A retrospective study (2007–2015) on brucellosis seropositivity in livestock in South Africa

Abstract In South Africa, brucellosis testing and record‐keeping are done by several laboratories, thus it is difficult to access any organized data to assess the status of the disease. This study evaluated the seropositivity for brucellosis using Rose Bengal test and complement fixation test in suspect cattle, sheep, goats and pigs sera submitted to Bacterial Serology Laboratory, Agricultural Research Council‐Onderstepoort Veterinary Research (ARC‐OVR) from nine provinces in the country during the period 2007–2015. This retrospective data analysis was conducted to estimate the occurrence of brucellosis in the country from the submitted samples, identify variables that affected seropositivity for brucellosis, investigate existing gaps in data recording and make recommendations on important variables to facilitate better data capture and inferences on brucellosis. Nine years of data were collated and analysed to detect association (seropositivity over time regarding animal species and location). Of the 764,276 animals tested, the distribution of samples was 90.50% (691,539/764,276), 5.19% (39,672/764,276), 3.92% (29,967/764,276) and 0.41% (3,098/764,276) for cattle, sheep, goats and pigs, respectively. The seropositivity for brucellosis by animal species was 6.31% (43,666/691,539, 95% CI: 6.26–6.37), 2.09% (828/39,672, 95% CI: 1.95–2.23), 0.63% (189/29,967, 95% CI: 0.55–0.73) and 0.13% (4/3,098, 95% CI: 0.05–0.33) in cattle, sheep, goats and pigs respectively. The data available did not capture information on the age, sex, breed and other host risk factors that would have been related to seropositivity for brucellosis. The data provide an understanding of the disease occurrence and confirm that brucellosis is enzootic in South Africa. Improved and standardized data collection can be used to pro‐actively drive, monitor, change or formulate policies to mitigate the challenges brought about by brucellosis in the livestock sector in South Africa.


| INTRODUC TI ON
Brucellosis is one of the most important and widespread zoonoses in the world (Kolar et al., 1995). Twelve Brucella species have been isolated of which B. abortus, B. melitensis and B. suis have been reported to affect livestock and humans while B. ovis affects only livestock (El-Sayed & Awad, 2018). In livestock, brucellosis is characterized by abortions and reproductive failure (Corbel & World Health Organization, 2006) but after abortion, the females can give birth again but continue to shed the pathogen (Lopes et al., 2010). Infections in humans cause additional losses (financial and disease burden), with prolonged clinical symptoms which could vary from months to years (Abdou, 2000;Corbel & World Health Organization, 2006;Zajtchuk & Bellamy, 1997). Brucellosis is an occupational disease and poses a risk to abattoir workers, employees in the meat packaging industry, veterinarians and farmers (Corbel & World Health Organization, 2006;Young, 1995).
In developed countries brucellosis is well controlled (Pappas et al., 2006) through routine domestic livestock surveillance, screening and animal vaccination programmes (Corbel, 1997;Maloney & Fraser, 2001). Brucellosis remains common in Africa, south and Central America, the Middle East, Asia, the Mediterranean basin and the Caribbean (Pappas et al., 2006). Brucellosis in the animal and human populations in South Africa date back to the early 19th century in various parts including Philippolis (Free State province), Steytlerville (Eastern Cape province) and Northern Cape province districts (Strachan, 1932;Van Drimmelen, 1949). In South Africa B. abortus infection was documented in 1913, when contagious abortion was observed to spread in cattle across the country and cases of 'camp fever' documented in humans (Thornton, 1936;Van Drimmelen, 1949). The fact that human cases of brucellosis were documented more than a century ago in South Africa could indicate that the disease had been circulating in the animal population. In the 1920s, only B. melitensis was isolated and suspected to be the cause of 'camp fever' or Malta fever in South Africa. Prior to these cases, it had been suspected in 1898 that goats may have been the source of suspected cases of 'camp fever' in 40 human patients around the Kimberley area (Northern Cape province) of South Africa (Strachan, 1932;Van Drimmelen, 1949). Brucellosis diagnosis had always been conducted with serological tests and bacteriological isolation (B. melitensis) from 1902 to 1911 (Strachan, 1932;Zammit, 1905), and these tests have yielded results from human blood (B. abortus) as well as from goat serum and milk samples in South Africa (Strachan, 1932). Later from 1956Later from to 1959 and B. melitensis were isolated from human blood samples as well (Schrire, 1962).
In South Africa brucellosis is a reportable and priority disease.
The control scheme is focused primarily to prevent the spread of bovine brucellosis and involves vaccination with B. abortus S19 and B. abortus RB 51 in cattle and B. melitensis Rev 1 in sheep, test and slaughter as well as prohibition of the movement of live animals from infected herds other than those for slaughter (OIE, 2016). The B. abortus S19 vaccine is used at the government recommended dose of 5 × 10 10 organisms on 4-8-month-old heifers (DAFF, 2016

| Study design
The study design was to acquire and collate diagnostic brucellosis data from samples collected from suspected cases of animal brucellosis tested at the Bacterial Serology Laboratory, ARC-OVR. Data abstraction was conducted by identifying complete and useable variables among available data set. Useable variables were filtered and curetted in Microsoft Excel 2007 version, and were stratified according to the year of testing, animal species, province and then related to the outcome of the serial testing conducted on the samples.

| Study area
South Africa is in the tropic of Capricorn in the southern hemisphere and the southernmost tip of the continent of Africa. The human population is estimated to be 58 million people with a surface area of 1,219,602 km 2 . The country has several distinct ecosystems and it is bounded by 2,798 km of coastline stretching along the South Atlantic and the Indian Oceans. In the north, its neighbouring countries are Namibia, Botswana, Zimbabwe and to the east and northwest are Mozambique and Swaziland. The country is divided into nine provinces and hosting the provincial laboratories and the Bacterial Serology Laboratory, ARC-OVR, which is in Gauteng province. The distribution of the number animals in the provinces (DAFF, 2018) during the study period is shown in Table S1.
The South African government announced that veterinary laboratories must be accredited to conduct the brucellosis serological

| Sampling
Records from 2007 to 2015 were retrieved from the Bacterial Serology Laboratory, ARC-OVR in South Africa. These data consisted of serological results of tests conducted on animal samples (serum) sent from farms, veterinary clinics, regional provincial laboratories and from the animal health officers from the nine provinces.
Suspicion of brucellosis was the primary criterion for the samples sent to the laboratory for testing and confirmation testing on the same animal could not be differentiated. Other samples sent for testing for brucellosis included animals from farms to confirm their brucellosis-free status and those destined to be exported to other countries.

| Laboratory tests data
Serial testing programme method was used to analyse all sera at ARC-OVR Laboratory (i.e. RBT positive followed by CFT). The RBT was used as a screening test for individual animals (Alton et al., 1988;OIE, 2009), while the CFT was used to confirm brucellosis in RBTpositive samples (Alton et al., 1988;OIE, 2009). Seropositivity for brucellosis was estimated based on samples seropositive on both RBT and CFT in series, and with titres of ≥1:30 for CFT. The correlation of the bacteriological and serological test results could not be investigated primarily because the two tests (serology and bacteriology) are conducted by different departments at ARC-OVR.

| Statistical analysis
The data were collated and managed in Microsoft Excel 2007 version and descriptive analysis was conducted using R (R Core Team, 2013).
The data were analysed based on the frequency of brucellosis seropositivity stratified by livestock species, province and year of testing. Analyses of measures of association of Brucella seropositivity with plausible risk factors and predictors (livestock species, year and provinces) were conducted using the Two X Two Table in OpenEpi ® (https://www.opene pi.com/Twoby Two/Twoby Two.htm).
For this purpose, the year '2007', ' KwaZulu-Natal Province' and 'pigs' were used as reference for comparison of risk within the categories.

| Demographic details
In our study we simply abstracted data available at the ARC-OVR Laboratory and no serological tests were conducted in the current study. The current study conducted data retrieval, filtration, curation and closely assessed the data to identify all the variables available

| D ISCUSS I ON
In South Africa, the existing surveillance system is weak in monitoring brucellosis in livestock on farms and/or abattoirs, or highrisk human population or from diagnostic laboratories records (DAFF, 2017;Padilla et al., 2010). Although Brucella species have been isolated from livestock (Caine et al., 2017;Kolo et al., 2019;Van Drimmelen, 1949, 1965 and humans (Schrire, 1962;Wojno et al., 2016), much is needed to be done to strengthen surveillance.
Low prevalence of brucellosis has been documented in previous studies in South Africa which include the 1.50% reported for cattle sampled at the Cato Ridge abattoir in Kwazulu-Natal province in 1984 (Bishop, 1984)  in Sub-Saharan Africa in livestock (Ducrotoy et al., 2017) were 1.00%-10.60% in cattle in countries where two serological tests were used in series (same criteria used in our study) were comparable to the 6.31% we detected. The analysis in this study identified the primary focus on bovine brucellosis as a gap in recording findings. Since the brucellosis scheme in South Africa mainly focuses on bovine, it was no surprise that the over-whelming majority of livestock tested was cattle, 90.50% (691,539/764,276).

Ref
The bovine brucellosis scheme is historic and based on the anecdotal belief that brucellosis is mainly a problem in cattle in the country. Since the brucellosis scheme in South Africa is biased towards the cattle population while sheep, goats and pigs have not received much attention, this could be mis-interpreted as though a higher seropositivity of brucellosis was detected in cattle compared to sheep, goats and pigs. The focus on bovine brucellosis could also be that cattle are considered economically productive based on meat and milk production as well as their export potential in South Africa compared to sheep, goats and pigs. It is relevant to mention that a limited level of control measures observed for cattle is being applied to some sheep and goat population in the country as B. melitensis outbreaks mainly identified with human brucellosis cases has been controlled in associated goat and sheep populations (Emslie & Nel, 2002;Kolar, 1987). The serological data of goats from ARC- Furthermore, it is impossible to ascertain the infecting Brucella species using serological tests, irrespective of the antigen (B. melitensis or B. abortus) or host species tested (Ariza, 1999;OIE, 2013aOIE, , 2013bSpink, 1956). This is because of the dominance and overlapping nature of the C epitope of smooth brucellae (Alonso-Urmeneta et al., 1998). This study emphasizes that bacteriological isolation is necessary to ascertain the infecting Brucella species and to understand the epidemiology when different host species are managed together or share grazing grounds and water sources.
However, identification and typing of Brucella species by conventional procedures are difficult and molecular methods are preferred for typing strains once these are isolated as described (Kolo et al., 2019;OIE, 2013a).
Based on the risk factor evaluation, it will appear that the restructuring of the testing scheme in 2010, from the period of announcement that 'Only the accredited laboratories can test for brucellosis', the odds ratios for Brucella seropositivity in livestock have increased, particularly in the sheep and cattle, and in the year 2010, for goats too. However, no significant increase in odds of seropositive sample has been noticed in pigs (Table 1). Whether this is due to (a). Increasing volume of export that mandated more samples to be It is of diagnostic relevance that the vaccination status of all the animals serologically tested (RBT and CFT) for brucellosis in the diagnostic laboratory is unavailable. More so, the live attenuated B.
abortus S19 vaccine used in the country has the potential to interfere with sero-diagnosis (OIE, 2009). However, the cattle tested were suspected clinical cases of brucellosis, most likely due to natural exposure to virulent Brucella field strains. Cattle that have been vaccinated with B. abortus S19 or sheep with B. melitensis Rev 1 between 3 and 6 months are usually considered to be infected if the sera give positive fixation at a titre of 30 or greater ICFTU/ ml when the animals are tested at an age of 18 months or older (OIE, 2013a). One way to avoid potential interference of vaccines in brucellosis sero-surveillance of testing is the recommended use of the rough B. abortus RB51 vaccine (Sowa et al., 1992), which do not induce antibodies detected in routine testing for brucellosis.
However, this vaccine is not supplied by the government and it is more expensive than B. abortus S19 vaccine but can be used for heifer or cow at any age. RB51 vaccine causing abortion has been reported but to a lesser extent than S19 (Palmer et al., 1996).
The failure of diagnostic laboratories to capture information on potential risk factors such as age, sex, breed, status of the herd from where the animals originate limits the application of the data for the country and veterinary services. Such data are imperative for the control and eradication of brucellosis as documented by others (Asante et al., 2019;Idrissi, 2014). Countries that have successfully eliminated or controlled brucellosis in livestock have achieved this by instituting control measures based on the demographic records of brucellosis as reported by others (Maloney & Fraser, 2001;Smirnova et al., 2013). The acquisition of important information on risk factors in animals tested for brucellosis at veterinary laboratories in the country will be invaluable when used in conjunction with the seropositivity data.
The finding that Gauteng province had the highest seropositivity for brucellosis 47.90% (21,389/44,687) (Table S3)   (iii) Sampling of the animals from the different provinces was biased because other than compulsory testing of dairy and stud cattle herds include mainly suspect cases of brucellosis that were tested, which may not be representative of cattle population in the country.

| LIMITATI ON S OF THIS S TUDY
(iv) Furthermore, the herd prevalence could not be determined based on the status of animals tested coupled with the lack of information on the herds of origin of the animals tested in the current study.

| CON CLUS IONS
In conclusion, the review and analysis of 9-year data on brucellosis from the Bacterial Serology Laboratory, ARC-OVR has demonstrated that brucellosis in livestock is endemic in South Africa.
Although the data lacked information on the vaccination status of the tested livestock, the fact that vaccinated animals are protected against brucellosis and thus will not exhibit clinical manifestation or require laboratory diagnosis, it is concluded that the seropositivity for the disease detected in our study reflects the natural exposure to Brucella spp. This study has provided baseline data that would lead to the improvement of informed policy to control the disease in the country. Although some vital information about the animals were missing, the information from our study can be used pro-actively to drive, monitor, change or formulate policies to mitigate the challenges brought about by brucellosis in the livestock sector of the country. This study identified the primary focus on bovine brucellosis, lack of linking brucellosis serological and bacteriological results, lack of centralized database and as mentioned the lack of vaccination status as gaps in the recording of data in South Africa.

| RECOMMENDATIONS
Considering that the data available at the Bacterial Serology Laboratory, ARC-OVR are grossly inadequate for analysis and drawing inferences on the epidemiology and important variables associated with brucellosis in livestock in South Africa, it is important that all samples from the provincial laboratories to the Bacterial Serology Laboratory, ARC-OVR should be accompanied by information such as animal sources, sex, age, breed and vaccination status of the animal.
To date in the country, vaccination of livestock against brucellosis with no stipulations on the types of vaccines allowed in the country, coupled by the fact that no compensation is paid to farmers of slaughtered brucellosis-positive livestock which may discourage their willingness to report the disease to the authorities. This practice therefore poses public health significance to farm/abattoir workers and consumers. It is therefore recommended that there is a need for standardized and target national Brucella program as well as the provision of resources for vaccination and indemnity.
At present in the country, with regard to brucellosis, animal health professionals including veterinarians and animal health assistants, are not trained or lack experience in disease tracebacks and programme implementation. There is therefore the need to train these personnel in testing for diseases such as brucellosis, collecting appropriate samples from suspect animals and to be proficient in disease prevention and control implementation.
Finally, brucellosis is a reportable disease in South Africa and our study has identified gaps, such as the lack of invaluable information on the livestock tested and the testing of only suspect cases and export livestock have the potential to contribute to the under-reporting of brucellosis in the country. It is therefore imperative to address these limitations to generate accurate data which will be essential for the development of an annual reporting and summary for program evaluation purposes.

ACK N OWLED G EM ENTS
The authors acknowledge Gauteng Department of Agriculture and Rural Development for funding this project. We also acknowledge our collaborators; Agricultural Research Council-Onderstepoort Veterinary Research (ARC-OVR), Faculty of Veterinary Sciences, University of Pretoria and all abattoirs that participated in the study.

CO N FLI C T O F I NTE R E S T
There is no conflict of interest with regards to this research.