study, collected samples and laboratory testing. EM did data analysis and wrote the manuscript. EB revised the content. Both authors read and approved the nal manuscript.  Overall HI titers for all samples. The log2(HI) titers for both positive and negative samples were included in these gures. Samples are strati ed A) regionally, B) seasonally, C) by age, and D) by sex. The samples were also strati ed by region then by E) season, F) age, and G) sex. The blue represents the boxplot and the grey represents the violin plot of the distribution of the titers. (*p-value < 0.001, **p-value < 0.0001). "protective" HI Titers. The log2 (HI) greater than 3, equating to "protective" levels of antibody titers, are represented in this gure. Samples are strati ed A) regionally, B) seasonally, C) by age, and D) by sex. The samples were also strati ed by region then by E) season, F) age, and G) sex. The blue represents the boxplot and the grey represents the violin plot of the distribution of the titers. (*p-value < 0.05, **p-value < 0.01).