1
CHAPTER I
GENERAL INTRODUCTION
1.1 Introduction and background
Brucellosis is a collective term that refers to the disease syndromes caused by bacteria of the genus Brucella, characterised by epizootic abortions, chronic endometritis, infertility, arthritis, orchitis or chronic infections in domestic animals. In humans the disease is characterised by septicaemia which manifests itself as recurrent fever, and localised chronic infections.
Brucella spp. are Gram-negative, coccobacillary, aerobic and facultative intracellular bacteria. There are nine distinct species which include: Brucella abortus (B. abortus), B. melitensis, B. suis, B. canis and B. ovis, B. neotomae, B. microti, B. ceti and B. pinnipedialis (Garritty et al., 2005; Foster et al., 2007; Scholz et al., 2008). Each of these brucellae have preferred natural hosts that include cattle (B. abortus); goats and to a lesser extent sheep (B. melitensis); pigs (B. suis); dogs (B. canis) and sheep (B. ovis) respectively (Quinn et al., 1999). B. ceti and B. pinnipedialis have cetaceans (whales and dolphins) and seals (pinnipeds) respectively, as their preferred natural hosts
(Garritty et al., 2005; Foster et al., 2007). B. neotomae that was originally isolated from a desert wood rat (Neotoma lepida) is believed to be non-pathogenic for cattle, sheep, goats and pigs (Garritty et al., 2005), while B. microti has been recently isolated from a vole, Microtus arvalis (Scholz et al., 2008).
2
The most important brucelloses in domestic animals are caused by B. abortus, B. suis and B. melitensis (Blood and Radostits, 1989). Four species, B. abortus, B. suis, B. melitensis and B. canis are associated with systemic disease in humans, and are thus recognized as zoonotic pathogens of public health significance (Quinn et al., 1999).
Brucellosis in humans is an occupational disease for shepherds, veterinarians, farmers, abattoir workers and laboratory personnel who handle infected animals and contaminated animal products (Moyer and Holocomb, 2005). In addition, B. abortus and B. melitensis are important food-borne pathogens that may be acquired by consuming raw milk and milk products such as soft cheese (Leclerc et al., 2002;
Kuplulu and Sarimehmetoglu, 2004). B. melitensis is the most virulent species for humans and accounts for the majority of cases of human brucellosis (Leclerc et al.,
2002).
Among the animal brucelloses, bovine brucellosis, caused by B. abortus biovars is the most important disease in many countries around the world due to its economic importance (McDermott and Arimi, 2002; Nicoletti, 1980; Silva et al., 2000; Taleski et al., 2002). The disease is characterised by abortion “storms” in pregnant cattle (Blood and Radostits, 1989) especially in naïve herds that are intensively managed (Bishop et al., 1994) and is thus a major contributor to low calf crop in cattle farming.
Occasionally, in cows, there is a decrease in milk yield and low reproductive efficiency due to high incidences of retained placenta (Blood and Radostits, 1989; Walker, 1999).
In affected bulls, orchitis and epidydimitis are typical presentations of the disease
(Bishop et al., 1994) with subsequent decline in fertility even though libido is retained
(Kumi-Diaka et al., 1980). Other causes of economic loss in areas where brucellosis is 3 endemic are due to the direct costs incurred by farmers when trying to control the disease (Nicoletti, 1980).
The prevalence and incidence of bovine brucellosis vary considerably between herds, areas and countries (McDermott and Arimi, 2002). The disease has been reported to occur in most countries in Africa (Chukwu, 1985; Faye et al., 2005) including the sub-
Saharan region (McDermott and Arimi, 2002). The disease has been reported to be endemic in some farming areas in Zimbabwe (Mohan et al., 1996). Together with bovine genital campylobacteriosis, brucellosis was established to be the most important cause of abortion and infertility in cattle in Zimbabwe during the 1970s (Swanepoel et al., 1975). The disease was originally perceived to be more important in dairy than in beef cattle (Manley, 1969), but was later established to be equally important in both cattle farming sectors (Swanepoel et al., 1976). The prevalence of bovine brucellosis is believed to be low in communal areas due to the extensive nature of cattle rearing although only very few studies have been conducted in this sector (Madsen, 1989).
However, the establishment of smallholder dairy schemes in communal areas using cattle purchased from commercial sector (Madsen, 1989), coupled with the subsequent restocking of communal cattle through purchases of commercial cattle has a possible influence on the dynamics of bovine brucellosis in communal areas. It is noteworthy that in most of these communal areas, mixing of cattle from different villages is likely due to sharing of grazing areas, water points and communal dip tanks used for control of ticks and tick borne-diseases. The practice of mixing cattle herds been reported to be an important risk factor for transmission of Brucella spp. (McDermott and Arimi,
2002).
4
Zimbabwe is prone to periodic droughts as a result of ephemeral and erratic rainfall patterns. One of the most severe droughts to affect the country occurred during the
1991/92 season and resulted in increased cattle mortalities in smallholder farms in communal areas mainly due to inadequate grazing pasture, water and livestock diseases.
This loss of livestock severely affected these farmers whose livelihood is heavily dependent on cattle farming. Cattle are a vital source of milk, meat, income, drought power for land tillage, transport and manure, in addition to their use in numerous social or cultural roles. Therefore, there is a need to enhance the control of important diseases of cattle that include brucellosis, anthrax, foot and mouth disease (FMD), and tick- borne diseases in order to ultimately improve food security to these resource-limited farmers. Brucellosis, bovine tuberculosis, anthrax and other zoonoses acquired through direct contact with animals or from consuming beef and milk are increasingly becoming diseases of public health concern especially in people who are immuno-compromised by HIV/AIDS, prevalent in the Southern African sub-region.
After the drought in 1991/92, the government of Zimbabwe, through the Agricultural
Rural and Development Authority (ARDA) and in collaboration with other organizations such as the Heifer Project International and Initiatives for Development of
Equity in African Agriculture, assisted the smallholder farmers to restock their herds.
This was a drought mitigation measure aimed at ensuring food security to communal farmers. During this period cattle of the Bos taurus breeds (both dairy and beef) were purchased from commercial farms and mixed with the Bos indicus in the communal areas. While some efforts were made to purchase animals only from brucellosis-free herds, it was not always feasible to establish the status of these animals before 5 translocation. This introduction of new animals to naïve herds was envisaged to be an important risk factor for transmission of Brucella spp. (McDermott and Arimi, 2002).
Since the last national brucellosis survey was conducted in 1987 (Madsen, 1989), it was imperative that new data be generated to establish the current status of bovine brucellosis in some of the areas that restocked their herds.
1.2 Hypothesis
The restocking of cattle herds is a risk factor for brucellosis in smallholder dairy farms in Zimbabwe
6
1.3 Aims and Objectives of the study
It was previously stated that the prevalence of brucellosis among cattle herds in communal areas in Zimbabwe is generally low, averaging about 3.0 % (Madsen, 1989).
This statement and the figure could not be relied upon because of the continuous stocking and re-stocking of cattle which have great impact on the prevalence of brucellosis. Thus, this project was aimed at investigating the epizootology and diagnostic approaches towards brucellosis in dairy cattle reared in smallholder farms in
Zimbabwe.
The study was formulated to investigate the following specific objectives:
1.3. 1. To establish the prevalence of antibodies against Brucella spp. in individual
cattle from smallholder farming areas;
1.3.2 To identify risk factors for Brucella spp. infection in individual cattle and herds
from smallholder farming areas;
1.3.3 To establish the prevalence and risk factors for abortions in dairy cattle from
smallholder farming areas;
1.3.4 To evaluate the different serological tests used for the diagnosis of brucellosis
under the Zimbabwean conditions; and,
1.3.5 To investigate the smallholder dairy herds by culture and isolation of Brucella
spp.
1.3.6 To characterize the Brucella spp from Zimbabwe using biochemical profiles and
polymerase chain reaction (PCR) technique 7
CHAPTER II REVIEW OF LITERATURE
2.1 General
Brucella spp. are small gram-negative aerobic bacteria that appear as short rods or coccobacilli, measuring 0.6 to 1.5µm x 0.5 to 0.7µm in size (Walker, 1999). They do not produce spores and have no flagella or true capsules (Holt et al., 1994) but instead have a rudimentary capsule-like envelope that has been demonstrated by electron microscopy for B. abortus and B. melitensis (Corbel and Brinley-Morgan, 1984;
Walker, 1999). Brucella cultures exist either as smooth, non-smooth variants of smooth cells or non-smooth (rough), but cultures are commonly designated either as smooth or rough on original identification (Corbel and Brinley-Morgan, 1984; Alton et al., 1988).
Although the structure of their cell wall is not completely elucidated, gross analyses have indicated in different proportions, the presence of proteins, carbohydrates, lipids, muramic acid and 2-keto-3-deoxyoctulsonic acid between the smooth and non-smooth species (Garritty et al., 2005). The external layer of the outer membrane of the cell wall comprises mainly lipopolysaccharides (LPS) interspersed with a variety of proteins and lipids (Garritty et al., 2005). Like other gram-negative bacteria, dominant surface antigents in both smooth and non-smooth strains are linked to the lipopolysaccharides
(Corbel and Brinley-Morgan, 1984; Walker, 1999). The smooth species; B. abortus, B. melitensis and B. suis possess two important surface antigens named A and M, which are present on these lipopolysaccharides, albeit in varying proportions among the species (Quinn et al., 1999; Walker, 1999; Garritty et al., 2005). The cultures of these
Brucella spp. agglutinate with absorbed monospecific A or M antisera (Alton et al., 8
1988). Only the permanently rough species, B. ovis and B. canis agglutinate with the R, anti-rough, monospecific serum (Quinn et al., 1999).
Beneath the cell wall is the periplasmic space, whose exact function is not clear, but believed to be the site of a variety of hydrolytic enzymes (Garritty et al., 2005).
Below the periplasmic space, the cytoplasmic membrane whose typical triple-layered lipoprotein structure resembles that of other gram-negative bacteria encloses the cytoplasm (Corbel and Brinley-Morgan, 1984). The Brucella cytoplasm, where the osmiophobic nuclear material is located, is homogenous and is interspersed with small vacuoles and polysaccharide-containing granules (Garritty et al., 2005).
2.2 Taxonomy and evolution of the family Brucellaceae
Nine conventional species of the brucellae are recognised, viz; Brucella abortus (B. abortus), B. melitensis, B. suis, B. canis, B. ovis, B. neotomae, B. ceti, B. pinnipedialis and B. microti (Alton et al., 1988; Foster et al., 2007; Scholz et al., 2008). In previous work, B. ceti and B. pinnipedialis were proposed to be B. cetaceae and B. pinnipediae respectively, but have recently been renamed (Cloeckaert et al., 2001; Foster et al.,
2007). The phylogenetic tree of the Brucella spp. is presented in Fig. 2.1. 9
Fig.2.1. Phylogenetic tree of the Brucella spp. (Garritty et al., 2005).
Recent studies based on the sequencing of the 16S rRNA gene, have delineated the family Brucellaceae that includes pathogens and soil organisms, only into three genera;
Brucella (type genus), Mycoplana and Ochrobactum (Garritty et al., 2005). On the basis of DNA sequences, the genus Brucella appears to be more closely related to
Ochrobactrum than Mycoplana (Garritty et al., 2005). While the genus Brucella contains facultatively intracellular and important pathogens of animals and humans, members of the genus Mycoplana are found in the soil as saprophytes (Garritty et al.,
2005). Some species of the genus Ochrobactrum survive as saprophytes in soil and a few have been isolated from human clinical specimens (Garritty et al., 2005). The veterinary significance of the genus Mycoplana and Ochrobactrum is questionable.
Molecular genetic studies have clearly demonstrated that the genus Brucella forms a discrete homogenous group which clearly belongs to the Alphaproteobacteria (Garritty et al., 2005). However, the use of 5S and 16S rRNA-DNA hybridisation (De Ley et al., 10
1987), and phenotypic tests such as multi-locus enzyme electrophoresis, antigenic relatedness and lipid composition, have revealed some genetic relationship with plant pathogens and symbionts such as Agrobacterium, Rhizobium and Phyllobacterium and distant relationship to intracellular animal parasites, Bartonella and Rickettsia and free- living bacteria like Caulobacter ( Moreno et al., 1990; 2002; Garritty et al., 2005).
2.3 Pathogenicity of Brucella spp.
Brucella spp are obligate intracellular parasites and each of the distinct species tends to have a preferred natural host (Quinn et al., 1999) as outlined in Table 2.1. These microorganisms are responsible for the several infectious conditions that result in placentitis and abortion in pregnant female animals, epididymitis and orchitis in male animals as well as localised chronic conditions such as hygroma, arthritis and bursitis.
Four of these species: B. melitensis, B. abortus, B. suis and B. canis, are infectious for humans, and are recognized as important zoonotic pathogens of public health concern.
They cause a disease syndrome that is characterised by the occurrence of recurrent fever
(Malta or Mediterranean fever) and non-specific focal conditions such as arthritis, orchitis and diskospondylitis (Quinn et al., 1999).
11
Table 2.1. Hosts, diseases and distribution of Brucella spp. (Quinn et al., 1999).
Species Host (s) Diseases Geographical distribution B. abortus Cattle Abortion and orchitis Biotypes: 1: Worldwide (common) Sheep, goats and pigs Sporadic abortion 2: Worldwide (not common) 3: India, Egypt, East Africa Horses Associated with bursitis (poll evil 5: Britain and Germany and fistulous withers) Other biotype are infrequently isolated Humans Undulant fever B. melitensis Goats, sheep Abortion Many sheep and goat rearing regions except New Zealand, Australia and North America Cattle Ocassional abortion and excretion in milk Humans Malta fever B. suis Pigs Abortion, orchitis, arthritis, Biotypes: spondylitis and herd infertility 1: Worldwide 2: Western and Central Europe Humans Undulant fever 3: USA, Argentina and Singapore 4:: Arctic Circle( Canada, Alaska and Siberia) in reindeer and caribou B. ovis Sheep Epididymitis in rams and sporadic New Zealand, Australia and some other sheep- abortion in ewes raising countries: USA, Romania, Czech Republic, South Africa, South America B. canis Dogs Abortion, epididymitis, North America and parts of Europe. Becoming discospondylitis and permanent worldwide but not common infertility in males Humans Undulant fever B. neotomae Desert wood rat Non-pathogenic for the wood rat USA (Utah) (Neotoma lepida) and has not been recovered from other animal species B. pinnipedialis Pinnipeds (PINNIPEDIA): ? localised and systemic infections Northen and Western Europe, Canada and Seals, Sea lions and ? Abortion USA Walruses B. ceti Cetaceans ? localised and systemic infections Northen and Western Europe, Canada and (ODONTOCETI): Whales, ? Abortion USA dolphins, porpoises B. microti Voles (Microtus arvalis) ? localised and systemic infections Europe
* Natural host given in bold print, ? = Disease status not certain 12
2.4 Identification of Brucella spp.
2.4.1 Isolation
Brucella spp. are chemo-organotrophic microorganisms requiring complex media containing several amino acids, thiamin, biotin, nicotinamide and magnesium salts, while X (haemin) and V (nicotinamide adenine dinucleotide [NAD]) factors are not required (Alton et al., 1988; Holt et al., 1994). Growth is inhibited on media containing bile salts, tellurite or selenite (Alton et al., 1988).
The growth of brucellae in simple nutrient liquid medium is usually poor unless these are supplemented with blood, serum or tissue extracts (Moyer and Holocomb, 2005).
Brucella spp. are more fastidious than other aerobes and hence their growth in liquid medium is generally poor, unless the medium is vigorously agitated to improve aeration
(Alton et al., 1988).
Most Brucella spp. can be isolated in unsupplemented, enriched peptone based media, or blood agar (Alton et al., 1988; Quinn et al., 1999). Good growth is obtained on
Brucella medium base (Oxoid), sucrose dextrose agar (Oxoid), tryptone soy agar or glycerol dextrose agar (Oxoid) supplemented with 5% bovine or horse serum (OIE,
2004; Moyer and Holocomb, 2005). Cultures of brucellae are usually established directly on solid media as it allows colonies to be isolated and recognised clearly, but liquid media may be used either for enrichment or for voluminous specimens (OIE,
2004). A non-selective biphasic Castaneda’s medium is recommended for the isolation of Brucella spp. from blood or other body fluids or milk where enrichment culture is advisable (OIE, 2004). 13
Brucella spp. are slow growing and the use of selective media is recommended for primary isolation from most clinical specimens because of the high numbers of overgrowing contaminants (Marin et al., 1996). Such selective media are prepared by incorporating antibiotics and bacteriostatic dyes onto basic enriched media such as
Brucella medium base (Oxoid). An example is Farrell’s medium (Oxoid), prepared by adding six antibiotics; bacitracin, vancomycin, nalidixic acid, polymixin B, nystatin and cycloheximide onto sucrose dextrose agar for the isolation of B. abortus from contaminated milk samples (Farrell, 1974). Farrell’s medium was found not to be an ideal medium for the isolation of B. melitensis, because the concentration of nalidixic acid and vancomycin in this medium have inhibitory effects on some strains (Marin et al., 1996). Therefore, the use of modified Thayer-Martin’medium supplemented with haemoglobulin (10g/l), colistin methanesulphonate, vancomycin, nitrofurantoin, nystatin and amphotericin B in tandem with Farrell’s medium is believed to enhance the chances of isolating B. melitensis (OIE, 2004). Recently, due to its carcinogenicity, cycloheximide has been removed from the Brucella selective supplements used in the
Farrell’s medium (Anon., 2005).
These antibiotic supplements of the Farrell’s medium are commonly used, in different combinations and proportions onto any one of the basal media such as Brucella medium base (Oxoid), Tryptone soya agar (Oxoid), Serum dextrose agar (Oxoid), Columbia blood agar (BioMerieux) and other medium bases, for the formulation of selective media for isolation of Brucella spp. Other types of selective media have at some stage been used in the isolation of Brucella spp. Selective BCYE (polymyxin, anisomycin, cefamandole) is commercially available (Raad et al., 1990; Moyer and Holocomb,
2005). Moyer and Holocomb (Moyer and Holocomb, 2005) reported the use of 14 chocolate agar containing selective supplements (BBL Laboratories) for the isolation of
Brucella spp. Terzolo et al. (Terzolo et al., 1991), used Skirrow’s agar to isolate B. abortus, B. suis, B. melitensis, B. canis and B. ovis from contaminated vaginal exudates and milk. Hornsby et al. (Hornsby et al., 2000) also found Skirrows agar, together with
Modified Kuzdas medium and Tryptone soya agar (TSA) suitable for the recovery of the vaccine strain B. abortus RB 51, while Farrell’s, Ewalt’s and Kuzdas and Morse’s medium were not suitable. Similarly, the use of new media such as rifampin Brucella medium and malachite Brucella medium (MBM), together with TSA, was found to enhance the recovery of B. abortus RB 51 (Hornsby et al., 2000).
For the isolation of Brucella spp. from milk samples, although solid media have been used successfully (Farrell, 1974), the use of enrichment media such as serum dextrose, tryptone soy or Brucella broth containing selective supplements of at least amphotericin
B and vancomycin should be used because the microorganisms are usually present in too low numbers to be detected on solid media (OIE, 2004).
2.4.2 Growth characteristics
Brucella spp. have an aerobic type of metabolism, using the cytochrome-based electron transport system with oxygen or nitrate as terminal electron acceptor (Holt et al., 1994).
On primary isolation, many strains would require supplementation with 5-10% CO2
(Alton et al., 1988; Holt et al., 1994). Although the growth of Brucella spp. may occur between 20 ºC and 40 ºC, growth occurs optimally at 37 ºC (Holt et al., 1994).
In static broth medium, when Brucella spp. are incubated at 37 ºC for seven days, smooth strains produce moderately uniform turbidity with pale powdery deposit (Alton et al., 1988). Non-smooth strains may produce granule deposit, variable turbidity and 15
pellicle formation (Alton et al., 1988). In semi-solid media, CO2-dependent strains produce a disc of growth a few millimetres below the surface, whereas CO2- independent strains produce uniform turbidity from the surface down to a depth of a few millimetres (Alton et al., 1988). On solid media, growth is not apparent until about 3 to
5 days of incubation (Quinn et al., 1999). Growth on selective media may be delayed by several days and some strains may not produce discernible colonies until about 14 days of growth (Alton et al., 1988; Walker, 1999). Brucella spp. produce colonies that are round, glistening, pin-point, 1-2 millimetres in diameter, with smooth margins (Alton et al., 1988; Quinn et al., 1999). Later they become larger and darker but remain clear
(Alton et al., 1988). On serum dextrose agar, or any other clear medium, when examined under low power microscope, Brucella colonies have a raised surface, translucent with entire margins and displaying a characteristic pale “honey drop-like” appearance (Corbel and Brinley-Morgan, 1984). On sheep blood agar, smooth strains of
Brucella spp. produce small, glistening, smooth and non-haemolytic colonies which become opaque with age (Quinn et al., 1999). Some strains of B. suis, may over time produce large colonies which appear mucoid. The colonies of non-smooth strains are dull, yellowish, opaque and when touched with an inoculation loop are found to be brittle (Quinn et al., 1999).
Brucella spp. may or may not grow on MacConkey agar depending on the fastidiousness of the strains (Alton et al., 1988). The more vigorous strains of B. abortus, B. melitensis and B. suis will grow on MacConkey agar, producing small, non- lactose fermenting colonies (Corbel and Brinley-Morgan, 1984).
16
2.4.3 Microscopic appearance
Brucella spp. are observed as gram negative cocci, coccobacilli or short rods. They are usually arranged individually and less frequently in pairs, short chains or small groups and do not usually exhibit bipolar staining ( Holt et al., 1994; Garritty et al., 2005).
2.4.4 Biochemical reactions
On the basis of biochemical tests, members of the genus Brucella are broadly defined as catalase positive, oxidase positive (except B. ovis), urease positive (except B. ovis and
B. neotomae), reduce nitrates to nitrites and do not exhibit motility in semi-solid media
(Alton et al., 1988; Quinn et al., 1999). Brucella spp., with the exception of B. neotomae, do not produce acid from carbohydrates in conventional peptone media (Holt et al., 1994). In addition, they do not produce indole, gelatinases, haemolysins, acetyl methyl carbinol (Voges Proskauer test), formic and acetic acids from glucose (Methyl red test) (Holt et al., 1994). A summary of the biochemical tests used to identify
Brucella spp. is given in Tables 2.2 and 2.3. The details of the scheme for identification of members of the genus Brucella and their differentiation from closely related bacteria are provided elsewhere (Barrow and Felthman, 1993; Holt et al., 1994; Garritty et al.,
2005).
2.4.5 Biotyping
The differences in biochemical and oxidative metabolic profiles (Meyer and Cameron,
1961), lysis by different bacteriophages and serological reactions with monospecific antisera, have enabled the subdivision of Brucella spp. into several biovars. B. abortus has been divided into seven biovars (1, 2, 3, 4, 5, 6 and 9) (Alton et al., 1988; OIE,
2004). B. abortus biovar 7 and 8 were dropped because they were reported to be mixed cultures (Alton et al., 1988; Subcommittee, 1988). B. melitensis has been divided into 17 three biovars (1, 2 and 3), while B. suis has five biovars (1, 2, 3, 4 and 5) (Alton et al.,
1988). While B. neotomae, B. ovis and B. canis are considered to be monovars, not much is known about marine Brucella isolates (Garritty et al., 2005). Details on biotyping are found in Table 2.4.
Table 2.2. Differential characteristics of Brucella spp. from some other gram negative bacteria (Alton et al., 1988).
Tests Brucella Bordetella Campylobacter Moraxella Acinetobacter Yersinia bronciseptica foetus species species enterocolitica
Morphology Small Small Coma-shaped diplococcoid diplococcoid Rods cocco- cocco-bacilli bacilli Motility at 37 - + + - - - ○C Motility at 20 - - - - - + ○C Lactose - - - va v - fermentation on MacConkey agar Acid -b - - - v + production on agar containing glucose Haemolysis - + - v v - on blood agar Catalase + + + v + +
Oxidase +c + + + - -
Urease +d + - v v +
Nitrate +e + + v - + reduction Citrate - + - - v - utilisation a Positive and negative species within the genus; b B. neotomae may show some fermentation; c Except B. ovis, B. neotomae and occasional B. abortus strains which are negative; d Except B. ovis and occasional B. abortus strains which are negative; e Except B. ovis which does not reduce nitrates to nitrites 18
Table 2.3. Differentiation of the species and biovars of the genus Brucellaa (Garritty et al., 2005; OIE, 2004).
Characteristics B. melitensis B. abortus biovars B. suis biovars B. B. B. biovars ovis neotomae canis 1 2 3 1 2 3 4 5 6b 7 9 1 2 3 4 5 Catalase + + + + + + + + + + + + + + + + + + + Oxidase + + + + + + + + + + + +e + + + + - - + Urease + + + + + +f + + + + + + + + + + - + + CO2 requirement - - - [+] [+] [+] [+] ------+ - -
H2S production - - - + + + + - [-] [+] + + - - - - - + - Growth on media containingc: Thionin + + + - - + - + + + + + + + + + + -d + Basic fuchsin + + + + - + + + + + + [-] - + [-] - [-] - [-] Agglutination with monospecific antisera A - + + + + + - - + + - + + + + - - + - M + - + - - - + + - + + - - - + + - - - R ------+ - +
aSymbols: +, positive; [+], positive for most strains, [-], negative for most strains, -, negative all strains. bFor more certain differentiation of biovar 3 and 6, thionine at 1:25, 000 (w/v) is used; biovar 3 gives a positive growth response, biovar 6 is negative. cDye concentration, 1:50, 000 (w/v). dGrowth will occur in the presence of thionine at a concentration of 1:150, 000 (w/v). eRapid reaction, most strains of B. suis test positive within 5 minutes fSome field strains of B. abortus may be negative 19
Table 2.4. Differentiation of the species of the genus Brucellaa (Garritty et al., 2005).
Characteristic B. B. B. suis biovar B. B. B.
melitensis abortus canis neotomae ovis
1 2 3 4 5
Lysis by phage at RTD:
Tb NL L NL NL NL NL NL NL PL NL
Wb NL L L L L L L NL L NL
Fi NL L PL PL PL L L NL L NL
BK2 L L L L L L L NL L NL
R/O NL PL NL NL NL NL NL NL NL L
R/C NL NL NL NL NL NL NL L NL L
Oxidation of substrates
L-alanine + + d - d - - d d d
L-Asparagine + + - d - - + - + +
L-Glutamic acid + + - d d d + + + +
L-Arabinose - + + + - - - d + -
D-Galactose - + d d - - - d + -
D-Ribose - + + + + + + + d -
L-Glucose + + + + + + + + + -
D-Xylose - d + + + + + - - -
L-Arginine - - + + + + + + - -
DL-Citrulline - - + + + + + + - -
DL-Ornithine - - + + + + + + - -
L-Lysine - - + + + + + - -
Meso-Erythritol + + + + + + + + -
RTD = Routine test dilution
Symbols: +, positive, -, negative; d, doubtful; NL, no lysis; PL, partial lysis
20
2.4.6 Molecular typing
Molecular typing of brucellae has been attempted using DNA-DNA or DNA-RNA hybridisation methods, polymerase chain reaction (PCR) based methods such as the repetitive extragenic palindromic PCR (REP-PCR) and the enterobacterial intergenic consensus sequences PCR (ERIC-PCR) (Mercier et al., 1996), the arbitrarily primed
PCR (AP-PCR) (Fekete et al., 1992) and the restriction fragment length polymorphism
PCR (RFLP-PCR) (Cloeckaert et al., 2001). These PCR based methods have been reviewed in detail by Bricker (Bricker, 2002). Recently, using the variable number of tandem repeats (VNTR) analysis, Bricker et al. (Bricker et al., 2003) found the technique to be the most discriminatory for Brucella spp. However, using this technique, Brucella spp. have been found to be highly homogenous. Other studies of the genome of Brucella spp. have demonstrated the existence of more than 90% homology
(Clavareau et al., 1998; Cloeckaert et al., 2001) and based on DNA-DNA hybridisation, a single species of B. melitensis has been proposed, with the other species being biovars
(e.g. B. melitensis biovar abortus) (Verger et al., 1985). This genomic similarity makes the differentiation of Brucella spp. difficult, and often a study of biological and physiological characteristics is required (Alton et al., 1988). The debate of whether the brucellae should comprise a single genospecies or multiple species has been a source of much controversy (Cutler et al., 2005). But the recent reappraisal of the Brucella spp. by review of their population structure and analysis of their genetic diversity by methods other than DNA-DNA hybridisation (Moreno et al., 2002) has reasserted the return to the pre-1986 taxonomy where the multiple species and biovars concept is used
(Osterman and Moriyon, 2003; Foster et al., 2007). The recent discovery of new
Brucella spp. from marine mammals has given further support for use of this nomenclature (Cloeckaert et al., 2001).
21
2.5 Domestic Animal Brucelloses
2.5.1 Introduction
Brucellosis is a collective term that refers to the disease syndromes caused by Gram- negative bacteria of the genus Brucella, characterised by epizootic abortions, chronic endometritis, infertility, arthritis, orchitis or chronic infections in domestic animals.
Brucellosis in cattle is an important zoonotic disease that has existed since antiquity
(Cutler, et al., 2005) and its economic importance in cattle farming in many countries of the world is well known. The disease was not described in detail until Bang in 1897 established that B. abortus was the cause of abortion in cattle (Nielsen, 2002). It is perhaps the most widespread and economically important of the zoonotic diseases in tropical and subtropical regions (Nicoletti, 1980).
In goats and sheep, brucellosis is primarily caused by B. melitensis is characterised by abortion, retained placenta, orchitis, epididymitis and rarely arthritis (OIE, 2004). B. melitensis is an important zoonotic infection causing undulant fever (also called Malta fever) in humans (Herr, 1994).
B. ovis is one of the most common causes of epididymitis and infertility in rams and a rare cause of abortion in ewes (Van Tonder et al., 1994). B. ovis infections are only important in sheep and not in goats.
The infection with B. suis in pigs, causes an acute or chronic disease that is characterised by abortions, stillbirths, heavy mortality in piglets, sterility in sows, and
22 orchitis in boars (Bishop and Bosman, 1994; Bishop et al., 1994; Blood and Radostits,
1989).
In dogs, brucellosis caused by B. canis is, characterised by abortion storms in females and testicular atrophy, epididymitis and infertility in males and generalised lymphadenitis in both sexes (Carmichael and Kenney, 1968; Oncel, 2005).
2.5.2 Aetiology
2.5.2.1 Bovine brucellosis (Bang’s Disease)
Bovine brucellosis is primarily caused by biovars of B. abortus and occasionally by B. melitensis in cattle kept closely together with goats and sheep (OIE, 2004). Although B. suis has been reported to cause mammary infection, it has not been associated with abortion in cattle (Ewalt et al., 1997). Worldwide, the majority of cases of brucellosis in cattle are attributable to B. abortus biovar 1 (Bishop et al., 1994; Mohan et al., 1996;
Quinn et al., 1999). B. abortus biotype 2 has a world wide distribution, but considered to be less frequent than biotype 1(Quinn et al., 1999). B. abortus biotype 3 has been reported from East Africa, Egypt and India; B. abortus biotype 5 has been isolated in
Germany and Britain; while the other biotypes are infrequently isolated (Quinn et al.,
1999). There are no proven differences in the pathogenicity of field strain biovars
(Nicoletti, 1980).
While the colonisation of the pregnant uterus in cows by the B. abortus strain 19 vaccine is unlikely due to the inhibition of growth by erythritol that is abundant in this organ, some erythritol-tolerant variants of B. abortus strain 19 have been reported to
23 cause abortions (Beckett and MacDiarmid, 1985). However, the persistence of B. abortus strain 19 vaccine-induced abortions in nature is not known.
2.5.2.2 Caprine and Ovine brucllosis
Caprine and ovine brucellosis is caused by B. melitensis biovars 1, 2 or 3 (OIE, 2004).
B. melitensis, the most pathogenic of all the Brucella spp, is morphologically indistinguishable from B. abortus but can be identified using serological methods
(Alton et al., 1988). Smooth cultures are usually exhibited and these are more pathogenic for laboratory animals than the rough mutant cultures (Alton et al., 1988;
Garritty et al., 2005).
2.5.2.3 Ovine brucellosis (Ovine Epididymitis)
The causative agent of ovine epididymitis has been established to be B. ovis (Buddle,
1956). The agent is morphologically similar to the other members of the genus, except that it stains blue with modified Koster’s stain, in contrast to the other Brucella spp. which stain pink (Garritty et al., 2005). Its cultures exist only in rough colonial phase, do not agglutinate with monospecific antisera for A and M surface antigens, but are agglutinated by antisera for the rough (R) surface antigen (Garritty et al., 2005).
2.5.2.4 Porcine brucellosis
Porcine brucellosis is caused by biotypes of B. suis which are morphologically similar to other Brucella spp. (Garritty et al., 2005).
24
2.5.2.5 Canine brucellosis
Canine brucellosis is caused by B. canis (Carmichael, 1966). The morphological characteristics of B. canis are similar to the other members of the genus, Brucella.
Similar to B. ovis, its cultures exist in non-smooth (rough) colonial phase and are agglutinated only by antisera for the R surface antigen (Garritty et al., 2005).
Occasionally, canine brucelloses caused by B. abortus, B. melitensis and B. suis have been reported (Shin and Carmichael, 1999; Wanke, 2004).
2.5.2.6 Equine brucellosis
Horses can be infected with B. abortus and B. suis, which are commonly associated with brucelloosis in this animal species (Quinn et al., 1999, Walker, 1999).
2.5.3 Epizootology
2.5.3.1 Bovine brucellosis
Bovine brucellosis is widespread and endemic in most countries in the world, especially where disease control is lacking. However, most parts of Northern and Central Europe,
Australia, New Zealand and Japan are believed to be free from the disease (OIE, 2004).
In these countries, the disease was eradicated through implementation of stringent disease control strategies that included test and slaughter policies.
Bovine brucellosis is reported to occur in most countries in Africa (Chukwu, 1985; Faye et al., 2005). The prevalence of the disease varies between countries, regions and farming sectors due to vast differences in terrain, climate, social customs, resources, livestock management and attitude towards disease control as shown in Table 2.5
(Nicoletti, 1984; McDermott and Arimi, 2002). In Southern Africa, the disease has been
25 documented to occur in Angola, Botswana, Lesotho, Mozambique, Namibia, South
Africa, Swaziland and Zambia, while Malawi and Madagascar could be free from the disease (Bedard et al., 1993; Bishop et al., 1994; Muma et al., 2006).
26
Table 2.5. Distribution of bovine brucellosis prevalence by country and production system
Production system Country Serological test % brucellosis applied prevalence (samples tested) 1. Pastoral/arid and semi-arid regions (Faye et al., 2005) Uganda RBPT 15.8 (10 529) (McDermott et al., 1987) Sudan SAT, CFT 20.0 (762) (Omer et al., 2000) Eritrea RBPT, CFT 8.5 (1294) (Werney et al., 1979) Somalia RBPT 9.5 (5056) 2. Livestock-subsistence crops/semi- arid regions (Jiwa et al., 1996) Tanzania SAT 10.8 (13 078) (Kadohira et al., 1997) Kenya ELISA 15.0 (640) (Msanga et al., 1986) Tanzania SAT 10.6 (17 758) 3. Cash and subsistence crops with livestock/subhumid regions (Bedard et al., 1993) Malawi SAT 0.3 (2 017) (Bishop et al., 1994) South Africa RBPT, CFT 1.5 (5 059) (Gallangher, 1973) Zambia RBPT, SAT, CFT 11.3 (1 879) (Kagumba and Nandokha, 1978) Kenya RBPT, SAT, CFT 12.8 (1 420) (Kagumba and Nandokha, 1978) Tanzania RBPT, SAT, CFT 1.8 (1 167) (Madsen, 1989) Zimbabwe (A) RBPT, SAT, CFT 1.4 (37 446) (Madsen, 1989) Zimbabwe (B) RBPT, SAT, CFT 3.0 (14 303) (Mohan et al., 1996) Zimbabwe (A) RBPT, SAT, CFT 0.92 (199 909) (Mohan et al., 1996) Zimbabwe (B) RBPT, SAT, CFT 2.7 (3754) 4. Crop-livestock/tropical highland zones (Bekele et al., 1989) Ethiopia RBPT 4.2 (1609) (Kadohira et al., 1997) Kenya ELISA 2.0 (374) 5. Subsistence crop with small livestock/humid regions (Kagumba and Nandokha, 1978) Uganda RBPT, SAT, CFT 2.4 (169) (Kubuafor et al., 2000) Ghana RBPT 6.6 (183)
Zimbabwe (A) = commercial farming sector, Zimbabwe (B) = communal farming sector
27
While there is ample data on the prevalence of brucellosis from such countries as South
Africa, Zambia and Malawi (Bishop et al., 1994; Muma et al., 2006), there are scanty reports from some of the Southern African countries (McDermott and Arimi, 2002).
In Zimbabwe, brucellosis was first suspected in cattle from dairy farms near Harare in
1906 and later confirmed in the Zambezi valley in 1913 (Bevan, 1931). Bevan is believed to be the first to demonstrate that B. abortus from infected cows was transmissible to humans causing a disease indistinguishable from undulant fever
(Anon., 1957). Since these early reports, the disease has been closely monitored and documented. Early studies indicated a seroprevalence of 12.3%, 5.2% and 3.0% in dairy, commercial beef and communal cattle respectively (Manley, 1969).
Consequently, the problem was suggested to be of greater magnitude in dairy than beef cattle. However, this was refuted by Swanepoel et al. (Swanepoel et al., 1976) who found the disease in 18.4% of cattle of mixed breeds tested and concluded that the disease was equally important in both cattle farming sectors. Together with
Campylobacter foetus, B. abortus was found to be the most important causes of infectious infertility and abortion in cattle of all breeds (Swanepoel et al., 1975).
Bovine brucellosis is believed to be endemic in some regions in Zimbabwe (Mohan et al., 1996), while some areas have presumably eradicated it through implementation of a national accreditation scheme that was enforced only for commercial dairy farms in the early 1980s (Madsen, 1989). This eradication scheme was only enforced for the commercial cattle farming sector, as these herds generally registered higher brucellosis prevalence than communal cattle, supposedly due to the extensive nature of cattle rearing in the latter (Madzima, 1987; Madsen, 1989; Mohan et al., 1996). It was also
28 perceived that a greater compliance would be obtained from commercial farms whose financial and infrastructural resource base would augment government effort to be able to implement the brucellosis accreditation programme.
The prevalence of bovine brucellosis in communal areas varies from province to province. The prevalence has been found to be lowest in Mashonaland and highest in
Matabeleland, Manicaland and Midlands (Madsen, 1989; Mohan et al., 1996). This variation in prevalence is believed to be due to the fact that some communal herds are kept closed whereas other cattle owners have purchased infected cattle to improve the genetics of their cattle (Madsen, 1989).
In cattle, there are no breed differences in susceptibility to brucellosis (Madsen, 1989).
Age has been found to be an important risk factor (McDermott et al., 1987). Sexually mature cows are more susceptible to infection than sexually immature heifers and calves (Cunningham, 1977). Therefore, breeds where sexual maturity is attained early appear to be at a higher risk of getting infected. However, this may be biased, as the main focus of bovine brucellosis has been linked to abortions. The difference in susceptibility to infection by Brucella spp. between bulls and cows is not known, as there are no controlled experimental studies conducted (Bishop et al., 1994).
2.5.3.2 Caprine and Ovine brucellosis
Caprine and ovine brucellosis is a disease of economic importance especially in the
Mediterranean and Middle East region and other parts of the world such as Africa,
Central America and Mexico where the incidence is very high and the diseaseis known to be enzootic (Herr, 1994; Banai et al., 2002; OIE, 2004). B. melitensis is the main
29 cause of ovine abortion in Turkey (Leyla et al., 2003). However, world wide, the extent of the distribution of the disease has not been accurately determined, despite the public health importance of B. melitensis (Blood and Radostits, 1989).
B. melitensis infection in goats and sheep is believed to be rare in Southern Africa.
While sporadic outbreaks were recorded in goats in South Africa (Herr, 1994) and
Zimbabwe (Madsen, 1989), the spatial and temporal distribution of the disease remains largely unknown. There are no documented reports of the disease from the other
Southern African countries.
Goats are highly susceptible to B. melitensis infection and there is no difference in breed susceptibility (Blood and Radostits, 1989). Sheep are less susceptible (Alton,
1987), but some breed differences are recognised. Maltese sheep are highly resistant to infection, while fat-tailed breeds such as Awassi and Kurdi that are used for milk production in the Middle East region are highly susceptible (Alton, 1987). In addition, there is no evidence to suggest difference in susceptibility between the sexes in both species (Herr, 1994). In contrast to cattle, it appears that susceptibility to B. melitensis in both goats and sheep is not influenced by pregnancy status as both pregnant and non- pregnant animals are equally susceptible (Herr, 1994).
2.5.3.3 Ovine brucellosis (Ram epididymitis)
The earliest reports of genital lesions in sheep due to B. ovis were reported in Australia and New Zealand (Buddle, 1953). Although the disease has been reported from other sheep rearing countries (OIE, 2004), the extent of its distribution has not been determined.
30
In Southern Africa, ovine epididymitis has been reported from South Africa, Namibia and Zimbabwe ( Van Tonder et al., 1994; Corbel, 1997). However, there are no documented reports of isolation of B. ovis in Zimbabwe.
Sheep are the only host animals that are naturally infected with B. ovis (Blood and
Radostits, 1989). Burgess and co-workers successfully infected goats with B. ovis, even though there was no evidence of clinical epididymitis (Burgess et al., 1985). In addition, experimental infection may produce sub-clinical infections in cattle, white tailed deer, guinea pigs, rabbits, mice and gerbils (Garritty et al., 2005).
2.5.3.4 Porcine brucellosis
B. suis is believed to occur in many countries (Blood and Radostits, 1989), but has not been reported in Southern Africa (Bishop and Bosman, 1994; Corbel, 1997).
The five known biotypes of B. suis vary in their host preference and virulence for pigs
(Bishop and Bosman, 1994). B. suis, biotypes 1, 2 and 3 affect pigs, while biotypes 4 and 5 have been associated with brucellosis in reindeer (Garritty et al., 2005). B. suis biotypes 1, 2 and 3 have also been associated with infections in humans and other animals species such as horses and cattle (OIE, 2004). B. suis biotype 2 has been especially important in wild pig boars (Sus scrofa) where it is believed to cause orchitis and can be transmitted to domestic pigs (Godfroid et al., 1994).
In pigs, susceptibility to infection with B. suis tends to increase with increasing age of the pigs (Blood and Radostits, 1989). However, clinical disease is more common in
31 piglets than in mature pigs (Bishop and Bosman, 1994). Differences in susceptibility to
B. suis infection among pig breeds is not known.
2.5.3.5 Canine brucellosis
B. canis was first recognised in the USA in 1966 as a cause of abortion in beagles
(Carmichael, 1966). Since then the disease has been reported to occur sporadically in
Europe, North and South America, Japan and China (Shin and Carmichael, 1999). It is becoming worldwide in distribution but the occurrence of cases tend to be sporadic
(Quinn et al., 1999). Although B. canis has been isolated from a dog in South Africa
(Gous et al., 2005), its occurrence in Africa is not clear.
2.5.3.6 Equine brucellosis
The extent of distribution of equine brucellosis is not really known. However, it is believed that the distribution of equine brucellosis follows that of cattle and to some extent swine brucellosis. Horses kept together with infected cattle are at a higher rsik of exposure to Brucella spp.
2.5.4 Transmission
2.5.4.1 Bovine brucellosis
The transmission of the organisms is mainly by direct or indirect contact of the mucous membranes with infective excretors (Quinn et al., 1999). Although cattle have been infected experimentally by conjunctival, vaginal and intramammary routes, the main route of infection in the field is the oral route (Cunningham, 1977). Thus, most cattle acquire infection by licking infected material, grazing on infected pasture or consuming other feedstuffs and drinking water contaminated by aborted material or uterine
32 discharges from an infected animal (Blood and Radostits, 1989). Less commonly, infection may occur via conjunctiva or by inhalation (Quinn et al., 1999). While calves born of infected dams may get infected in utero, the majority of them will clear the infection in a few months while a few may remain latently infected (Blood and
Radostits, 1989). Although calves suckle and ingest large numbers of viable organisms from colostrum, it is unlikely that they will be infected by this route as colostral antibodies appear to be protective (Cunningham, 1977).
Transmission by coitus is unlikely, but semen from infected bulls that is used for artificial insemination could be a source of infection (Franklin, 1965; Lambert et al.,
1963). Although Brucella spp. can survive for many days on grass in cold temperate climate (Blood and Radostits, 1989), they are unlikely to survive for a long time in hot and dry tropical climates. Therefore, the period between contamination of the pasture and ingestion by susceptible animals is critical in the transmission of Brucella spp.
2.5.4.2 Caprine and Ovine brucellosis
Transmission of B. melitensis between animals is believed to be through ingestion of food or water contaminated with infected material from the placenta, foetal fluids and vaginal discharges expelled by infected ewes or when they abort or have full-term parturition (OIE, 2004). However, inhalation of contaminated aerosols is believed to be the most important mode of transmission of B. melitensis, a phenomenon that is likely to be accentuated by kraaling of goats at night (Herr, 1994).
Shedding of the organism may occur in milk of affected ewes and does (Alton et al.,
1988). Suckling kids and lambs ingest large doses of B. melitensis, but most of them
33 will recover spontaneously (Herr, 1994) probably because the colostral antibodies tend to be protective (Cunningham, 1977).
Although B. melitensis may be present in semen and urine of affected rams and billy goats, males do not seem to play an important role in the dissemination of the disease
(Alton, 1987; Herr, 1994).
2.5.4.3 Ovine brucellosis (Ram Epididymitis)
B. ovis is believed to be introduced into flocks through introduction of infected rams
(Burgess et al., 1985). Infection spreads from the infected rams to ewes through coitus
(Burgess et al., 1985). B. ovis may be excreted from infected rams even before the development of lesions (Burgess et al., 1985). The incidence of B. ovis was found to increase with increasing age of rams (Murray, 1969; Olsen and Stoffregen, 2005).
There was evidence that B. ovis may be transmitted from ewes to rams during mating
(Murray, 1969). However, this is believed to be only by mechanical means when ewes which harbour the microorganisms, are mated in succession by different rams during the same heat period (Van Tonder et al., 1994). While ram to ram transmission is likely to occur through homosexual contact (Baggley et al., 1985), ewe to ewe transmission has not been demonstrated (Van Tonder et al., 1994).
2.5.4.4 Porcine brucellosis
In naïve herds, introduction of infected boars is believed to be an important risk factor for B. suis infection because transmission is believed to occur through coitus (Blood and
Radostits, 1989). However, transmission through ingestion of material contaminated
34 with semen or vaginal discharges is likely to be a more common route of spread of infection within infected herds (Blood and Radostits, 1989).
2.5.4.5 Canine brucellosis
Transmission is believed to be via both the oral and sexual routes (Shin and Carmichael,
1999; Wanke, 2004). The organisms are excreted in vaginal secretions, both during oestrus and parturition, abortion, and post partum where the organisms can be found in large numbers (Wanke, 2004). Both sexes excrete the B. canis in urine but males have been reported to excrete significantly higher levels than females (Wanke, 2004).
However, relatively lower levels of B. canis exist in urine than in semen (Shin and
Carmichael, 1999). Hence, environmental contamination from urine is likely to be minimal unless urine is contaminated with seminal or prostatic fluid. Nevertheless, a contaminated environment is considered to be an important source of infection for canine brucellosis particularly in boarding kennels (Shin and Carmichael, 1999).
Males secrete the bacteria in semen and may potentially infect clean bitches during mating (Wanke, 2004). B. canis is secreted in milk (Wanke, 2004), making transmission through ingestion of milk possible. However, the role of milk in the transmission of B. canis to puppies is ambiguous (Wanke, 2004).
2.5.4.6 Equine brucellosis
The mode of transmission of Brucella spp. in horses is believed to be through ingestion, or inhalation of contaminate aerosols or by direct contact of mucous membranes with infected material. Horse-to horse transmission may be accomplished by direct contact
35 with an infected horse with open infection (fisyulous withers or poll eveil) during grooming or mating.
2.5.5 Pathogenesis
2.5.5.1 Bovine brucellosis
The pathogenesis of brucellosis is poorly understood and has intrigued researchers for a long time. Similarly, the virulence factors of Brucella spp. are unknown. Although there have been preliminary reports of involvement of toxins, fimbriae and plasmids, none of these has been confirmed, and like other gram-negative bacteria, lipopolysaccharides
(LPS) are presumed to play an important role ( Ficht, 2003; Delrue et al., 2004).
It is widely known that B. abortus has a predilection for the pregnant uterus, udder, testicle and accessory male sex glands, lymph nodes, joint capsules and bursae (Blood and Radostits, 1989). Hence, brucellosis is usually a disease of the sexually mature animals (Quinn et al., 1999). It is believed that soon after entry in the host Brucella spp. are engulfed by phagocytic cells in which they multiply and get transported to regional lymph nodes where initial localisation takes place (Walker, 1999) with subsequent development of hyperplasia and infiltration of inflammatory cells (Anon, 2006). They subsequently enter the circulatory system via the thoracic duct for dissemination to parenchymatous organs and other sites (Quinn et al., 1999). Colonisation of the foetus and placenta occurs rapidly and the factors that control this tropism are unknown but believed to be allantoic fluid factors (Walker, 1999) that include erythritol and steroid hormones. The hormones are secreted by the pregnant uterus and stimulate luxuriant growth of Brucella spp. (Quinn et al., 1999). The rapid multiplication of Brucella spp. results in the development of severe ulcerative endometritis of the intercotyledonary spaces with development of yellowish gelatinous fluid (Cunningham, 1977; Walker,
36
1999). The allantochorion, foetal fluids, and placental cotyledons are next invaded and the villi are destroyed (Cunningham, 1977). The cotyledons are frequently necrotic, yellow-grey in colour and covered by thick brown exudates (Walker, 1999). However, there is a considerable variation in the nature of the uterine and placental lesions in both experimental and natural B. abortus infections (Bishop et al., 1994). Depending on the severity of the placentitis, abortion, premature birth or birth of a viable or non-viable calf may result (Bishop et al., 1994). The exact causal mechanism of abortion is not known, but believed to be due to the interference with foetal circulation due to placentitis, or the direct effect of endotoxins, or directly from foetal stress due to inflammation of foetal tissues (Enright et al., 1984; Walker, 1999). Endotoxins of
Brucella spp. may induce the production of cortisol leading to a decreased progesterone production and an increase in the oestrogen production (Enright et al., 1984). Decreases in progesterone levels and increase in oestrogen levels are known to induce a premature parturition ( Enright et al., 1984; Anon, 2006).
In chronically infected animals, Brucella spp. may localise in sites other than the uterus.
This often results in the development of granulamatous foci in the lymphatic tissues, liver, spleen, bone marrow and other locations (Carter and Chengappa, 1991). In general, such lesions are characterised by development of a typical granulomatous type of inflammatory reaction (Walker, 1999).
2.5.5.1.1 Intracellular survival
The genetic basis of virulence of Brucella spp. is not fully understood. The coining of the term “stealthy bug” (Kohler et al., 2003) couldn’t have been a better description of elusive nature of Brucella spp. One of the mechanisms through which these organisms
37 are able to cause persistent infection in the host is through their ability to survive inside macrophages, which would normally kill and destroy other bacteria. Attempts to explain the basis of this intracellular survival has been dealt with at length elsewhere
( Ficht, 2003; Kohler et al., 2003; Celli and Gorvel, 2004). Basically, they are adapted to surviving in the phagosome as their natural living niche believed to be associated with the rough endoplasmic reticulum (ER) (Celli and Gorvel, 2004). Recent studies have indicated that this phagosome contains acidic environments, is low in nutrients, contains cholesterol, looks different from any existing organelles, and the name
“brucellosome” has been coined for this structure (Kohler et al., 2003) (Fig. 2.2). In these brucellosomes, Brucella spp. are able to produce virulence genes (VirB) which promote multiplication of the organisms in such environments (Kohler et al., 2003).
Fig. 2.2. Characteristic properties of (a) an early vacuole and (b) the replicative niche of Brucella spp. Experimental results suggest that the early phagosome is very acidic and poor in nutrients, resulting in induction of VirB and genes encoding stress proteins. VirB participates in the creation of the ‘brucellosome’ characterised by: (1) absence of fusion with lysosomes, (2) neutral pH, and absence of certain nutrient components. Source: (Kohler et al., 2003).
38
2.5.5.2 Caprine and Ovine brucellosis
Primary bacteraemia is preceded by multiplication of the micro-organisms at the site of entry followed by localisation in the lymph nodes, the udder and the uterus (Collier and
Molello, 1964). Bacteraemia in goats, but not in sheep usually results in the development of systemic signs (Blood and Radostits, 1989). The localisation and colonisation of B. melitensis in the placentomes would lead to placentitis and subsequent abortion (Collier and Molello, 1964). B. melitensis also colonises the foetus resulting in foetal death (Collier and Molello, 1964). However, the factors that control the tropism in the gravid uterus are not known - but similar to B. abortus infections this could largely be influenced by the presence of erythritol (Quinn, et al., 1999; Walker,
1999).
2.5.5.3 Ovine brucellosis (Ram epididymitis)
In the ram, an initial bacteraemia with a mild systemic reaction is followed by localisation of the organism in the epididymis (Blood and Radostits, 1989). The factors that affect this tropism are not known. The earliest evidence of infection occurring after about two weeks of infection is the presence of inflammatory cells in semen (Rahaley and Dennis, 1984). The secretion of B. ovis in the semen is intermittent and is likely to occur three weeks after infection (Blood and Radostits, 1989). The detailed sequence of events of the pathogenesis of the lesions in the epididymis and the associated sex glands is given elsewhere (Van Tonder et al., 1994). Briefly, the initial inflammatory reaction destroys the epithelial cells of the epididymis, resulting in the leakage of semen into the interstitial tissues where it provokes a further inflammatory reaction (Van Tonder et al.,
1994). Ultimately, chronic granulomatous reactions leading to a reduction in both the quantity and quality of semen (Baggley et al., 1984). Infertility may result from total
39 cessation of spermatogenesis or obstruction of the epididymis by granulomas (Van
Tonder et al., 1994).
In ewes, the microorganisms are believed to gain entry through the vaginal mucosa
(Van Tonder et al., 1994). In pregnant ewes, the colonisation of the uterus and the placenta by B. ovis is unlikely to occur and the organism rarely causes placentitis and abortion (Meinershagen et al., 1974). The most plausible reason for this is that the growth of B. ovis is inhibited by erythritol, present in the gravid uterus (Quinn et al.,
1999). However, B. ovis was reported to be able to survive only in the uterus, resulting in low grade pyogenic infection and subsequent foetal death in spite of little or no foetal invasion (Collier and Molello, 1964). An experimental trial suggested the occurrence of repeat breeding in ewes presumably due to pregnancy failure (Hughes, 1972), but the significance of this finding under natural conditions has not been explored further.
2.5.5.4 Porcine brucellosis
Following initial multiplication at the site of entry, bacteraemia subsequently occurs. B. suis does not seem to have organ predilection but often localises in the genital tract, mammary gland, joints and other sites of the skeletal system (Bishop and Bosman,
1994). Most infections of B. suis are characterised by progressive granulomatous lesions (Bishop and Bosman, 1994). The more common manifestations of localisation are abortion and infertility due to localisation in the uterus; lymphadenitis; arthritis and lameness due to bone localisation, and posterior paralysis due to osteomyelitis (Blood and Radostits, 1989).
40
2.5.5.5 Canine brucellosis
The routes of entry for B. canis are the genital, oronasal or conjunctival mucosa
(Wanke, 2004). After the brucellae gain entry, they are believed to be taken up by macrophages and transported to regional lymph nodes and genital organs from where they multiply (Wanke, 2004). This is followed by bouts of recurrent bacteraemia and dissemination to target organs. This subsequently leads to the development of pathological lesions such as placentitis and abortion in females, and epididymitis and prostatitis in males (Shin and Carmichael, 1999).
2.5.5.6 Equine brucellosis
The pathogenesis of equine brucellosis is not clear as there are few studies documenting the disease in this species. However, it is believed that the disease is brought about as a result of chronic localised infection. These chronic localised infections often occur in burase, for example, the supraspinous and atlantal bursae, joints, muscles and tendons
(Anon. 1986; Quinn et al., 1999). The presence of brucellae would elicit a chronic pyogranulomatous reaction, similar to the chronic infections observed in cattle.
2.5.6 Clinical signs
2.5.6.1 Bovine brucellosis
The incubation period of bovine brucellosis varies markedly depending on the size of the infective dose, and the age, sex, stage of gestation and immunity of the affected animals (Bishop et al., 1994). Unlike other classical infectious diseases, the definition of incubation period in bovine brucellosis appears to be controversial as such, and several definitions have been proposed. The incubation period may be defined as the period between exposure and abortion (Brinley Morgan, 1977), the period between exposure
41 and the first appearance of clinical disease, or the period before the first serological evidence of infection can be detected (Nicoletti, 1980). It is interesting to note that some animals may not fit into any of these definitions because they may neither show clinical signs nor seroconversion (Lapraik, 1975). In female animals, the first definition is usually adopted since abortion is the most common presentation (Walker, 1999). In pregnant females, this incubation period is inversely proportional to the stage of development of the foetus (Thomsen, 1950). The clinical disease is characterised by late term “abortion storms” in the fifth month of gestation or later in up to about 90% of the animals at risk in susceptible non-immune herds (Cunningham, 1977; Blood and
Radostits, 1989). Retained placenta is a common sequel and this would appear to be more related to premature parturition rather than to the infection of the uterus
(Cunningham, 1977; Walker, 1999). A small proportion of cows never abort even though foeti may be infected (Cunningham, 1977). Sometimes, some of these full term calves exasperated with stress of infection die within a week of birth (Bishop et al.,
1994). Occasionally, in lactating cows, a decrease in milk yield is observed (Blood and
Radostits, 1989).
In succeeding years, although extensive infection is present in the herd, abortions become fewer and fewer because most animals abort only once (Cunningham, 1977).
This is probably explained by the presence of some degree of immunity, albeit some of these animals remain infected and can shed the organism in subsequent parturitions
(Brinley Morgan, 1977; Quinn et al., 1999).
42
Fig. 2.3: Unilateral Brucella
abortus-induced hygroma in a cow.
In bulls, infection may not be apparent ( Lambert et al., 1963; Bishop et al., 1994). Infection is often acquired early in life as calves and retained into adulthood (Franklin, 1965). The most common findings are usually unilateral and rarely bilateral epididymitis and orchitis (Walker, 1999). Although it is believed that brucellosis is not an important cause of infertility in bulls (Bishop et al., 1994), it could be that in chronically infected animals with bilateral lesions, it is probable that fertility would be affected. A decline in fertility in infected bulls has been documented, even though such bulls retained a normal libido (Kumi-Diaka et al., 1980).
In both sexes, hygromata, involving one or more leg joints (Fig. 2.3), are frequently observed in African countries (McDermott et al., 1987; OIE, 2004). However, such lesions may be found in animals that have been vaccinated with B. abortus S19 vaccine
(Corbel et al., 1989).
2.5.6.1.1 Latency
Some heifer calves that acquire infection in early life, test negative to serological tests conducted at six months of age and yet abort during first pregnancy (Cunningham,
1977). There is growing evidence that in some calves born of infected dams, hidden and localised foci of viable organisms remain even though they test serologically negative
43
(Lapraik, 1975). This condition is referred to as “latency” or hidden infection in which an animal exhibits no signs of infection. Pregnancy reactivates infection due to the production of erythritol which stimulates the proliferation of Brucella spp. (Quinn et al.,
1999). Such heifers could spread infections to susceptible animals in herds which they join (Lapraik, 1982).
2.5.6.1.2 Immunity
2.5.6.1.2.1 Humoral immunity
When cattle are vaccinated with B. abortus strain 19, a live attenuated vaccine, immunoglobulin M (IgM) develops earlier than IgG, being first detected at about 5-7 days and reaching a peak at 13-21 days (Brinley Morgan, 1967). On the other hand, IgG is first detected 14-21 days post vaccination and reaching a peak after 28-42 days
(Brinley Morgan, 1967). Two IgG isotypes, IgG1 and IgG2 are produced and the latter in small amounts (Nielsen, 2002). When cattle are challenged with virulent strains of B. abortus a similar pattern is observed except that the IgG reaches a higher maximum level and persists for much longer periods (Brinley Morgan, 1967). In chronic brucellosis, some animals may have high levels of IgG1 that agglutinates poorly and can mask the normally efficient agglutinating properties of any IgM present (Quinn et al.,
1999). Imunoglobulin A may be produced, but the concentration is very low and these are only important in secretions such as milk in which it is a major component (Duncan et al., 1972). Most cross-reacting antibodies, from exposure to micro-organisms other than Brucella spp. consist mainly of IgM (Nielsen, 2002).
44
2.5.6.1.2.2 Cell-mediated immunity
Advances in the field of immunology have clearly demonstrated that the level of immunity to intracellular pathogens cannot be assessed only on the basis of the level of circulating antibodies (Nelson, 1977). There is growing evidence that cell-mediated immunity against B. abortus involves antigen-specific T-cell activation, CD4+, CD8+ T- cells, in addition to humoral responses (Golding et al., 2001; Oliveira et al., 2002). Host protection against B. abortus is believed to be mediated by Th1 immune response (Zhan et al., 1993). B. abortus triggers the host antigen presenting cells (APC) to secrete interlukin-12 (IL-12), which in turn causes Th0 cells to differentiate into Th1 cells that secrete gamma interferon (IFN-γ) that up-regulates macrophage killing mechanisms
(Zhan et al., 1993). In addition, IL-12 produced by APC triggers natural killer (NK) cells to become killer cells and secrete IFN-γ (Golding et al., 2001). Cytokines secreted by CD4+ help to activate CD8+ T-cells and B-cells, stimulating their differentiation into cytotoxic T-cells and plasma cells (Golding et al., 2001). The cytotoxic T-cells that secrete IFN-γ are able to kill B. abortus infected macrophages (Oliveira et al., 2002).
However, there is limited knowledge on the nature of antigens involved in the stimulation of the protective cellular immunity against brucellosis (Oliveira et al.,
2002).
2.5.6.2 Caprine and Ovine brucellosis
The most common sign of B. melitensis infection in sheep and goats is abortion, usually in late pregnancy during which up to 60% of the pregnant animals may abort (Herr,
1994). Goats may show systemic signs such as fever, diarrhoea and weight loss which may be followed by mastitis, lameness and hygroma (Blood and Radostits, 1989). The number of aborting animals declines in subsequent breeding even though most of these
45 will be infected. Viable kids are infected and in some cases the disease persists in latent form until sexual maturity when clinical signs become evident (Waghela, 1980).
However, kids that are weaned early from their dams may be free from infection (Blood and Radostits, 1989). Orchitis is a common sequel to B. melitensis infection in rams and billy goats (Herr, 1994).
2.5.6.3 Ovine brucellosis (Ram epididymitis)
The incubation period of ram epididymitis varies from about eight weeks to 12 weeks or longer ( Webb et al., 1980; Plant et al., 1986). The most consistent clinical sign of epididymitis is evidenced by swelling of the tail, more often than the head of the epididymis (Van Tonder et al., 1994). Epididymitis may be unilateral or bilateral
(Hughes and Claxton, 1968; OIE, 2004) acute or chronic (Blood and Radostits, 1989).
In acute epididymitis, rams usually show no evidence of inflammation, albeit most of these excrete B. ovis in semen (Blood and Radostits, 1989; Dargatz et al., 1990). When it becomes palpable, it is indicative of a chronic condition. In chronic epididymitis, there is gross enlargement of the affected parts with reduced mobility of the testes due to fibrous tissue formation (Van Tonder et al., 1994). The testes are rarely involved, although in some cases these may be atrophied and reduced in size (Blood and
Radostits, 1989; Van Tonder et al., 1994). While the libido of affected rams is unaffected, they may have reduced fertility or may be completely sterile (Van Tonder et al., 1994; OIE, 2004).
Under experimental conditions, B. ovis infection in pregnant ewes has been reported to cause abortion in 33% of the infected animals (Meinershagen et al., 1974). This is probably ascribed to the fact that B. ovis has low pathogenicity for the pregnant uterus
46
(Collier and Molello, 1964). Therefore, infected sheep rarely abort (Libal and
Kirkbride, 1983) and as a result, the prevalence of B. ovis-induced abortion under field conditions seems to be low (Meinershagen et al., 1974).
2.5.6.4 Porcine brucellosis
Infection of the genital tact in sows results in irregular oestrus, abortions, stillbirths, small litters and birth of weak piglets (Bishop and Bosman, 1994). Abortion may occur early or at any time during gestation (OIE, 2004). Chronic metritis characterised by nodular inflammatory thickening and abscessation and mucoid endometritis may be present (Blood and Radostits, 1989). In boars, B. suis infection is characterised by the development of a chronic granulomatous orchitis with subsequent interference of sexual activity which may be temporary or permanent (Blood and Radostits, 1989). In both sexes affected pigs may suffer swollen joints and tendon sheaths, lameness, and occasionally posterior paralysis (OIE, 2004).
While mortality may be high in piglets, mature animals rarely succumb to the disease
(Bishop and Bosman, 1994; Blood and Radostits, 1989). A significant proportion of the affected pigs will recover from clinical disease, but many remain permanently infected
(OIE, 2004).
2.5.6.5 Canine brucellosis
Clinical signs of generalised canine brucellosis are not clearly evident since they may resemble signs of other systemic infections. Moreover clinical signs vary according to the organ affected (Wanke, 2004). The most common clinical sign of canine brucellosis in pregnant bitches is abortion, between 45 and 55 days of gestation in 75% of the cases
47
(Carmichael and Kenney, 1968; Shin and Carmichael, 1999). Aborted foetuses are usually autolysed and show characteristic lesions of generalised bacterial infection, while the bitches continue to excrete brownish or grey discharge for long periods of time (Carmichael and Kenney, 1968; Wanke, 2004). Puppies may be born alive but usually die within a few hours or days of parturition (Carmichael and Kenney, 1968).
In males, the most frequent manifestation is the occurrence of epididymitis and prostatitis (Shin and Carmichael, 1999). In acute infection, the epididymis often increases in size and accompanied by evidence of pain and accumulation of fluid in the tunica (Wanke, 2004). In chronic infection, the epididymis ultimately decreases in size and testicles appear atrophied (Shin and Carmichael, 1999; Wanke, 2004). The dogs become sterile, presumably due to an autoimmune response that results in the production of anti-sperm antibodies (Shin and Carmichael, 1999).
In both sexes, severe lymphadenitis involving the retropharyngeal and inguinal lymph nodes is often present, although other lymph nodes may be affected (Wanke, 2004).
2.5.6.6 Equine brucellosis
In horses, brucellosis is usually a chronic localised infection. The common sites for these chronic infections include the supraspinous and atlantal bursae. The horse typically presents with gross swellings which may open up and exhibit pus-dreaining sinuses from these bursae and these coditions are commomnly called “fistulous withers”
(supraspinous bursitis) and “poll evil” (atlantal bursitis) respectively (Walker, 1999).
The bursitis in early stages consists of a distension of the supraspinous bursa with a
48 clear, straw-coloured, viscid exudate (Anon., 1986). Occasionally, abortions have been reported (Anon., 1986).
2.5.7 Diagnosis
2.5.7.1 Bovine brucellosis
2.5.7.1.1 Clinical diagnosis
The diagnosis of bovine brucellosis on the basis of clinical signs is usually difficult.
Abortion in the third trimester of pregnancy is suggestive of brucellosis, although other causes of abortion should be ruled out. Several infectious diseases such as Rift Valley fever (RVF), salmonellosis, leptospirosis and listeriosis could cause abortion “storms” in cattle (Blood and Radostits, 1989).
If present, carpal hygroma in some animals could be indicative of prior exposure to B. abortus. However, there are other factors associated with this condition. Moreover, the occurrence of carpal hygroma appears not to be a consistent finding.
2.5.7.1.2 Laboratory diagnosis
2.5.7.1.2.1 Culture and isolation
The culture, isolation and identification of Brucella spp. from an aborted foetus, placenta, vaginal discharge, or milk remain the only test to definitely diagnose brucellosis in individual animals or herds. Under field conditions, obtaining aborted foetuses or placenta may be difficult. Often, the aborted foetuses or aborted materials are autolysed if they are located, and in most cases they may not be located. If the aborted foetus is located and still fresh, suitable specimens for the culture of Brucella
49 spp. include the stomach contents, pieces of the foetal liver, spleen and lung or the placental cotyledons (Alton et al., 1988; Bishop et al., 1994).
When aborted material cannot be utilized, culture and isolation of Brucella spp. is usually performed on vaginal discharges or milk. A major disadvantage of relying on culture and isolation is that some cases are misdiagnosed as negative because Brucella spp. are slow growing and fastidious micro-organisms that are easily overgrown by contaminating bacteria. Hence, for practical reasons, the only method relied upon for herd diagnosis is demonstration of antibodies in animals previously exposed to antigens of Brucella spp. In recent years, molecular methods such as the PCR have been suggested as alternative gold standard tests to confirm brucellosis (Bricker, 2002).
However, such methods may be too expensive to be relied upon in routine diagnosis of bovine brucellosis, especially in resource-poor countries, and the question always remains if the PCR detects viable bacteria or not.
2.5.7.1.3 Serology
Serological techniques for the diagnosis of bovine brucellosis have been in existence for over a century. Consequently, a multitude of serological tests have been developed and each has its own special applications and limitations (Alton et al., 1975; Mikolon et al., 1998). In reality, only a few of these serological tests are considered to have acceptable sensitivities and/or specificities. For example, the Rose Bengal plate test
(RBT), indirect immunosorbent assay (i-ELISA), competitive immunosorbent assay (c-
ELISA), complement fixation test (CFT) and the fluorescence polarisation assay (FPA) are recommended for the purpose of international trade of livestock (OIE, 2004). The choice of which test to use in brucellosis surveillance programmes especially in
50 developing countries, depends on several factors that include specific objectives of the programme, cost of setting up the test, technical competence and application adaptability of the technique. Table 2.6 gives some information about the main tests used.
Table 2.6. Summary of the serological tests for bovine brucellosis (Quinn et al., 1999).
Serological test Principal Comments immunoglobulin class identified
HERD TEST
Brucella milk ring test IgM, IGg1, IgA Conducted on bulk milk from a herd. If positive sera are collected from individual cows in the INDIVIDUAL TESTS herd and subjected to one or more of the tests. Plate agglutination tests:
Rose Bengal plate test and the Card test IgG1, IgM Useful screening tests. Antigen buffered to pH
3.65-4.0 and this allows IgG1 to cause agglutination.
Brucella serum agglutination test (SAT) IgM, IgG2 (IgG1) Widely used test, but often IgG1 fail to agglutinate, so false negatives may occur.
Brucella complement fixation test (CFT) IgG1, IgM One of the most specific of the serological tests. Enzyme linked immunosorbent assays Has capacity to detect all Comparative new tests for Brucella antibodies (ELISA) immunoglobulins but proving reliable, especially the competitive ELISA Fluorescence polarisation assay (FPA) Has capacity to detect all A new test. It is very sensitive and specific immunoglobulins
SUPPLEMENTARY TESTS
Coombs antiglobulin test IgG1, IgG2, IgM Very sensitive test, will detect ‘incomplete’ antibodies that do not react in the Brucella SAT. Rivanol precipitation IgG These are designed to differentiate the “non- Mercapto-ethanol treatment IgG specifics reactions by destroying IgM, the antibody most commonly produced by adult vaccination
Serological tests mainly rely on the detection of antibodies produced against lipopolysaccharides (LPS) of both smooth and rough Brucella spp. The three smooth
51 species; B. abortus, B. melitensis and B. suis which contain the O-polysaccharide (OPS) as part of the lipopolysaccharides (LPS) are diagnosed serologically using either a whole cell antigen or smooth lipopolysaccharide (SLPS) prepared by chemical extraction (Nielsen, 2002). The rough species; B. canis and B. ovis, which contain no detectable OPS are mainly diagnosed using rough lipopolysaccharides (RLPS) or protein antigens (Nielsen, 2002). Because common epitopes are present in all smooth
Brucella spp., virtually all serological tests for antibody to these bacteria utilise B. abortus antigen (OIE, 2004).
Brucella spp. share similar antigens with those of other Gram negative bacteria such as
Francisella, Campylobacter, Vibrio cholerae, Salmonella, Pasteurella and Yersinia enterocolitica serotype O:9 ( Quinn et al., 1999; Nielsen et al., 2004). Therefore, non- specific serological reactions due to cross reacting antibodies produced against other bacteria are commonly encountered in Brucella serology. The occurrence of such non- specific serological cross reactions become significant especially towards the end of eradication programmes when the prevalence of brucellosis is low (Godfroid et al.,
2002).
Serological diagnosis of brucellosis was first accomplished using an agglutination test
(Wright and Smith, 1897). This is similar to the serum agglutination test (SAT)
(standard tube agglutination test) that mainly detects the immunoglobulin type M (IgM) and also the IgG2 which are the major agglutinating antibodies (Alton, 1977; Nielsen,
2002). Although its sensitivity is good, its specificity is low due to crossreactions by
IgM antibodies produced against B. abortus S19 vaccine and other bacteria closely
52 related to Brucella spp. (Nielsen et al., 1996a). Hence the use of this test for the purpose of international trade of livestock is discouraged (OIE, 2004).
Several modifications of the original agglutination test have been made to increase the specificity and among them is the use of an acidified antigen in the Rose Bengal/card test (Nicoletti, 1967) and the buffered plate agglutination test in which stained B. abortus antigens are used at low pH of 3.65. The low pH of the assay prevents agglutination with IgM and encourages agglutinations with IgG1, thereby reducing non/specific reactions (Nicoletti, 1967). While the RBT is reported to be highly sensitive and and an ideal test for screening individual animals, its specificity is often low (Alton et al., 1975) thereby requiring further confirmation of sera using more specific tests.
The CFT was later developed and its two forms, the hot and cold CFT were standardised (Hill, 1963) to detect mainly the complement fixing antibody IgG1 (Alton,
1977). Assays that predominately measure IgG1 are considered to be the most useful
(Nielsen, 2002). Among other problems, the CFT failed to distinguish between B. abortus S19 vaccine-induced antibodies and those from natural infection in addition to the occasional occurrence of serum samples that activate complement in the absence of antigen (Nielsen, 2002). Furthermore, the CFT is cumbersome, requiring experienced technicians to carry out. Nevertheless, the test is commonly used in many brucellosis eradication programmes and is a prescribed test for international trade of livestock
(OIE, 2004).
The primary binding assays, specifically the i-ELISA were developed to increase test sensitivity, though the specificity was found to be low (Nielsen et al., 1989). To
53 alleviate the shortfalls of low specificity created by i-ELISA, a c-ELISA was developed and validated (Nielsen et al., 1995). The c-ELISA is based on a monoclonal antibody specific for the O-polysaccharides (OPS) of the smooth lipopolysaccharides (SLPS) chains. The monoclonal antibody competes with antibodies present in the test serum.
Although the c-ELISA is highly sensitive (it detects all antibody isotopes, IgM, IgG1,
IgG2 and IgA), the monoclonal antibody is selected based on its affinity to exclude most antibodies resulting from B. abortus S19 vaccine without loss of sensitivity
(Muma et al., 2007). Hence, the test can distinguish antibodies produced against the vaccine strain and those against B. abortus field strains (Nielsen et al., 1989; Nielsen,
2002). Although the binding assays are fairly rapid to perform, they require several manipulations and may be prone to error. Further, the need to automate the test makes it more expensive for developing countries. Consequently, such tests are not suitable for field use.
A homogenous assay, the FPA, designed to detect and measure both small molecules
(Fernando et al., 1992) and macromolecules (Hasoda and Yasuda, 1989), was later adapted and validated for detection of antibodies against field Brucella spp. in cattle
(Nielsen et al., 1996a; Dajer et al., 1999; Samartino et al., 1999; McGiven et al., 2003), sheep and goats (Minas et al., 2005; Ramirez-Pfeiffer et al., 2006) , pigs (Nielsen et al.,
1999) and humans (Lucero et al., 2003). The FPA is based on the rotational differences between a small fluorochrome labelled antigen molecule in solution and the antigen molecule complexed with its antibody. The rate of rotation is inversely proportional to the size of the molecule, and smaller molecules rotate faster than larger molecules. The rate at which a molecule rotates is assessed by measuring the time it takes to rotate through a given angle, using polarised light in two planes (Muma et al., 2007). Thus,
54 small antigen molecules in solution will rotate rapidly. If an antibody binds to the small antigen molecules, the size of the molecule increases and this slows down the rotational movement. This rotational difference can be measured using a fluorescence polarisation analyser.
The FPA has been validated for rapid detection of antibodies to Brucella spp. not only in serum, but also whole blood or milk (Nielsen et al., 1996a; Samartino et al., 1999).
The FPA has been shown to have high sensitivity and specificity and the ability to differentiate antibody responses due to vaccination with B. abortus S19, similar to the c-ELISA (Nielsen et al., 1998). The assay is rapid, simple, done in solution in a single tube with no washing or precipitation steps and requires simple equipment. This makes the test adaptable to field use and this could offer a better turn-around time especially in developing countries.
Several other serological tests that have been evaluated for the detection of antibodies against Brucella spp. include; the USDA card test, rapid automated presumptive test,
Mexican Rose Bengal plate test, French Rose-Bengal plate test, USDA standard plate test, USDA buffered acid agglutination test, USDA and Mexican rivanol tests (Mikolon et al., 1998). MacDiarmid and Hellstrom (MacDiarmid and Hellstrom, 1987) reported a
Brucella tuberculin test (Brucellin test) that despite its relative low sensitivity, showed a high degree of specificity by being able to differentiate B. abortus S19 vaccinated animals from those infected by virulent strains of B. abortus. However, the majority of these serological tests are rarely used for routine diagnosis of brucellosis.
55
It is interesting to note that up till now, no single test is claimed to have perfect sensitivity and specificity. An ideal test should be sensitive, specific and be able to detect all stages of infection (McGiven et al., 2003). Unfortunately, none of these possesses attributes of an ideal test. It is therefore a common practice in brucellosis eradication programmes to use such tests in combination, either in parallel or in series in order to improve their performance. The parallel test programme maximises on sensitivity, while a serial programme maximises on specificity. Thus, a serial programme would become more appropriate towards the end of a brucellosis eradication programme when more false positive reactions are likely to occur
(Godfroid et al., 2002).
In Zimbabwe, a serial testing programme using the RBT as a screening test, followed by confirmation using the SAT and CFT, is used in routine serological diagnosis of bovine brucellosis and the guidelines for interpretation of the results are given
(Madsen, 1989). A major constraint of this programme is the difficulty in differentiating antibodies induced by B. abortus S19 from those produced against field strains of B. abortus (Nielsen, 2002). In addition, the test programme allows doubtful reactors to be treated as negative (Madsen, 1989) and this could prolong the brucellosis eradication campaign because such animals could be latently infected (Lapraik, 1982).
2.5.7.1.4 Milk Ring Test
The milk ring test (MRT) was developed by Fleischner in 1937 (Brinley Morgan, 1967) as an adaptation of the serum agglutination test to detect the presence of antibodies against Brucella spp. in milk (Hunter and Allan, 1972). It is commonly used to monitor brucellosis using bulk tank milk, and is thus recommended as a screening test (OIE,
56
2004), but its sensitivity can easily be affected by pooling of milk samples.
Modifications of the original MRT have been done in the USDA and Mexican milk ring tests. Other tests such as the ELISA and the FPA have been used successfully to detect antibodies to Brucella spp. both in individual and bulk milk samples (Nielsen and Gall,
2001).
2.5.7.2 Caprine and Ovine brucellosis
Caprine and ovine brucellosis may be suspected on the basis of clinical signs but the various causes of abortion, lameness, hygroma and orchitis need to be ruled out (Blood and Radostits, 1989). The presumptive diagnosis of ovine brucellosis may be made from impression smears of the placenta, abomasal contents, and vaginal discharges stained with the Stamp’s or Modified Ziehl Neelsen stain (Herr, 1994). Unequivocal diagnosis of the disease is done through culture and isolation of the aetiological agent from the aborted foetus, vaginal discharge or milk. Blood cultures may be performed in goats since bacteraemia persists for up to a month following infection (Blood and
Radostits, 1989). However, for practical reasons, culture and isolation of the causative agent is seldom done for the diagnosis of the disease in sheep or goat flocks. Although a species specific PCR technique for diagnosis of B. melitensis has been reported (Lelya et al., 2003), high input costs may be prohibitive for the use of this technique as a diagnostic tool in developing countries. Consequently, serological methods are often relied upon.
Several serological tests have been evaluated for the diagnosis of B. melitensis infection in sheep and goats (Blasco et al., 1994; Jacques, 1998; Minas et al., 2005; Ramirez-
Pfeiffer et al., 2006). The main serological tests used for the diagnosis of B. melitensis
57 infection in sheep and goats are the RBT as a screening test and the CFT (Alton et al.,
1988) as the confirmatory test. But these tests are not specific enough to discriminate serological reactions due to B. melitensis Rev. 1 vaccine and Yersinia enterocolitica O:9 from those due to natural infection (OIE, 2004). The SAT is least reliable due to the occurrence of false positives (Waghela et al., 1980; OIE, 2004). However, the need to improve test accuracy culminated in the development of the i-ELISA (Jacques, 1998), the c-ELISA (Nielsen et al., 2005) and FPA (Minas et al., 2005) whose ability to identify infected animals appears to be better than that of the RBT and CFT.
2.5.7.3 Ovine brucellosis (Ram epididymitis)
The diagnosis of B. ovis requires the correlation of laboratory tests with flock history and clinical examination (Plant et al., 1986). Clinical diagnosis is complicated because there are several causes of ram epididymitis (Van Tonder et al., 1994).
The staining of affected tissues and semen with Stamp’s method is helpful in diagnosing B. ovis infection (Meinershagen et al., 1974). Bacteriological culture of semen and uterine exudate (Van Tonder et al., 1994) and affected tissue such as the placenta, the ampulla, the tail of the epididymis and the seminal vesicle is recommended to confirm the diagnosis (Plant et al., 1986).
The serological tests such as the RBT and CFT, indirect haemagglutination and ELISA have been used for the diagnosis of B. ovis infection ( Webb et al., 1980; Burgess and
Norris, 1982; Lee et al., 1985). Of these, the CFT seems to be the most reliable test although its sensitivity appears to be lower than that of the ELISA (Lee et al., 1985).
Although false positives have been obtained with the CFT in infected flocks (Hughes
58 and Claxton, 1968; Searson, 1986), no false positives have been recorded in flocks free of the disease (Hughes and Claxton, 1968).
2.5.7.4 Porcine brucellosis
Although abortion storms in piggeries are common, the clinical diagnosis of porcine brucellosis is complicated because in most instances the causes are several and these are often not resolved (Saunders, 1958). The confirmation of diagnosis of porcine brucellosis is based on culture and isolation of the causative agent because to date none of the conventional tests in use are reliable for the diagnosis of the disease in individual animals (OIE, 2004). The development of a B. suis-specific PCR in recent years (Fayazi et al., 2002) looks promising and may complement the shortfalls of these conventional diagnostic techniques. This technique has not been adopted widely, and studies on evaluation of the sensitivity and specificity of this PCR technique under different laboratory conditions have apparently not been conducted. Moreover, such tests are suitable for the diagnosis of B. suis in individual animals. Therefore, for brucellosis diagnosis in a herd, serological testing is recommended and the SAT is commonly used
(Quinn et al., 1999). Unfortunately, this test is not reliable due to the occurrence of low titres and non-specific reactions (Quinn et al., 1999). Serological cross reactions may occur in pigs that are infected with Y. enterocolitica serotype O: 9 (Wrathall et al.,
1983).
Experiences have shown that the RBT and CFT are the tests of choice (Quinn et al.,
1999). However, a low sensitivity has been reported for the CFT (Rogers et al., 1989).
A competitive ELISA test utilising the O-lipopolysaccharide antigens of B. suis was developed and it performed better than these conventional tests (Nielsen et al., 1999).
59
The FPA which has been shown to be able to eliminate serological cross reactions, in addition to being more robust was later validated for the serological diagnosis of porcine brucellosis (Nielsen et al., 1999).
2.5.7.5 Canine brucellosis
Canine brucellosis is often an insidious disease and is difficult to diagnose clinically
(Gordon et al., 1985). The only method to provide definitive diagnosis is by culture and isolation of the agent, but this is not always possible.
Serology is used to diagnose canine brucellosis but the limitations of the serological tests may make the diagnosis inaccurate. In addition, the diagnosis is further complicated by the occurrence of unstable levels of serum antibodies (Wanke, 2004).
Nevertheless, tests such as the 2-mecarptoethanol tube agglutination test (METAT), rapid slide agglutination test, agar gel immunodiffusion test (AGID) and the CFT are recommended for the detection of antibodies to B. canis (Gordon et al., 1985; Quinn et al., 1999). A diagnostic scheme where the METAT is used as a screening test and the
AGID as a confirmatory test has been proposed (Gordon et al., 1985). This is because the METAT has a high sensitivity while the AGID is highly specific (Wanke, 2004).
ELISAs have shown promising results due to superior sensitivity (Baldi et al., 1994), but these have not been used widely for the diagnosis of canine brucellosis.
2.5.7.6 Equine brucellosis
The resumptive diagnosis of equine brucellosis is based on the presenting clinical signs.
However, other causes of chronic draining abscesses have to be ruled out. Other micorganisms such as Actinomyces bovis have been isolated together with B. abortus
60 from poll evil and fistulous withers (Quinn et al., 1999). Nevertheless the isolation of B. abortus serves to confirm the diagnosis.
The use of serology in the diagnosis of equine brucellosis has not gained much ground.
Serological tests such as the SAT have been used to demonstrate the presence of serum antibodies (Anon., 1986). However, further work is required to establish the relationship between serum antibody titres and infection since most sera showed low antibody titres (Anon., 1986).
2.5.8 Treatment
2.5.8.1 Bovine brucellosis
Antibiotic treatment of bovine brucellosis is usually futile and normally not undertaken.
Different treatment regimens using agents such as trace elements, vitamin mixtures, and antimicrobial agents such as phenol, azo and flavine dyes, have been tried, but all have yielded mixed results (Milward et al., 1984). Similarly, different antibiotic combinations have been evaluated for treatment of bovine brucellosis but without much success ( Milward et al., 1984; Blood and Radostits, 1989). A combination of oxytetracycline and streptomycin was found to successfully treat 71.4% of the infected animals, while sulphonamides or penicillin were found to be less effective (Milward et al., 1984). Under in vitro conditions, B. abortus have been found to be sensitive to gentamicin, kanamycin, tetracyclines and rifampin (Timoney et al., 1988), but the efficacy of these antimicrobials in vivo have not been comprehensively evaluated.
61
2.5.8.2 Caprine and Ovine brucellosis
The treatment of B. melitensis infection is not usually undertaken, and probably not recommended due to the possibility of exposure to humans from handling infected animals.
Although B. melitensis has been found to be sensitive to aminoglycosides, rifampin and tetracyclines in vitro, the efficacy of antimicrobial therapy or prophylaxis of sheep and goats has not been reported (Timoney et al., 1988).
2.5.8.3 Ovine brucellosis (Ram epididymitis)
Antibiotic treatment of B. ovis infection is usually not undertaken (Blood and Radostits,
1989). In experimentally infected rams, a combination of oxytetracycline and streptomycin given parenterally produced good results (Marin et al., 1989; Dargatz et al., 1990). However, although antibiotic therapy was apparently effective as it stopped excretion of B. ovis in semen, it did not resolve clinical epididymitis (Marin et al.,
1989).
2.5.8.4 Porcine brucellosis
In most piggeries, the treatment of porcine brucellosis is not advisable because antibiotic therapy is considered to be ineffective. In vitro, B. suis have been found to be sensitive to aminoglycosides, rifampin, and tetracyclines, but the clinical efficacy of these antimicrobial agents in the treatment of infection is very low (Timoney et al.,
1988). Treatment of B. suis infection using chlortetracycline or a combination of streptomycin parenterally and sulphadiazine orally were shown to be ineffective (Blood and Radostits, 1989).
62
2.5.8.5 Canine brucellosis
Antibiotic treatment of dogs in kennels or where dogs cannot be isolated is not recommended because cure is difficult to achieve (Shin and Carmichael, 1999). A four week continuous treatment using a combination of tetracycline and streptomycin or dihydrostreptomycin, administered within the first three months of infection have been found to give successful therapy (Shin and Carmichael, 1999; Wanke, 2004). However, recrudescence of infection after the cessation of antibiotic treatment is not uncommon
(Wanke, 2004).
2.5.8.6 Equine brucellosis
The most successful treatment is complete dissection and removal of of the infected bursa (Anon., 1986). The earlier this is carried out the better, because the prognosis in chronic cases is bad. Sodium ididide therapy is of little value (Anon., 1986).
2.5.9 Control
2.5.9.1 Bovine brucellosis
Bovine brucellosis has been controlled and successfully eradicated in some countries through vaccination, coupled with stringent test and slaughter policies. In many countries, the practice of purchasing animals to improve genetics and intensive management systems often makes the control of bovine brucellosis difficult due to exposure to infection of many highly susceptible animals (Nicoletti, 1984). Similarly, in developing countries in the subtropics, control of the disease is complicated by such practices as communal grazing, pastoralism and non-controlled livestock trade
(McDermott and Arimi, 2002; Nicoletti, 1984). Under such management, hygienic
63 measures as segregation of purchased animals or keeping parturition cows separated from the herd is impractical.
2.5.9.1.1 Control by vaccination
Several vaccines have been developed and are licensed and available for use in some countries. Back in 1906, Bang observed that cattle could be protected from infection by immunising them with live virulent cultures of Brucella organisms (Bishop et al.,
1994). However, the use of such vaccines could potentially spread the disease when used in other susceptible herds. Therefore a safer, live attenuated vaccine, B. abortus,
S19 was later introduced and found to be very effective in controlling bovine brucellosis (Nelson, 1977). Vaccination with B. abortus S19 by itself will not eradicate bovine brucellosis, but it raises the level of immunity for individual animals such that following exposure to virulent strains of B. abortus, undesirable consequences of brucellosis are minimised (Nelson, 1977). The use of B. abortus S19 vaccine should only be recommended where the prevalence of the disease is high and cessation of vaccination should be considered when the prevalence is reduced to 0.2% or less (Alton et al., 1988). Nevertheless, B. abortus S19 vaccine has been the most widely used vaccine in the control of bovine brucellosis (Schurig et al., 2002).
The normal practice of using a standard dose of 5x1010 viable organisms per dose
(Bishop et al., 1994), to vaccinate calves between 3 to 6 months of age has been reported to give long term immunity and benefits of re-vaccination has not been firmly demonstrated (Berman and Irwin, 1952), contrary to what has been reported (Nicoletti et al., 1978). Moreover, antibody titres would decline to a point where 6-8 months after vaccination it is rare to find IgG in the serum (Nelson, 1977). This will be an added
64 advantage in countries where test and slaughter is practiced since occurrence of B. abortus S19 cross-reacting antibodies will be minimised. Although some studies have advocated for the use of a reduced dose (2x 108 – 3x 109 organisms/dose) (Bishop et al.,
1994) to vaccinate adult animals to control bovine brucellosis (Alton and Corner, 1981), the benefits of this practice are debatable (Nelson, 1977). A major set-back of using B. abortus S19 vaccine in adult cattle is that significantly more animals will have persistent antibody titres than those vaccinated as calves (Nelson, 1977; Beckett and
MacDiarmid, 1985). This will interfere with serological tests in herds where test and slaughter is being practiced. In addition, the use of B. abortus S19 has been associated with abortions in cows vaccinated during pregnancy (Beckett and MacDiarmid, 1985), sterility problems in males, occasionally with low levels of protection (Nelson, 1977) and arthropathy (Corbel et al., 1989).
A variety of vaccines prepared from killed cells of Brucella spp. have been tried and tested (Schurig et al., 2002), but with the exception of B. abortus strain 45/20 (McEwen and Priestley, 1938), the practical use of these preparations has been very limited. B. abortus 45/20 was found to offer protection comparable to that of B. abortus S19 if administered as double doses in adjuvant (McEwen and Priestley, 1940). The need for a booster and the irritant nature of the adjuvant might make this vaccine more expensive to use and less desirable than B. abortus S19. Moreover, like any other killed vaccine, the use of B. abortus 45/20 may be associated with low level of cell-mediated immunity which is critical in protection against infection with Brucella spp. (Oliveira et al., 2002).
A potential vaccine candidate, B. abortus M-strain was discontinued from trials because the strain offered low protection (Huddleston, 1946).
65
A rough mutant B. abortus RB51 has been a promising vaccine candidate, lacking the antibody inducing antigens but still giving a similar cellular protection as B. abortus
S19 (Schurig et al., 1991). However, its efficiency over B. abortus S19 remains a subject of debate (OIE, 2004). Similar to B. abortus S19, the B. abortus RB51 vaccine has been reported to cause placental infection and placentitis, and abortion in vaccinated cattle ( Palmer et al., 1996; OIE, 2004) as well as infections in humans (OIE, 2004).
The use of DNA vaccines in farm animals has not gained any ground even though they have shown much promise in laboratory animals (Schurig et al., 2002). Consequently, the usefulness of these vaccines to control animal brucellosis is a subject for future research.
The use of Brucella spp. vaccines, to successfully control bovine brucellosis has not been extrapolated to wildlife animals as few studies have been conducted. There is limited evidence that these vaccines may produce as good results as they do in domestic animals (Davis and Elzer, 2002).
2.5.9.1.2 Control programme on a herd basis
The strategies to control bovine brucellosis may differ from herd to herd depending on such factors as the level of infection present and general immune status of the herd.
During an abortion storm, the test and disposal of reactors may be unsatisfactory because the spread of infection occurs faster than disposal is possible (Nicoletti, 1984).
Hence, isolation of infected animals, disposal of aborted foetuses, placentas and uterine discharges, and subsequent disinfection of the contaminated areas is recommended
(Blood and Radostits, 1989).
66
In heavily infected herds, all calves should be vaccinated using the recommended vaccines such as B. abortus S19 or B abortus RB51. In serologically positive herds, positive reactors should be culled (Blood and Radostits, 1989). In herds where infection is light, vaccination of calves may be optional, but herds should be monitored regularly using milk ring tests (Blood and Radostits, 1989). Maintenance of closed herds would provide beneficial results if eradication of the disease is the ultimate goal.
2.5.9.1.3 Control of bovine brucellosis in Zimbabwe
In Zimbabwe, the control programme for bovine brucellosis depends heavily on presumptive diagnosis of infection by serological tests and the subsequent recommendation for slaughter of the infected cattle (Anon., 1995). Accreditation scheme, aimed at control and possible eradication, has been legislated for the commercial dairy farming sector in the 1980s (Madsen, 1989; Mohan et al., 1996). To be accredited as brucellosis free, a farm needs to pass three negative tests conducted at three monthly consecutive periods, on all dairy cattle on the farm that are over 18 months of age. An accreditation certificate is then issued which is valid for one year after which renewal is based on a single serological test conducted on all animals above
18 months of age. If positive animals are detected, they are culled through slaughter and the whole process repeated again (Mohan et al., 1996). The farms are monitored monthly by milk ring test (MRT) conducted on bulk milk. If a positive test is recorded, the accreditation certificate is forfeited. Individual animals have to be tested serologically to identify reactors. To be re-accredited, the whole process of three-month interval serological testing and culling of positive animals is repeated. In addition, vaccination of calves between 3 to 10 months of age, using B. abortus S19 strain was recommended (Anon., 1995).
67
2.5.9.2 Caprine and Ovine brucellosis
Effective control of caprine and ovine brucellosis is based on proper hygiene at kidding or lambing and safe disposal of infected material and culling of infected animals (Blood and Radostits, 1989). However, where resources are limited, it is recommended that entire flocks rather than individual animals are destroyed.
In certain circumstances, vaccination is recommended using Elbeg’s B. melitensis, Rev.
1, a live attenuated vaccine ( Elber, 1981; Banai et al., 2002). Although B. suis strain 2 vaccine has been advocated for vaccinating sheep against B. melitensis infection, it has been demonstrated that B. melitensis Rev 1 gives a better protection (Verger et al.,
1995). The use of a killed vaccine, H38, prepared from B. melitensis biovar 1 has been reported, but this vaccine has been associated with protection failures (Alton, 1987).
2.5.9.3 Ovine brucellosis infection (Ovine Epididymitis)
In infected flocks, separation of young and old rams is believed to reduce the risk of infection in young rams (Timoney et al., 1988). Although the culling of rams with clinical epididymitis will reduce the incidence of the disease, it will not eliminate infection (Baggley et al., 1985). An effective control programme is based on serological test and slaughter of positive reactor animals (Murray, 1969).
Vaccination or a combination of vaccination and elimination of positive animals may be adopted (Van Tonder et al., 1994). B. melitensis Rev., 1 vaccine has been found to be very effective when used in young rams at weaning (Van Heerden and van Ransburg,
1962). B. abortus S19 has been found to offer unsatisfactory results as it has been associated with epiphysitis in young rams (Claxton, 1968). A killed B. ovis in saline-oil
68 adjuvant offered solid protection, but had side effects at the site of injection (Claxton,
1968).
There is no evidence that vaccination of ewes is of value in reducing spread (Blood and
Radostits, 1989). Further research is required to assess the usefulness of this approach in the control of ovine epididymitis.
2.5.9.4 Porcine brucellosis
The control of porcine brucellosis has been based on test and slaughter of reactor animals (OIE, 2004). However, the occurrence of false negative animals is high using either bacteriological culture or serology (OIE, 2004) and this complicates the stamping out policy.
Vaccination of pigs against B. suis has not been practised widely in many countries as no suitable vaccine is available (Blood and Radostits, 1989). Nevertheless, the use of an orally administered vaccine prepared from B. suis strain 2 used to immunise not only pigs, but also cattle, sheep and goats, against animal brucellosis, has been reported in
China and other countries (Deqiu et al., 2002).
2.5.9.5 Canine brucellosis
There are no suitable vaccines for control of canine brucellosis. Prevention of infection and elimination of infected dogs constitute the principal control strategy (Shin and
Carmichael, 1999). In areas where canine brucellosis is endemic, annual serological testing of the breeding stock and all dogs introduced to kennels should be recommended.
69
2.5.9.6 Equine brucellosis
For prophylactic control, it is davisable to keep horse separate from Brucella-infected cattle, and horse with discharging fistulous withers from cattle (Anon., 1986). Brucella vaccines are not helpful in the control of equine brucellosis.
2.6 Human brucellosis
2.6.1 Introduction
Brucellosis is an important disease that has serious implications on the health of humans. Human brucellosis remains the most common zoonotic disease world wide with more than 500 000 new cases annually (Pappas et al., 2006). Brucellosis is usually a serious occupational hazard affecting primarily meat-packing employees, farmers, butchers, livestock producers and veterinarians (Chukwu, 1987).
2.6.2 Aetiology
Four species, B. abortus, B. suis, B. melitensis and B. canis are associated with systemic brucellosis in humans, and are thus recognised as important zoonotic pathogens of public health significance (Quinn et al., 1999). The causative agent of Malta fever, B. melitensis, was first discovered by Bruce on the island of Malta in 1887, when it was isolated from the spleen of a soldier who had died of the disease (Hall, 1989). The agent was named Micrococcus melitensis (L. honey, sweet isle) (Nicoletti, 2002) but was later re-named B. melitensis..
B. melitensis is highly pathogenic for humans and accounts for the majority of cases and all the three biovars are equally involved (Doganay and Aygen, 2003; Pappas et al.,
2006). On the other hand, most of the human infections due to B. abortus are caused by
70 biovar 1 that is widely distributed in cattle in most countries. However, there are no proven differences in the pathogenicity of field strain biovars (Nicoletti, 1980).
Occasional cases of infection due to B. abortus S19 vaccine strain have been reported in vaccination accidents (Nelson, 1977). Of the B. suis biovars, only 1 and 3 are commonly associated with human brucellosis (OIE, 2004). B. suis biotype 2 is generally considered to be non-pathogenic for human (Garritty et al., 2005). Cases due to B. canis are infrequent, but important especially in laboratory workers (Doganay and Aygen,
2003; Wanke, 2004).
2.6.3 Epidemiology
Although brucellosis due to B. melitensis is a world wide zoonosis, it predominates in
Mediterranean countries, the Middle East and Latin America (Lucero et al., 2005).
Human brucellosis has also been reported in many African countries (Chukwu, 1985;
McDermott and Arimi, 2002; Pappas et al., 2006).
Interestingly, human brucellosis due to B. melitensis has not been reported in
Zimbabwe. It is quite probable that B. melitensis is not present in Zimbabwe since the only sporadic cases of the disease in goats were reported more than a decade ago
(Madsen, 1989).
Brucellosis due to B. abortus is distributed world wide, especially in developing countries in Africa, Asia and Latin America (Pappas et al., 2006). In Zimbabwe, the occurrence of human brucellosis caused by B. abortus was demonstrated by Bevan in
1913, when people were infected by consuming unpasteurised milk from infected cows
(Chukwu, 1987). Bevan is believed to be the first person to demonstrate that B. abortus
71 from infected cows was transmissible to humans to cause a disease indistinguishable from undulant fever (Anon., 1957).
However, the incidence of B. abortus infection in humans in Zimbabwe is not known. It could be that a lot of cases remain undiagnosed since brucellosis is difficult to detect clinically (McDermott and Arimi, 2002), due to the numerous nature of causes of recurrent fevers. Alternatively, largely due to rigorous brucellosis control measures that include milk and meat hygiene (Madsen, 1989), the incidence of human brucellosis is probably low.
The incidence of B.suis infection in humans is not known. It was reported to be fairly common in the USA in the 1970s probably due to rigorous control of B. abortus
(Pappas et al., 2006).
Although B. canis affects man, there are few reported cases (Quinn et al., 1999).
Moreover many illnesses, generally from close contact with infected animals, is often mild and usually not noticeable (Shin and Carmichael, 1999). Isolated cases have been however, reported in some parts of the world (Lucero et al., 2005).
Brucellosis is readily transmissible to humans through contact with infected material, especially in people who handle infected animals and animal products contaminated with Brucella spp. (Kuplulu and Sarimehmetoglu, 2004). The portal of entry is usually via cuts and abrasions on the skin, and the conjunctiva (Doganay and Aygen, 2003).
Inhalation of aerosols is also considered an important mode of transmission (Chukwu,
1987). Brucella spp. may be food-borne and consumption of contaminated food
72 products of animal origin such as meat, milk, blood, cheese, ice cream, cream and butter may be important sources of infection (Doganay and Aygen, 2003; Kuplulu and
Sarimehmetoglu, 2004).
2.6.4 Pathogenesis
The pathogenesis of human brucellosis is similar to that seen in animals, except that their site of predilection is not the gravid uterus. Brucella spp. infect both phagocytic and non-phagocytic cells (Doganay and Aygen, 2003). An initial bacteraemia results in spread of the bacterium to regional lymph nodes, joints, testes in males and several unspecific sites. The occurrence of recurrent periodic bacteraemia results in further dissemination of the microorganisms to other areas (Doganay and Aygen, 2003).
Chronic infection often results in the development of granulomata consisting of epithelioid cells, polymorphonuclear leukocytes, lymphocytes and some giant cells
(Doganay and Aygen, 2003).
2.6.5 Immunity
An infection with Brucella spp. induces both humoral and cell-mediated immunity and the latter appears to be the principal mechanism of recovery (Doganay and Aygen,
2003). The humoral response is characterised by the production of both the IgM and
IgG in sequence akin to that in cattle. After antibiotic treatment, the titres gradually come down, with a faster decrease of IgG antibodies than IgM antibodies (Doganay and
Aygen, 2003).
73
2.6.6 Clinical signs
Clinical signs of human brucellosis vary, and are often non-specific. Commonly observed signs include recurrent fever, weakness, depression, low libido, sweating, arthralgia, lethargy, enlarged lymph glands, joint pains, anorexia, nausea, vomiting and weight loss (Chukwu, 1987; Corbel, 1997; Doganay and Aygen, 2003; Lucero et al.,
2005). Abortion in pregnant women is not a common feature of human brucellosis, most probably due to the absence of growth stimulants for Brucella spp. in the gravid uterus.
2.6.7 Diagnosis
The most specific diagnostic test for human brucellosis is the culture and isolation of the causative microorganism. The blood broth culture in 10% CO2 is the simplest and most often utilised bacteriologic procedure (Diaz and Moriyon, 1989). Although the success rate is considered to be variable, three blood cultures drawn over a 24 hour period, particularly from febrile patients are generally sufficient (Diaz and Moriyon,
1989). In the case of focal complications, culture material, if possible should be collected from the affected places such as liver, lymph node, abscess, synovial fluid or cerebrospinal fluid (Doganay and Aygen, 2003).
Although the only definitive diagnostic test is bacteriologic isolation of the causative microorganism, cultures are not always positive, and serological methods must be used as indirect proof of the diagnosis (Diaz and Moriyon, 1989). Several serologic tests such as the SAT, 2-Mercaptoethanol test (TMET), RBPT, the anti-Brucella Coombs test, CFT and ELISA have been used successfully to detect antibodies against Brucella spp. (Diaz and Moriyon, 1989). Serological cross reactions can be seen between
74 infections caused by Brucella spp. and Y. enterocolitica O:9 infection due to sharing of epitopes in the two micoorganisms (Doganay and Aygen, 2003). A c-ELISA offers advantage because in addition to having a higher sensitivity, its specificity is superior to the conventional tests because the assay uses a monoclononal antibody specific for a polysaccharide on the smooth lipopolysaccharide molecule of Brucella spp. (Lucero et al., 2003).
2.6.8 Treatment and control
Antibiotic therapy of human brucellosis is difficult and requires prolonged continuous treatment. This is probably explained by the occurrence of intermittent bacteraemia and that Brucella spp. are obligate intracellular micro-organisms.
Different treatment regimens have been proposed (Ariza et al., 1985; Akova et al.,
1993; Corbel, 1997), but the treatment recommended by the World Health Organisation
(WHO) is rifampin 600 to 900 mg and doxycycline 200 mg daily for a minimum of six weeks (Doganay and Aygen, 2003). However, a combination of intramuscular streptomycin and oral tetracycline gives fewer relapses than the rifampin-doxycycline combination (Ariza et al., 1985; Shin and Carmichael, 1999). Thus the streptomycin- tetracycline regimen may be preferred for the treatment of human brucellosis. On the other hand quinolones in combination with rifampin have been found to be as effective as the streptomycin-tetracycline combination (Akova et al., 1993). Infections with complications such as neurobrucellosis or endocarditis are treated with a combination therapy with rifampin, doxycycline and ceftriaxone for 2-3 weeks and yield satisfactory results (Doganay and Aygen, 2003).
75
It is generally recognised that the prevention of human brucellosis is best achieved by control or eradication of the disease in animals, combined with adequate heat treatment of potentially contaminated food products (Schurig et al., 2002). In addition, a lot is achieved through education campaigns. Pasteurisation of milk removes the risk of spread through milk and milk products (Kuplulu and Sarimehmetoglu, 2004).
There are no safe and effective vaccines for use in preventing human brucellosis. A derivative of B. abortus S19, 19-BA, given intradermally by scarification was tried, but gave limited protection for a relatively short duration (Schurig et al., 2002). Therefore, control and eradication of animal brucellosis appears to be the only practical means of eliminating human brucellosis.
2.7 Brucellosis of Wildlife and Marine mammals
2.7.1 Aetiology
2.7.1.1 Wildlife brucellosis
B. abortus has been documented to be the major cause of brucellosis in wild animals such as the African buffalo (Syncerus caffer), Eland (Taurotragus orynx), Waterbuck
(Kobus elipsiprymnus), the lechwe (Kobus leche kafuensis) in Southern Africa and the bison (Bison bison) in North America (Condy and Vickers, 1972; Coetzer et al., 2000;
Gall et al., 2000; Muma et al., 2006).
Elsewhere in Europe, B. suis biovar 2 has been associated with infections in wild pig boars (Sus scrofa) (Godfroid et al., 1994; Garritty et al., 2005). B. suis biovar 4 has been isolated in the reinderr and while B. suis biovar 5 has been isolated from rodents.
76
2.7.1.2 Brucellosis of marine mammals
Unidentified Brucella spp. were found to cause infections in marine mammals ( Ross et al., 1994; Foster et al., 1996; Ross et al., 1996). Therefore, an expansion of the six
Brucella spp. has been proposed to include one (B. maris) (Garritty et al., 2005) or two
(B. pinnipediae and B. cetaceae) new species to categorize these strains (Jahans et al.,
1997). The new approved names of these Brucella spp. were B. pinnipedialis and B. ceti respectively (Foster et al., 2007). The proposal to assign these isolates into two new species, was based on animal host of origin as well as differences in CO2 –dependence, primary growth on Farrell’s medium and their metabolic activity on galactose (Foster et al., 1996; 2002). B. ceti include those strains isolated from cetacean mammals (whales, porpoises and dolphins) while B. pinnipedialis include strains from pinnipeds (seals and sea lions) (Foster et al., 2002; 2007). The total number of Brucella spp. has been expanded to nine, with the recent isolation of B. microti from voles (Microtus arvalis) from Europe (Scholz et al., 2008).
2.7.2 Epizootology
2.7.2.1 Wildlife brucellosis
In Southern Africa, the isolation of B. abortus has been confirmed in the buffalo
(Syncerus caffer), the waterbuck (Kobus elipsiprymnus) and the eland (Taurotragus oryx) (Condy and Vickers, 1972; Coetzer et al., 2000). Serum antibodies to Brucella spp. have been detected in a number of wildlife species, both captive and free-living.
Condy and Vickers (1972) carried out sero survey on 28 species and a total of 2320 individual animals (free living) wild animals in Zimbabwe. Of these 159 reacted positively to SAT, with the highest cases from zebras (Equus burchelli) (24%), eland
(Taurotragus oryx) (16.9%), buffalo (Syncerus caffer) (14.7%) impala (Aepyceros
77 melampus) (9.7%). High seroprevalence in buffaloes and wildebeests (Conncahaetes taurinus) was later confirmed by Madsen and Anderson (1995).
In Zambia, Rottcher (1978) examined a total of 439 sera from 37 species using SAT and recorded 6.5% seropositive reactors in Kafue lechwe antelope (Kobus leche kafuensis),
9.6% in puku antelope (Kobus vardonii), 8.5% in wildebeest (Connchaetes taurinus),
10.5% in buffalo (Syncerus caffer), 8.6% in bushbuck (Tragelaphus scriptus), 18% in zebra (Equus burchelli), 12.5% in impala (Aepyceros melampus), and 16% in eland antelope (Taurotragus oryx). These findings were later confirmed, Pandey et al., (1999).
Elsewhere on the African continent, brucellosis has been shown to be prevalent in areas of wildlife-livestock interaction (Sachs et al., 1968, Nicoletti, 1980, Jiwa et al., 1996;
Muma et al., 2006). Therefore in these interface areas, these wild animals may be important reservoirs of brucellosis of domestic animals. However, the possibility of a bimodal transmission of B. abortus is possible with either the domestic animals or wildlife acting as reservoirs of B. abortus.
2.7.2.2 Brucellosis of marine mammals
There are three groups of mammals adapted to life in sea water; the cetaceans (whales, dolphins and porpoises), sirenians (dugongs and manatees) and pinnipeds (seals, sea lions and walruses). Several reports have documented the isolation and characterisation of Brucella spp. or serologic evidence of infection in a wide variety of marine mammals from such areas as the Pacific, Atlantic and Arctic oceans around Canada and the
United States of America, the Scottish coasts, North Eastern England, the North Sea, and the Barents Sea ( Ross et al., 1994; 1996; Foster et al., 1996; 2002; Nielsen et al.,
2001; Godfroid, 2002). These Brucella spp. were isolated from cetaceans; the
78 bottlenose dolphins (Tursiops truncates), Atlantic white-sided dolphins
(Lagenorhynchus acutus), striped dolphins (Stenella caeruleoalba), common dolphins
(Delphinus delphis), harbour porpoises (Phocoena phocoena) and a minke whale
(Balaenoptera acutorostrata), pinnipeds; common seals (Phoca vitulina), hooded seals
(Cystophora cristata), grey seals (Halichoerus grypus) and a European otter (Lutra lutra) (Foster et al., 1996; Ross et al., 1996; Jahans et al., 1997; Clavareau et al., 1998;
Bricker et al., 2000;;; Foster et al., 2002; Godfroid, 2002).
Although the incidence and distribution of infections attributable to these Brucella spp. remains to be determined, serological results and bacterial isolations may suggest a wide distribution in a variety of marine mammal species (Foster et al., 2002; Ross et al.,
1996).
2.7.3 Pathogenesis and clinical signs
2.7.3.1 Wildlife brucellosis
It is likely that the pathogenesis of brucellosis in wild life resembles that in cattle, because the clinical presentation suggests the occurrence of both the acute and chronic infections. The production of erythritol by the gravid uterus of ungulates is likely to stimulate massive growth of brucellae, and this is likely to result in the occurrence of abortions (Walker, 1999). In buffaloes, the occurrence of abortions (acute brucellosis) and carpal hygromas and orchitis in bulls have been reported (Coetzer et al., 2000).
2.7.3.2 Brucellosis of marine mammals
The significance of the presence of Brucella spp. in marine mammals is largely not known. The isolation of these Brucella spp. in different visceral organs (Foster et al.,
79
1996) might be indicative of systemic dissemination of infection. Although their pathogenicity remains speculative these Brucella spp. might display pathogenic mechanisms similar to the other brucellae affecting terrestrial mammals. Abortion and meningoencephalitis have been reported in dolphins, albeit in the majority of cases, the animals did not show evidence of clinical disease (Godfroid, 2002).
2.7.4 Diagnosis
2.7.4.1 Wildlife brucellosis
The diagnosis of brucellosis has bee accomplished by using conventional culture techniques (Condy and Vickers, 1972) and serological tests (Madsen and Anderson,
1995; Pandey et al., 1999; Gall et al., 2000). Serological tests that have been evaluated for use in wildlife include the SAT, RBT, c-ELISA, FPA, CFT and i-ELISA (Mdasen and Anderson, 1995; Pandey et al., 1999; Gall et al., 2000). However, the shortcomings of some of these tests encountered in the diagnosis of bovine brucellosis are likely to be experienced in wildlife. Hence the choice of which test to use will be deterimined by the test performance indices, application adaptability and the costs of setting up the test.
2.7.4.2 Brucellosis of marine mammals
The isolation and identification of Brucella spp. from several organs such as lymph nodes, uterus, testes, spleen, mammary glands and cutaneous lesions (Foster et al.,
1996), has yielded definitive evidence of infection in marine mammals.
Serological evidence of exposure to Brucella spp. in a variety of species of these marine mammals has been demonstrated successfully using serological tests such as the RBT,
SAT, i-ELISA and the c-ELISA ( Ross et al., 1996; Nielsen et al., 2001).
80
2.7.5 Control
2.7.5.1 Wildlife brucellosis
The control of brucellosis in wild life is difficult to implement because of the nature of the life style of some of these animals. It is likely that in gregarious species such as buffaloes and wildebeests, brucellosis could easily be mainatained in their populations because of the existence of large herds. The only feasible method is to limit their contact with domestic animals through erection of game fences.
Although vaccination has been used extensively in the control of bruccelosis in domestic animals, it has not been used to control wildlife brucellosis (Davis and Elzer,
2002). The problems associated with use of vaccines in wildlife include, the existence of large herds of free-living animals, and none of the live vaccines currently used in livestock have been fully validated for safety and efficacy in wildlife (Davis and Elzer,
2002). Hence, parallel effort has to be made to find suitable vaccines for wildlife and at the same time to find cost effective and safe systems to deliver the vaccines.
2.7.5.2 Brucellosis of marine mammals
Little is known about the routes of transmission of marine Brucella spp.; their pathogenicity and epidemiology. Similarly, the risk of infection posed by these Brucella spp. to terrestrial animals is not known. As in other wild animals, control measures are difficult to implement. However, since these Brucella spp. pose a zoonotic risk to humans (Sohn et al., 2003), caution and good hygienic practice is advisable to people who handle stranded marine mammals or meat from these animals (Nielsen et al.,
2001).
81
CHAPTER III
MATERIALS AND METHODS
3.1 Prevalence of antibodies to Brucella species in individual dairy cattle and herds from smallholder farms in Zimbabwe
3.1.1 The study population
Cattle are animals of extreme importance for the people of Zimbabwe. In 2001,
Zimbabwe had an estimated 6 million cattle, and 60-80% (3 million) of these were in the smallholder farming sector (Anon., 2001) (Table 3.1). The smallholder cattle farms keep small herd sizes, usually less than 100 animals per herd. The latest census figures from the Department of Veterinary Services estimated that in 2001,the selected study areas had a combined cattle population of 680 400 (Anon., 2001). Cattle breeds kept in these areas included pure exotic dairy, pure indigenous beef and crosses of beef and dairy (both Bos indicus and Bos Taurus cattle).
Table 3.1. Cattle population of Zimbabwe by farming sector from 1991 to 2001, given in thousands (% of all cattle)*.
Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Communal cattle 4271 3805 3426 3918 3380 3470 3574 4108 4389 4526 5127
(71.4) (69.9) (70.2) (73.9) (71.7) (72.3) (74.5) (72.5) (72.3) (73.2) (79.7)
Commercial cattle 1714 1642 1451 1383 1331 1329 1226 1560 1680 1660 1305
(28.6) (30.1) (29.8) (26.1) (28.3) (27.7) (25.5) (27.5) (27.7) (26.8) (20.3)
Total 5985 5447 4877 5301 4711 4799 4800 5668 6069 6186 6432
* (Anon., 2001)
82
3.1.2 The study areas
The study was conducted in Gokwe south, Marirangwe, Mushagashe, Nharira, Rusitu
Valley and Wedza smallholder dairy cattle farming areas located in different agro- ecological regions of Zimbabwe (Fig. 1). The details of these study areas are outlined
(Table 3.2). These areas were specifically selected for this study because they; (1) were representative areas of the five different agro-ecological regions of Zimbabwe, (2) kept mixed cattle breeds of Bos taurus (originally from commercial farms) and Bos indicus
(indigenous Zebu) origin, (3) were members of the smallholder dairy schemes in their areas, (4) were not vaccinating against brucellosis using B. abortus S19 vaccine.
Gokwe south, from the western part of Zimbabwe, is an arid area; Mushagashe, a semi- arid area and both are characterised by low, erratic and ephemeral rainfall (<600mm per annum); Marirangwe, Wedza and Nharira, moderately wet areas (600-800mm per annum) and Rusitu Valley, a highland wet area (800-1200mm per annum).
The predominant type of agriculture in these smallholder areas is small-scale crop- livestock farming on small farms (median farm size 10ha). Prior to 1980, the areas of
Marirangwe, Mushagashe and Nharira existed as self-contained small-scale farms
(formerly designated as African Purchase Areas), mixed with some communal areas and practising semi-intensive crop-livestock agriculture, while Gokwe, Rusitu and Wedza were communal areas (formerly designated as Tribal Trust Lands) practising subsistence crop-livestock agriculture. All the study areas except Gokwe established smallholder dairies in the early 1980s (Madsen, 1989). In Gokwe, smallholder dairies were introduced in 1991. The establishment of the smallholder dairies was done through the Dairy Development Programme (DDP), a division of the parastatal company, the
Agricultural Rural Development Authority that provided funds for construction of milk
83 collection centres, and the purchase of requisite equipment and Bos taurus breeds of dairy cattle from commercial farms. These exotic cattle were mixed and subsequently cross-bred with the indigenous Zebu (Bos indicus) breeds. At the time of the study, most of the farms kept small cattle herd sizes (median herd size 15 cattle).
The cattle management type as prescribed by DDP was generally similar for all the study areas that were involved in the smallholder dairy schemes. This involved grazing of cattle on separate pastures with own supplies of drinking water. Therefore, unlike other smallholder farms in communal areas where there is a lot of commingling of cattle, making the definition of a herd difficult under these conditions of management, the type of cattle management in our study areas permitted us to regard the individual farms as independent herds. A farm was classified as a piece of land allocated to a single house hold for farming purposes and were demarcated from others by perimeter fencing.
84
Fig. 3.1. The location of study areas in relation to the agro-ecological regions of
Zimbabwe. Most of commercial farms are found in regions 1-3 that are characterised by good soils and moderate to high rainfall (800-1200 mm/ year). The smallholder farms are mostly located in regions 3-5 that are characterised by low, erratic and ephemeral rainfall (<600 mm/ year).
Table 3.2 Details of geographical locations and climatic conditions of study areas
(adapted from Chenje et al., 1998).
District of Study Area Agro-ecological zone Mean temperature (°C) Annual origin (Range) rainfall (mm)
Gokwe South Gokwe III- IV 26 (18-37) 450-750
Chikomba Nharira III 22 (18 -34) 550-750
Seke Marirangwe IIa 21 (16-30) 750-1000
Wedza Wedza IIb 22 (16-33) 750-1000
Masvingo Mushagashe IV 25 (20-35) 450-650
Chipinge East Rusitu I 15 (10-26) 1000-1200
85
3.1.3 The study design
A cross sectional study was carried out using a stratified sampling procedure to select herds (farms) and individual animals per herd. Herds were regarded as the primary sampling units and stratified according to the study area in order to control for selection bias.
3.1.4 Selection of herds and animals
A tentative list of the approximate number of herds in each study area was obtained from the local veterinary or agricultural office. Herds that were recruited for the study were randomly selected from the list of farms in each study area with an adjustment for a finite population as described (Dohoo et al., 2003) (Fig. 3.2).
1. Sample size for estimating proportions or means from an infinite population
2 n = Zα pq ------L2 2. Sample size for estimating proportions or means from a finite population
n¹ = 1 ------1/n+1/N Zα = the value of Zα required for confidence = 95% (1.96) p = a priori estimate of the disease prevalence q = 1-p L = the precision of the estimate (the allowable error), eg. 5% or 10% N = the size of the population
Fig. 3.2. Formulae used to calculate sample sizes of cattle herds and individual animals for estimation of mean prevalence of brucellosis
The prevalence of the disease was estimated to be approximately within 10% of the true value with 95% confidence, assuming that the inter-herd and individual animal prevalence of brucellosis was 25% and 15%, respectively (Madzima, 1987; Madsen,
1989). The estimated sample size was 179 smallholder dairy farms based on the above
86 assumptions, although 203 smallholder dairy farms were eventually sampled. Only herds with a minimum of ten animals were eligible for the study. As suggested by
Muma (Muma et al., 2006), herds that were reared in close proximity of each other were grouped together and considered as one.
The optimum number of cattle to be sampled from each herd was decided based on the average herd size, the estimated individual animal disease prevalence and the combined sensitivity and specificity of a serial testing using Rose Bengal plate test (RBT) and competitive enzyme immunosorbent assay (c-ELISA). Using the estimated individual values for sensitivity and specificity of these tests (McGiven et al., 2003; Nielsen et al.,
1995; Samartino et al., 1999), the serial sensitivity of 88.2% and specificity of 99.8% were determined as suggested (Noordhuizen et al., 1997). A minimum of 24 animals per herd was assumed. Since the herd sizes were anticipated to be small in these farms and to balance on the resources available, it was decided to sample at least 8 cattle from each herd and a 25% sampling fraction from herds with > 40 cattle. The sampling fraction would give herd sensitivity and specificity of approximately 89.7% and 100% respectively when the cut-off point for classifying a herd as positive is at least single positive reactor animal to Herdacc™ (Jordan, 1995). Real random sampling of individual animals within selected farms was not achievable due to absence of animal restraining facilities and and also because some aanimals are semi-wild and ard not accustomed to routine restraining procedures.
Each farm was visited once for sampling animals and data collection. The sampling was either done on farm or at communal dip tanks used for dipping cattle for control of ticks and tick-borne diseases. The animals were randomly selected based on the
87 identifications supplied on the stock cards. They were confined in pens or crushes, and further restrained using casting ropes. To minimise the number of false positive reactors due to maternal antibodies, only animals that were at least two years old at the time of the study were eligible for sampling. From each sampled animal, 15-20ml of blood was collected from a jugular or the cocceagial (tail) vein. The blood was left to clot at ambient temperature (25-30°C) for about 25 minutes and then kept on ice (~ 4ºC) and transported for storage at –20°C until use.
3.1.5 Data collection
Information on individual animal variables such as age, sex, origin of the animal
(locally raised or purchased) was recorded separately for each sample on sample submission forms.
3.1.6 Laboratory Tests
Clotted blood samples were centrifuged at 3000 x g for 15 minutes. 2 ml of each serum sample was collected into cryo-tubes and stored at -20°C. Before testing, sera were allowed to equilibrate to room temperature (25-27°C).
3.1.6.1 Rose Bengal plate test (RBT)
Rose Bengal test (RBT) performed as described by (Alton et al., 1975) was used to screen all samples for antibodies to Brucella spp. The test was performed on round- bottomed welled pyrex plates. Generally, 25µl of the serum were mixed with equal amounts of stained Rose Bengal antigen (pH 3.65). The samples were mixed on a rocker for five minutes. Results were read visually and recorded as positive if visible
88 agglutinations were present and negative if not. The antigens, positive and negative controls used were procured from Central Veterinary Laboratory, Weybridge, UK.
3.1.6.2 The competitive ELISA (c-ELISA)
All sera positive on RBT were further tested using the SvanovirTM Brucella-Ab c-
ELISA test kits (Svanova Biotech, Uppsala, Sweden). The c-ELISA was performed using the SvanovirTM Brucella-Ab c-ELISA test kits (Svanova Biotech, Uppsala,
Sweden), used according to the manufacturer’s instructions. Briefly, the test was carried out in 96 well polystyrene plates (Nalge Nunc, Denmark) that were pre-coated with Brucella spp. lipopolysaccharides (LPS) antigen. Serum diluted 1:10 was added to each well, and immediately followed by equal volumes of pre-diluted mouse monoclonal antibodies specific for a common epitope of the O-polysaccharide (OPS) of the smooth LPS molecule. The reactivity of the mouse monoclonal antibody was detected using goat antibody to mouse IgG that was conjugated to horseradish peroxidase. Hydrogen peroxidase substrate and ABTS chromogen were developed for
10 minutes. The reaction was stopped using 1M H2SO4. Optical densities were read at
450nm using a Titertek Multiscan® PLUS reader (Flow Laboratories, UK). All controls were run in duplicates. Antibody titres were calculated as percentage inhibition (PI) defined by the ELISA kit manufacturer as;
(Mean OD value of sample or control PI = 100 − ______x 100 (Mean OD value of conjugate control (cc)
Sera were calssified as positive and negative according to the manufacturer’s recommendations. All sera with PI values of ≥3 0% were classified as positive while those <30% were negative.
89
In a serial testing protocol, a serum was considered positive for antibodies to Brucella spp. if it was positive for both the RBT and c-ELISA.
3.1.7 Statistical analysis
The recording and editing of data was performed using Microsoft Excel®. In order to improve the estimation of the prevalence of antibodies to Brucella spp., calculation of sampling weight adjustments which take into consideration the probability of selection into the final sample, for each animal from the different strata (area and herd of origin) were performed as described (Dohoo et al., 2003). Statistical analyses were performed using Stata SE/9.0 for Windows (Stata Corp. College Station, TX, USA). Specifically, the Stata survey (svy) command which takes into account the sampling weights was used to calculate individual sero-prevalence estimates according to the study areas, sex and age categories.
For the herd-level data-base, calculations of sampling weight adjustments for the different strata (herds and study areas) were done using Microsoft Excel ® as described
(Dohoo et al., 2003). The proportion of infected cattle herds for each study area was determined using the survey comand of Stata. The effect of herd size, farm size and stocking density on the distribution of Brucella seropositive cattle herds was determined by manual calculations in Stata.
90
3.2 The risk factors for infection with Brucella spp. in individual cattle and
herds from smallholder farms in Zimbabwe
3.2.1 Selection of herds and animals
The details of the criteria for selection of study areas, their geographical location and climatic conditions have been described in sections 3.1.1 to 3.1.2 (above). The same farms that were selected for prevalence studies were also selected for these risk factor studies. The details of their selection criteria and sampling of both farms and individual animals were presented in section 3.1.3. Similarly, the details of serological testing for the presence of antibodies to Brucella spp. were presented in section 3.1.4.
3.2.3 Epidemiological data collection
Herd level data that included: herd structure, size, history of purchases of animals and farm management practices were collected using an interviewer-administered structured questionnaire (Appendix 1). Based on previous reports of brucellosis, and on biological relevance, only information on important potential risk factors that are believed to influence the spread of Brucella spp. infections between and within cattle herds were collected. A total of 17 herd-level predictor variables that included: herd size, farm size, stocking density, management practice of keepeing mixed cattle breeds or pure breeds, type of grazing (communal or own pasture), source of drinking water, method of tick control (communal or own dip tank), presence of a purchased animal, purchase of animals in the past three years, keeping cattle together with sheep and goats, use of a calving pen, keeping cattle in confined pens at night, hiring animals from neighbours for use, method of breeding (natural or artificial), source of bull for breeding (own or hired), keeping records and knowledge of bovine brucellosis were included on the
91 questinnaire. The herds selected using the criteria in sections 3.1.1 to 3.1.3 (above) were visited once between 2004 and 2005. The inteview was conducted to the owner or keeper of the cattle herd.
3.2.4 Serological testing
Serology on individual cattle sera collected was carried out as described in Section
3.1.5.
3.2.5 Statistical analysis
3.2.5.1 The logistic regression analysis
The logistic regression analyses were conducted to investigate both the individual animal- and herd-level risk factors for infection with Brucella spp. For the individual animal level logistic regression analysis, the Brucella-seropositive status of cattle
(negative = 0, seropositive =1) was used as the outcome variable with age and sext as the predictor variables. For herd-level data, two separate of logistic regression models were built. The first model, the outcome was the binomial herd-level data, Brucella seropositive status (no, seropositive animals = 0, ≥1; yes, seropositive animals =1) and the explanatory variables identified in univariable analysis. The second model, a negative binomial regression analysis was used on the total number of cattle per farm that were positive for antibodies to Brucella spp. as the outcome variable and the number of cattle tested as the population at risk (exposure variable). The model is suitable for assessing associations between risk factors and level of infection within herds.
The predictor variables to be included in the regression models were selected in univariable anlysis using the Pearson’s χ2 or Fisher’s exact tests for teting the
92 association between the outcome and and potential categorical risk factors, while a
Kruskal Wallis test was used for continuous variables. The predictor variables were assessed for collinearity by cross tabulations using the two sided Fisher’s exact test.
Only variables that had a P-value <0.20 in univariable analysis were offered to the regression model. Fisher’s exact test was used to test for the association between the
Brucella-seropositive status of cattle (negative = 0, seropositive =1) and potential categorical risk factors, while a Kruskal Wallis test was used for continuous variables.
The continuous variables were tested for collinearity using correlation analysis and scatter plots while categorical variable were cross tabulated using the two-tailed Fischer exact test. Variables considered to be of biological relevance with a P-value <0.2 from univariable analyses were further tested in multivariable regression analyses.
The predictor variables were assessed for collinearity by cross tabulations using the two sided Fisher’s exact test.
The models were manully constructed using a forward selection process as described by
(Dohoo et al., 2003). The logistic regression models were evaluated for goodness-of-fit using Hosmer-Lemeshow test while their sensitivities and specificities determined using receiver operating characteristic (ROC) curves (Dohoo et al., 2003). The negative binomial model was assessed by comparing the predicted outcome to the observed values.
The interpretation of the coefficients, odds ratio and the intercept of the regression models were done as suggested by Dohoo and co-workers (Dohoo et al., 2003). In
93
general, the coefficients (β1) of dichotomous predictors represent the the amount that the log odds of a disease increase (or decrease) when the factor is present. These can easily be converted into odds ratios by exponentiating the coefficients. The odds ratios
(OR) are much simpler to interprete as compared to coefficients.
For a continuous predictor, the coefficient (β1) represents the changes in the log odds of disease for one unit increase in the predictor. The OR represents the factor by which the odds of disease are increased (or decreased) for each one-unit increase in the predictor.
For categorical predictors, the coefficient for each indicator variable (categorical predictors are represented as indicator variables), represents the the effect of that level compared to the category (baseline) not included in the model.
The intercept represents the logit of the probability of disease if all of the “risk factors”
(predictors) are absent (ie. equal to zero). This can be expressed as:
β1 = ln ( p0 ) ------1-p0)
Where in cross-sectional or cohort studies, p0 equals the probability of diasese in the non-exposed group (Dohoo et al., 2003).
94
3.3 Prevalence and risk factors for abortions in cows from smallholder farms
naturally infected with Brucella species
3.3.1 Study areas
The details of the selection of study areas, the sampling plan, sampling of herds and individual animals were described in previous sections (3.1.1 to 3.1.3).
3.3.2 Epidemiological data collection
Herd level data that included: herd structure, size, history of purchases of animals and farm management practices were collected using an interviewer-administered structured questionnaire (Appendix 1) as detailed in section 3.2.4.
3.3.3 Serological tests
The processing and testing of serum samples was described previously (Section 3.1.5).
For determination of antibody titres, all sera that were positive on both the RBT and c-
ELISA were further tested using the serum agglutination test (SAT), which was carried out as described by Brinley Morgan, (1967). The antigens for the SAT were obtained from Onderstepoort Veterinary Institute, South Africa. The classification of the sera into doubtful, low and strong reactors was done according to the Zimbabwean norms as outlined by Madsen (1989).
3.3.4 Data analysis
Data from questionnaires on herd-level risk factors from questionnaires and laboratory analyses were stored in a computer data-base and calculations of sampling weight adjustments for the different strata (herds and study areas) were done using Microsoft
95
Excel ® as described (Dohoo et al., 2003). Statistical analysis was performed using
Stata/ SE 9.0 for Windows (StataCorp. College Station, Texas, USA). The proportions of individual animals and herds that recorded abortions in the past three years were determined for each study area using survey command of Stata, with adjustments for primary sampling units and strata as described (Dohoo et al., 2003).
3.3.4.1 Logistic regression analyses
Pearson’s χ2 or Fisher’s exact tests were used to test the significance of the association between individual animal or herd’s abortion status and categorical risk factors in univariable analyses. Continuous variables were assesed by the Kruskal Wallis test.
Variables we considered to be of biological relevance and with P-values <0.2 were offered to the multivariable logistic regression models. Variables were checked for collinearity using correlation analysis and scatter plots.
Two separate regression models were built. For individual animals, the effect of age and
Brucella sero-positive status (0 = negative, 1 = positive) on abortions were investigated in a logistic regression model using the abortion status (0 = no, 1= yes) as the dependent variable. A second multivariable logistic model was built to analyse the effect on herd abortions, of the various herd-level risk factors such as stocking density, type of breed
(mixed or pure) and livestock management practices. The two models were manually constructed by forward-selection as described by (Dohoo et al., 2003). The logistic regression models were evaluated for goodness-of-fit using Hosmer-Lemeshow test while their sensitivities and specificities were determined using ROC curves (Dohoo et al., 2003).
96
3.4 Comparison of the Fluorescence polarisation assay with the Rose Bengal
test and the competitive ELISA for the serological diagnosis of bovine
brucellosis in smallholder cattle farms in Zimbabwe
3.4.1 Sera and whole blood
Serum (n = 789) and whole blood (n = 351) were randomly collected between May
2004 and November 2005, from cattle from smallholder farms in Gokwe South,
Marirangwe, Mushagashe, Nharira, Rusitu valley and Wedza. The details of the selection of these study areas, the sampling pland and sampling of herds and individual animals have been described (Section 3.1.1 to 3.1.3).
3.4.2 Serological tests
3.4.2.1 Rose Bengal plate test (RBT)
The RBT was performed as described previously (Section 3.1.5.1).
3.4.2.2 The competitive ELISA (c-ELISA)
The c-ELISA was performed as described previously (Section 3.1.5.2
3.4.2.3 The Fluorescence polarisation assay (FPA)
The fluorescence polarisation assay was carried out essentially as described by Nielsen et al., (1996a). B. abortus S1119.3 O-polysaccharide hydrolysed to an average molecular weight of 22 kDa and conjugated to fluorescein isocyathionate was used as the antigen. Briefly, 2ml of 0.0M Tris-HCl buffer, pH 7.2 incorporating 0.15 M NaCl,
15mM EDTA and 0.05% Igepal NP40 were added to 10 x 75 mm borosilicate culture tubes (VWR™) and 20µl of serum or 40µl of whole blood were added to the buffer,
97 mixed well and incubated at room temperature (25-27 ○C) for about 15 seconds. The background activity of the serum or whole blood sample was measured using a fluorescence polarization analyzer (Diachemix ® LLC, Milwaukee, WI53202, USA).
Further, 10µl of antigen were added and a second absolute reading was obtained after incubating at room temperature for a minimum of 2 minutes. The machine automatically subtracts the first reading from the second to give a net reading of roational time which is converted into millipolarisation (mP) values. The higher the mP units, the longer it took for the molecule to travel through the angle, indicating an antibody-antigen reaction. Data was expressed as milipolarisation (mP) units. The instrument was calibrated according to manufacturer’s specification using reference sera (Viral Antigens, Inc, Memphis, USA).
3.4.3 Data analysis
The recommended cut-off value of 90 mP (Nielsen et al., 1996a) was initially used for the FPA. The data were recorded as positive or negative based on the c-ELISA results, which was used as a reference standard. Data were plotted in receiver operating characteristics (ROC) curves to determine suitable cut-off value using Stata 9.0 software (Stata Corp. College Station, TX, USA). The cut-off value of 91.0 mP was determined assuming that false negative and false positive results were of equal importance.
The sensitivities and specificities of FPA and RBPT relative to the c-ELISA, or FPA relative to the combined serial RBPT/c-ELISA were determined. To assess levels of agreement among test combinations, kappa statistics and McNemar X2 tests were calculated.
98
3.5 Bacteriological investigations of individual cattle by culture and isolation,
and characterization of some Brucella spp. from Zimbabwe by biochemical
profiling and AMOS PCR
3.5.1 Culture and isolation of Brucella spp. from clinical specimens
Culture and isolation of Brucella spp. was made from milk and other samples collected from cattle from smallholder farms in Gokwe, Mushagashe, Marirangwe, Wedza,
Nharira, and Rusitu valley between 2004 and 2005. The selection of cows for the milk samples was not randomised, because it was desired to maximise chances of isolating
Brucella spp. by including all cows with history of abortion.
Milk samples were collected in universal bottles and kept cold at 40C until firther processed. In the laboratory, the milk samplese were centrifuged at 3000 x g for 15 minutes and cultured from the cream and the sediment into Bijou bottles containing
TSB incorporating antibiotic supplements (Oxoid). They were subsequently plated onto onto Farrell’s medium (Oxoid) after 3 dsys’ incubation. In addition, they were cultured on Mueller-Hinton agar for assessment of colonial morphology. All plates were incubated at 37°C under 10% CO2. Brucella spp. were identified using the procedure of
Alton and co-workers (Alton et al., 1988).
3.5.2 Brucella isolates for characterisation using biochemical profiles and AMOS-
PCR
Further Brucella isolates were obtained from the culture collections of both the
University of Zimbabwe Bacteriology laboratory and the Central Veterinary Laboratory
(Table 3.5.1). All isolates were then inoculated into TSB with 5% glycerol, frozen and
99 exported to Norway for further characterisation. Reference Brucella strains were obtained from the culuture collection of the National Veterinary Institute, Oslo,
Norway.
Table 3.5.1 Brucella isolates used in the study collected from different regions of
Zimbabwe.
Field Brucella spp. Reference number Specimen of origin Farm name, type and (Year isolated) place of origin B. abortus B1-2-2676 (1994)a Aborted foetus Mazowe (S), MC B. abortus B4-11-438 (1998)a Aborted foetus Mhuri (S), Gokwe, MD B. abortus B5(1999)a Hygroma Chinamhora (S), MSE B. abortus B6-304 (1997)a Aborted foetus Pilosoff (C) MN B. abortus B7-307 (1997)a Aborted foetus Chikurubi Prisons (C) MSE B. abortus B8-2160 (1996)a Aborted foetus Greyling (C) MW B. abortus B9-2260 (1996)a Aborted foetus Hensman (C) MW B. melitensis B10-6419(1988)a Aborted foetus (goat) Muzarabani (S)MC B. abortus B12-gl-55(?)a Aborted foetus (C), NE B. abortus B14-(2005)b Milk Mulanjeni (S), Gokwe, MD B. abortus B15-H-56(?)a Aborted foetus (C), NE B. abortus B16-494-64 (?)a Aborted foetus (C), NE B. abortus B20- (2006)b Milk Lulaka (S), Gokwe, MD B. abortus B21-93-35 (?)a Aborted foetus (C), NE Reference Brucella spp. B. abortus 1 544 - NVI B. abortus 2 86/8/59 - NVI B. abortus 3 Tulya - NVI B. abortus 4 292 - NVI B. melitensis 1 16M - NVI B. melitensis 3 Ether - NVI B. suis 1 1330 - NVI B. suis 4 40 - NVI B. canis RM-6/66 - NVI B. ovis 63/290 - NVI B. neotomae 5K-33 - NVI aObtained from culture collection from the Central Veterinary Laboratory or University of Zimbabwe bIsotaed during the course of the project (S) = smallholder farm, (C) = Commercial farm, NE= Farm of origin not established. Provinces of origing: MC = Mashonaland Central, MSE = Mashonaland East, MD = Midlands, MW = Mashonaland West, MN = Matabeleland North NVI = National Veterinary Institute, Norway
100
Prior to use, all strains from culture collection were stored either as lyophilised or in a -
80°C deep freezer. Lyophilised isolates were re-constituted and cultured in tryptone soya broth (TSB) (Oxoid) and subsequently sub-cultured onto Farrell’s medium (Oxoid) and assessed for purity on bovine blood agar (Oxoid).
3.5.3 Biochemical profiles
The phenotypic and molecular characterisation tests of the Brucella spp. were done at the National Veterinary Institute in Norway and were performed by the principal author, in compliance with the requirements for handling Brucella spp. The protocol for phenotypic characterisation of Brucella strains was essentially as described by Alton et al. (Alton et al., 1988) albeit with some modifications. Briefly, the strains were checked for production of urease, catalase, oxidase, H2S, indole and sensitivity to thionin
(20µl/m and 40µl/ml) and basic fuchsin (20µl/ml). Further tests for carbon dioxide requirements, H2S and urease production, dye sensitivity, phage sensitivity, and agglutination with A, M and R monospecific antisera were carried out at the Central
Veterinary Laboratory, Weybridge, UK.
3.5.4 Characterisation by AMOS-PCR
3.5.4.1 Extraction of Brucella DNA
The suspected Brucella species were grown on Farrell’s agar (Oxoid) and incubated for
○ 48 hours at 37 C under 10 % CO2. To yield adequate DNA, a few colonies from a pure culture were harvested using a sterile loop, and suspended in 200 µl of sterile distilled water in Eppendorf tubes. A homogeneous suspension was made by stirring with the inoculation loop. Bacterial cells were lysed by heating the tubes at 100 ○C for 10 minutes on a QBT2 heating block (Grant Instruments, UK). To separate the DNA,
101 killed bacterial cells were centrifuged at 15, 700 x g for 10 minutes. The supernatant containing crude DNA template was pipetted into new sterile Ependorf tubes and the sediment discarded. The concentration of the extracted crude DNA was measured using a computerised ND-1000 V3.0 spectrophotometer (NanoDrop Technologies Inc.,
USA). The DNA was stored at –20 ○C until use.
3.5.4.2 The AMOS-PCR
The B. aborus, B. melitensis, B. ovis and B. suis (AMOS) multiplex PCR was done essentially as described previously (Bricker and Halling, 1994) but with minor modifications of the assay environment. Briefly, PCR assay reaction mixture consisted of the following: 1 x PCR buffer (Applied Biosystems), 3mM MgCl2, 200µM (each) of the four deoxynucleotide triphosphates (dNTPs) (Finnzymes Oy, Espoo, Finland), and the 5 sets of primers (0.2 µM each) of B. abortus, B. melitensis, B. ovis, B. suis and
IS711-specific primer (Fig. 3.5.2). One and half unit (1.5 U) of AmpliTaq Gold® DNA polymerase (Applied Biosystems) per 45 µl reaction mixture was added before the reaction mixture was dispensed into MicroAmp vials (Applied Biosystems). A total of 5
µl DNA template of killed bacteria (estimated at 200ng/ml) was added per 45 µl reaction mixture. The PCR was performed with a PTC-200 Peltier Thermocycler
(Roche Molecular Systems Inc, Almelda, USA). Amplification was performed for 35 cycles, each cycle comprised denaturation at 95 ○C for 1 minute and 15 seconds, annealing at 60 ○C for 2 minutes, and extension at 72 ○C for 2 minutes. A final elongation period for 5 minutes at 72 ○C for product extension was provided before storage of samples at 4 ○C. The PCR products were separated by electrophoresis using
1.5 % agarose gel (w/v) (BDH Electran) at 100V for 1.5 hours. Gels were stained with ethidium bromide and photographed using a gene snap camera connected to computerised software (Syngene Pvt Ltd, UK).
102
Table 3.5.2 Sequences of the oligonucleotide primers for the AMOS-PCR
Primer Sequence (5’ -3’)
B. abortus-specific primer GAC-GAA-CGG-AAT-TTT-TCC-AAT-CCC
B. melitensis-specific primer AAA-TCG-CGT-CCT-TGC-TGG-TCT-GA
B. ovis-specific primer CGG-GTT-CTG-GCA-CCA-TCG-TCG
B. suis-specific primer GCG-CGG-TTT-TCT-GAA-GGT-TCA-GG
IS711 TGC-CGA-TCA-CTT-AAG-GGC-CTT-CAT
103
CHAPTER IV RESULTS
4.1 Prevalence of antibodies to Brucella species in individual cattle and herds from smallholder farms in Zimbabwe
4.1.1 Individual animal descriptive results
A total of 1440 animals from 203 farms (median herd size = 17 cattle/herd) from the six study areas were tested for the presence of antibodies to Brucella spp. (seroprevalence)
(Table 4.1.1).
The mean individual animal seroprevalence by age and sex for the six study areas are shown in Table 4.1.2. The highest (12.6%; 95% CI: 3.9%, 21.4%) and the lowest (3.6%;
95% CI: 1.4%, 5.8%) individual animal seroprevalence were recorded from Gokwe and
Wedza smallholder farms respectively, with expressed variability across sex and age groups. Weighting of prevalence estimates was perceived to be necessary to obtain proper population based prevalence estimates.
104
Table 4.1.1. The distribution of farms, individual cattle sampled detailing their categories and age groups
Study area Total number Total number Animals Distribution of ages and categories of of farms of farms sampled animals sampled targeted for actually Age Total sampling sampled Gokwe 29 30 265 2-4 145 smallholder 4.5-5 46 dairy 5.5-7 57 cooperative >7 17 Females 233 Males 32 Marirangwe 29 28 306 2-4 64 small-scale 4.5-5 41 farms 5.5-7 109 >7 91 Females 245 Males 60 Mushagashe 19 15 133 2-4 35 small-scale 4.5-5 30 farms 5.5-7 38 >7 30 Females 122 Males 11 Nharira 39 40 272 2-4 102 smallholder 4.5-5 58 dairy 5.5-7 79 cooperative >7 33 Females 254 Males 18 Rusitu 40 65 354 2-4 136 Valley dairy 4.5-5 96 development 5.5-7 82 cooperative >7 40 Females 338 Males 16 Wedza 23 25 111 2-4 49 centre dairy 4.5-5 27 development 5.5-7 28 cooperative >7 7 Females 107 Males 4 Total 179 203 1440
105
Table 4.1.2. Distribution of Brucella seroprevalence in smallholder dairy cattle (n =
1440) by age, sex and study area calculated using the survey estimators in Stata with prevalence adjusted for primary sampling (2004-2005).
Study area Category Percent seroprevalence (95% CI)
Gokwe All animals 12.6 (3.9, 21.4) 2-4 6.9 (2.2, 11.7) 4.5-5 20.5 (4.3, 36.6) 5.5-7 22.7 (2.9, 42.5) >7 6.2 (0.0, 18.3) Females 13.2 (3.2, 23.1) Males 8.8 (0.0, 20.3) Marirangwe All animals 3.6 (1.7, 5.5) 2-4 9.3 (0.0, 18.7) 4.5-5 5.4 (0.0, 12.7) 5.5-7 1.6 (0.0, 4.2) >7 1.6 (0.0, 3.9) Females 3.4 (1.4, 5.3) Males 4.6 (0.0, 10.0) Mushagashe All animals 5.7 (2.6, 8.7) 2-4 12.1 (2.5, 21.6) 4.5-5 4.3 (0.0, 11.3) 5.5-7 2.2 (0.0, 6.6) >7 4.5 (0.0, 13.4) Females 6.0 (3.1, 9.0) Males 0.0 (-) Nharira All animals 6.1 (2.9, 9.3) 2-4 6.2 (1.9, 10.6) 4.5-5 9.5 (0.0, 20.3) 5.5-7 5.9 (0.4, 11.3) >7 0.0 (-) Females 4.6 (2.0, 7.2) Males 23.5 (0.0, 47.1) Rusitu All animals 3.6 (1.4, 5.8) Valley 2-4 6.8 (2.1, 11.5) 4.5-5 0.9 (0.0, 2.8) 5.5-7 3.3 (0.0, 7.0) >7 0.0 (-) Females 3.4 (1.2, 5.7) Males 7.8 (0.0, 19.3) Wedza All animals 2.3 (0.0, 5.3) 2-4 2.7 (0.0, 7.9) 4.5-5 0.0 (-) 5.5-7 4.6 (0.0, 12.9) >7 0.0 (-) Females 2.4 (0.0, 5.4) Males 0.0 (-) Mean Total 5.6 (4.4, 6.8)
106
4.1.2 Herd-level descriptive results
Of the 203 smallholder cattle farms that were sampled, 52 (25.0%; 95% CI: 18.1,
31.9%) were positive for antibodies to Brucella spp. There was a marked variation in herd level Brucella seroseroprevalence among the study areas (Table 4.1.3).
Table 4.1.3. The sampling weight-adjusted seroprevalence of antibodies to Brucella spp. of cattle from smallholder farms in the respective study districts of Zimbabwe
(2004-2005).
Study area Total farms Total positive Percent positive
sampled farms farms (95%CI)
Gokwe 30 12 39.9 (19.7, 60.1)a
Marirangwe 28 10 30.5 (12.0, 49.0)a
Mushagashe 15 6 53.3 (21.1, 85.0)a
Nharira 40 13 31.1 (15.1, 47.1)a
Rusitu Valley 65 9 14.1 (5.2, 23.1)b
Wedza 25 2 9.3 (0, 22.0)b
Total 203 52 25.0 (18.1, 31.9)
Key: a, b Shows study areas with significantly different seroprevalence P<0.05. Study areas with the same asterisks were not statistically different (P>0.05).
The highest (53.3%; 95% CI: 21.1%, 85.0%) and the lowest (9.3%; 95% CI: 0%, 22.0%) herd-level seroprevalence were recorded in Mushagashe and Wedza respectively.
107
4.2 The risk factors for infection with Brucella spp. in individual cattle and herds from smallholder farms in Zimbabwe
4.2.1 Individual animal risk factors
Individual animal factors such as sex and age had mixed effects on the odds of the animals testing positive for antibodies to Brucella spp. Although there were more males that tested positive for antibodies to Brucella spp., in univariable analysis, this variable was not significant (P>0.2) and was not offered to logistic regression model. There was no significant difference (P>0.05) between purchased and locally raised cattle (Table
4.2.1). The distribution of seropositive animals appeared to be age related. The prevalence of antibodies to Brucella spp. tended to increase with age, peaking at 4-5 years and then decreased with increasing age of cattle as depicted in Figure 4.2.1. This was further supported by the results of the logistic regression model age was categorized into four groups (2-4 years, 4.5 -5 years, 5.5 – 7 years and >7 years) (Table
4.2.2). There were significantly (P= 0.03) more young animals (2-4 years) that were positive compared to old animals >7 years (3.2, 95% CI: 1.1, 9.1) (Table 4.2.2).
No association between the presence of newly acquired animals and brucellosis status was demonstrated and this variable was subsequently dropped from the model (OR =
1.2, 95% CI: 0.6, 2.4).
108
Table 4.2.1. Distribution of Brucella seropositive reactor smallholder dairy cattle (n =
1440) by age group, sex and origin (purchased or locally-raised) with prevalence adjusted for primary sampling unit and weights (2004-2005).
Risk factor Level Number tested Positive Percentage sero-prevalence and
95% confidence interval (CI)
Age category* 2– 4 years 531 36 6.7 (4.3, 9.0)
4.5 – 5 years 298 18 6.1 (2.0, 10.2)
5.5 – 7 years 393 23 5.5 (2.7, 8.4)
> 7years 218 4 1.3 (0.0, 2.7)
Sex Female 1296 70 5.0 (3.4, 6.6)
Male 144 11 11.3 (4.0, 18.6)
Origin of Locally raised 1269 67 5.3 (4.0, 6.5) animal* Purchased 171 14 8.2 (4.1, 12.3)
Note: *These values had Fisher’s exact or Kruskal Wallis P value ≤ 0.2 and were identified as risk factors for inclusion in the multivariable analysis
109
Table 4.2.2. Results of the logistic regression analysis for identification of individual animal risk factors in dairy cattle from smallholder farms in Zimbabwe (2004-2005)a. bResults given with beta (b), standard errors (S.E.), and odds ratio (OR) with 95% confidence intervals (CI).
Variable Level Logistic regression
SE (b) P-value 95% CI b OR
Constant -1.97 0.22 0.000 - -
Area Gokwe - - - 1.0 -
Nharira -0.64 0.32 0.04 0.53 0.28, 0.98
Wedza -1.98 0.74 0.007 0.14 0.03, 0.59
Rusitu -1.3 0.35 0.000 0.27 0.13, 0.55
Marirangwe -0.98 0.36 0.007 0.38 0.18, 0.77
Mushagashe -0.74 0.44 0.09 0.48 0.20, 1.13
Age 2-4 years - - - 1.0 - category 4.5-5 years 0.03 0.3 0.93 1.03 0.57, 1.87
5.5-7 years -0.03 0.28 0.91 0.97 0.55, 1.69
>7years -1.17 0.55 0.03 0.31 0.11, 0.9
aOverall data of the model: Log likelihood = -296.3, LR chi2(8) = 31.1, P = 0.0001, number of observations = 1440. Hosmer-Lemeshow chi2(15) = 17.7, Prob > chi2 = 0.3
110 seroprevalence Brucella
Age of the animals (years)
Fig. 4.2.1. Lowess smoother graph and scatter plots showing the relationship between
Brucella seroprevalence and age. For Gokwe and Nharira, the seroprevalence initially
increased with age and further declined with increasing age. The other study areas
showed a general decrease of Brucella seroprevalence with increasing age. For
Marirangwe, few data points in the old age category tend to distort this relationship.
111
4.2.2 Herd-level risk factors
The effect of farm size, herd size and stocking density on the distribution of Brucella seropositive farms is shown in Table 4.2.3. More sero-positive herds were from larger farms that tended to have larger cattle herds and consequently greater stocking densities than small farms with small herd sizes.
Table 4.2.3. The effect of herd size, farm size and stocking density on the distribution of Brucella seropositive dairy cattle herds from smallholder farms in Zimbabwe (2004- 2005).
Variable Seropositive farms Seronegative farms Minimum Mean Median Maximum Minimum Mean Median Maximum Farm size in 2.00 44.57 10.00 305.00 2.00 32.61 8.00 229.00 hectares (ha) Herd sizea 10 22.96 19 78 10 15.57 14.00 74 (cattle numbers per farm) Stocking 0.10 2.69 2.13 13.00 0.07 1.74 1.38 10.33 density (cattle/ha)b ap< 0.2 in Kruskal Wallis test but dropped due to collinearity with stocking density. bp< 0.2 in Kruskal Wallis test and presented to the model.
The distribution of sero-positive farms was influenced by an increase in herd size.
When herd size was divided into quartiles, the risk of a herd being infected increased
(OR = 4.9, 95% CI: 2.0; 12.1); if a farm was from the fourth quartile (20 or more animals per herd) when compared to a herd from the first quartile (11 to 12 animals per herd).
Table 4.2.4 shows the potential herd-level risk factors that were identified using univariable analysis. Predictor variable with P<0.2 were presented to the logistic regression models.
112
Table 4.2.4. Potential risk factors for the occurrence of antibodies to Brucella spp. in dairy cattle herds from smallholder farms in Zimbabwe (2004-2005). Variable Herds Brucellosis status P-value Negative (n = 151) Positive (n = 52) Management practicea . 0: Breeds not mixed 48 5 0.002 1: Mixed breeds 103 47 Type of grazing 0: Communal grazing 105 31 0.231 1: Own grazing 46 21 Method of tick control 0: Use communal dip tank 143 49 1.000 1: Use own dip tank 8 3 Source of drinking water 0: Communal water source 125 45 0.664 1: Own supply of water 26 7 Purchase of animals in the past three years 0: No 75 21 0.264 1: Yes 76 31 Cattle kept with sheep and goatsa 0: No 112 29 0.015 1: Yes 39 23 Use of animals 0: Uses own animals 140 45 0.255 1: Hire animals for use 11 7 Farmer’s knowledge of brucellosisa 0: No 100 44 0.013 1: Yes 51 8 Farm records kept 0: Yes 60 16 0.213 1: No 91 36 Use of calving pens 0: No 148 51 1.000 1: Yes 3 1 Use of cattle pen to confine cattle at night 0: No 3 1 1.000 1: Yes 148 51 Type of breeding 0: Use artificial insemination 2 0 1.000 1: Use natural service 149 52
Type of natural breeding 0: Use hired bull 47 14 1: Use own bull 104 38 0.604
Herd size quartiles 1: 10-11 cattle 49 10 2: 12-15 cattle 45 7 0.000 3: 16-19 cattle 35 13 4: ≥20 cattle 22 22 aP<0.2, presented to the multiple logistic regression models
113
4.2.2.1 Multivariable regression analysis
Of the five herd-level variables presented to the multivariable regression analysis (the practice of keeping mixed breeds, keeping cattle together with sheep and goats, farmer’s knowledge of brucellosis, stocking density, and herd size), only the practice of keeping mixed breeds and stocking density were accepted by the final model and both were good predictors of infection with Brucella spp. (Table 4.2.5).
Table 4.2.5 Final multiple logistic regression model showing the effects of herd-level risk factors on Brucella seropositivity in dairy cattle herds from smallholder farms (n=203) in Zimbabwe (2004-2005)a. bResults given with beta (b), standard errors (S.E.), and odds ratio (OR) with 95% confidence intervals (CI).
Variable Level Multiple Logistic regression
b SE (b) P-value Odds Ratio 95% CI
Constant -3.42 0.83 0.00 - - Area Gokwe - - - 1.0 - Nharira 0.21 0.6 0.73 1.23 0.38, 3.99 Wedza -1.73 0.87 0.05 0.18 0.03, 0.98 Rusitu valley -1.12 0.60 0.06 0.33 0.10, 1.06 Marirangwe 1.11 0.72 0.12 3.04 0.74, 12.4 Mushagashe 1.5 0.83 0.07 4.49 0.88, 2.88 Management Single breeds - - - 1.0 - effect Mixed breeds 2.14 0.58 0.00 8.48 2.72, 6.47
Stocking low - - - 1.0 - density high 0.35 0.11 0.001 1.42 1.15, 1.77
aOverall data of the model: Log likelihood = -92.30, LR chi2(7d.f) = 46.42, p = 0.00, number of observations = 203. Hosmer-Lemeshow chi2(8) = 5.37, Prob > chi2 = 0.72, ROC = 0.80.
114
The model had a good predictive ability (Area under the ROC curve = 0.80). Similarly,
these results were also obtained for the negative binomial regression model for counts
of positive animals per herd (Table 4.2.6). There were no significant interactions
between the main effects and post-fit testing did not reveal major influence of outliers
on the model.
Table 4.2.6 Final multiple negative binomial regression model showing the effects of individual animal level risk factors on Brucella seropositive dairy cattle herds from smallholder farms in Zimbabwe (2004-2005)a.
Variable Level Multiple negative binomial regressionb SE (b) P-value 95% CI b Risk Ratio
Constant -2.25 0.59 0.00 - - lnalpha -0.38 0.47 - - - - alpha 0.69 0.32 - - - - Area Gokwe - - - 1.0 - Nharira -0.54 0.4 0.17 0.58 0.27, 1.27 Wedza -2.17 0.78 0.01 0.11 0.02, 0.53 Rusitu valley -1.48 0.42 0.00 0.23 0.10, 0.52 Marirangwe -0.15 0.49 0.76 0.86 0.33, 2.26 Mushagashe 0.05 0.58 0.94 1.04 0.34, 3.25 Mnagement Single breeds - - - 1.0 - effect Mixed breeds 1.83 0.50 0.00 6.21 2.35, 16.45 Stocking low - - - 1.0 - density high 0.16 0.06 0.01 1.17 1.04, 1.31
aOverall data of the model: Log likelihood = -143.12, LR chi2(7 d.f.) = 48.12, P = 0.00, number of observations = 203. Likelihood ratio-test of alpha =0: Chibar2 (01) = 11.88, p= 0.00 bResults given with beta (b), standard errors (S.E.), and incidence risk ratio with 95% confidence intervals (CI).
115
4.3 Prevalence and risk factors for abortions in cows from smallholder farms
naturally infected with Brucella species
4.3.1 Descriptive statistics
A total of 1291 cows from 203 herds selected from the six study areas were tested. The antibodies to Brucella spp. were detected in 5.0 % (95% CI: 3.4, 6.6) cows. The herd- level Brucella seropositivity was 21.1% (95% CI: 14.6, 27.6). The mean prevalence of abortions at individual cow-level was 3.2% (95% CI: 1.8, 4.6). The highest (63.2%,
95% CI: 44.1%, 82.2%) and lowest (14.6%, 95% CI: 2.9%, 26.3%) numbers of herds recording abortions were recorded in Gokwe and Nharira, respectively (Table 4.3.1).
Individual animals that were sero-positive had higher odds (OR= 9.9, 95% CI: 4.5,
19.8) of having a history of abortion than sero-negative animals. Some of these positive animals had high titres of antibodies to Brucella spp. (Table 4.3.2) and a few of these showed reactions higher than 3392 IU (data not shown). A total of 22.1% (95%CI: 15.5,
28.7) of the herds had recorded abortions in the previous three years (Table 4.3.1). The highest (63.2%, 95% CI: 44.1%, 82.2%) and lowest (14.6%, 95% CI: 2.9%, 26.3%) numbers of farms recording abortions were recorded in Gokwe and Nharira respectively.
116
Table 4.3.1. The sampling weight-adjusted prevalence of abortions in individual cows and herds from smallholder dairy farms in Zimbabwe (2004-2005).
Area Total No. of cows Median Individual Herd-level
hereds sampled herd size abortion abortion
sampled prevalence prevalence
(95% CI) (95% CI)
Gokwe 30 233 14 9.9 % 63.2 %
(2.5, 17.3) (44.1, 82.2)
Marirangwe 28 243 19 5.2 % 29.8 %
(0.0, 10.9) (7.2, 52.4)
Mushagashe 15 122 17 0.6 % 25.1 %
(0.0, 1.7) (0.0, 64.9)
Nharira 40 248 16 1.4 % 14.6 %
(0.0, 3.0) (2.9, 26.3)
Rusitu 65 338 13 2.4 % 17.0 %
(0.7, 4.4) (7.3, 26.6)
Wedza 25 107 12 3.2 % 17.9 %
(0.0, 6.7) (1.5, 34.3)
Overall 203 1291 14 3.2 % 22.1 %
(1.8, 4.6) (15.5, 28.7)
117
Table 4.3.2. The SAT antibody titres of 32 Rose Bengal/c-ELISA positive cows from
Gokwe, Nharira and Rusitu Valley smallholder farming areas.
Classification based on Doubtful Low positive Strong positive Zimbabwean norms reactors reactors reactors Sera dilution 1:20 to 1:80 1:80 to 1:160 >1:160 Antibody titres (IU) 16.5 to 212 212.5 to 848 > 848 Total positive sera 8 10 14 % positive sera 25.00 31.25 43.75
Generally, the number of positive cows decreased with increasing age (Fig. 4.3.1). In contrast, a change from age groups 2-4 years to 5.5-7 years was associated with an increase in odds of abortion (OR = 4.7, 95% CI: 2.0, 11.1), but the risk subsequently decreased with increasing age. Thus, the oldest age group (>7 years) was associated with lower numbers of animals that had aborted (Fig. 4.3.1).
Fig. 4.3.1. Relative distribution of Brucella seropositive and aborted animals by age and group.
118
4.3.2 Logistic regression analyses
For individual animals, multiple logistic regression analysis revealed district of origin, age group and exposure to Brucella spp. as risk factors for abortion (Table 4.3.3).
Table 4.3.3. A multiple logistic regression model showing the effect of individual animal risk factors on abortions in dairy cattle from smallholder farms in Zimbabwe (2004-2005). Results given with beta (b), standard errors (S.E.), and odds ratio (OR) with 95% confidence intervals (CI). Variable Level Multiple logistic regressionb b SE (b) P-value Odds Ratio 95% CI Constant 2.1 0.6 0.000 - - Area Gokwe - - - 1.0 - Nharira -1.9 0.6 0.003 0.2 0.0, 0.5 Wedza -0.8 0.7 0.200 0.4 0.1, 1.6 Rusitu -0.9 0.4 0.03 0.4 0.2, 0.9 Marirangwe -1.0 0.5 0.03 0.4 0.2, 0.9 Mushagashe -2.4 1.1 0.02 0.1 0.0, 0.7 Age 2-4 years - - - 1.0 - category 4.5-5 years 1.0 0.5 0.05 2.7 1.0, 7.3 5.5-7years 1.5 0.4 0.001 4.7 2.0, 11.1 >7 years 1.1 0.6 0.65 3.1 0.9, 10.2 Brucella negative - - - 1.0 - sero-status positive 2.2 0.4 0.000 8.7 4.0, 18.9
aOverall data of the model: Log likelihood = -160.8, LR chi2(9d.f) = 62.2, P = 0.0000, number of observations = 1291. Hosmer-Lemeshow chi2 (8) = 3.4, Prob > chi2 = 0.91, ROC = 0.8.
Of the eight potential risk factors identified in univariable analysis (Table 4.3.4), only study area, farm’s Brucella seropositive status and purchase of animals in the past three years were included in the final multiple logistics regression model (Table 4.3.5).
Brucella seropositive herds were associated with higher odds of recording abortion (OR
= 3.0, 95% CI: 1.4, 6.6) when compared to seronegative herds. Farms that purchased
119 animals in the past 3 years were marginally at a higher risk of recording abortions (OR
= 2.3, 95% CI: 1.1, 4.9) compared to those that did not.
Table 4.3.4. A crude description of the potential risk factors for the occurrence of abortions in dairy cattle from smallholder farms in Zimbabwe (2004-2005).
Variable Herds Abortion status P-value Negative (n = 155) Positive (n = 48) Management practicea 0: Breeds not mixed 45 8 0.095 1: Mixed breeds 110 40 Type of grazing 0: Communal grazing 100 36 0.220 1: Own grazing 55 12 Method of tick controla 0: Use communal dip tank 149 43 0.135 1: Use own dip tank 6 5 Source of drinking water 0: Communal water source 129 41 0.825 1: Own supply of water 26 7 Purchase of animals in the past three yearsa 79 17 0.070 0: No 76 31 1: Yes Cattle kept with sheep and goatsa 0: No 115 26 0.012 1: Yes 40 22 Use of animalsa 0: Uses own animals 144 41 0.144 1: Hire animals for use 11 7 Farmer’s knowledge of brucellosisa 0: No 115 29 0.072 1: Yes 40 19 Farm records kept 0: Yes 60 16 0.609 1: No 95 32 Use of calving pensa 0: No 154 45 0.042 1: Yes 1 3 Use of cattle pen to confine cattle at night 3 1 0: No 152 47 1.000 1: Yes Type of breeding Use artificial insemination 1 1 0.418 Use natural service 154 47 Type of natural breeding 0: Use hired bull 49 12 0.472 1: Use own bull 106 36 Herd size quartilesb 1: 10-11 cattle 54 5 2: 12-15 cattle 34 18 0.001 3: 16-19 cattle 39 9 4: ≥20 cattle 28 16 aP<0.2, presented to the multiple logistic regression models bP<0.2, dropped due to collinearity
120
Table 4.3.5. A multiple logistic regression model showing the effect of herd level risk factors on abortions in cattle from smallholder cattle farms in Zimbabwe (2004-2005). Results given with beta (b), standard errors (S.E.), and odds ratio (OR) with 95% confidence intervals (CI).
Risk factor Level Multiple logistic regressionb b SE (b) P-value Odds Ratio 95% CI Constant -0.49 0.44 0.269 - - Area Gokwe - - - 1.0 - Nharira -2.2 0.6 0.000 0.1 0.0, 0.4 Wedza -1.96 0.71 0.006 0.14 0.0, 0.6 Rusitu -1.59 0.52 0.002 0.20 0.1, 0.6 Marirangwe -1.53 0.60 0.011 0.22 0.1, 0.7 Mushagashe -3.20 1.14 0.005 0.04 0.0, 0.4 Brucella Negative - - - 1.0 status Positive 1.1 0.4 0.006 3.0 1.4, 6.6 Purchase of No - - - 1.0 - cows Yes 0.8 0.4 0.04 2.3 1.1, 4.9 aOverall data of the model: Log likelihood = -93.8, LR chi2(7d.f) = 34.5, P = 0.0000, number of observations = 203. Hosmer-Lemeshow chi2 (7) = 5.4, Prob > chi2 = 0.6, ROC = 0.7.
Both individual animal level (Area under the ROC curve = 0.8) and herd-level (Area under the ROC curve = 0.7) logistic regression models had fairly good predictive abilities. There were no interactions identified between the main effects and post-fit testing did not reveale major influences of outliers on the models.
121
4.4 Comparison of the Fluorescence polarisation assay with the Rose Bengal test
and the competitive ELISA for the serological diagnosis of bovine brucellosis in
smallholder cattle farms in Zimbabwe
The receiver operating characteristic (ROC) curve (Fig. 4.4.1) for the fluorescence polarisation test using serum samples suggested an optimum cut off point of 91.0 mP resulting in a sensitivity and specificity combination of 73.9% and 98.4% respectively and similar to what has been recorded for the RBPT (Table 4.4.1). A cut off point for
FPA using stored whole blood which gave the optimum values for sensitivity and specificity was determined to be 91.3 mP. While the sensitivity (61.1 %) of the FPA using stored whole blood was moderate, the specificity (98.4 %) was similar to that recorded using serum samples.
Relative to the RBPT/c-ELISA positive and negative sera, the sensitivity and specificity of the serum FPA was 87.3% (95% CI: 78.0%, 93.8 %) and 99.2% (95% CI:
98.5%, 99.6) respectively. The relative sensitivity and specificity of the stored whole blood FPA was 70.6 (95% CI: 52.5%, 84.9 %) and 98.6% (95% CI: 96.8%, 99.6 %) respectively.
122
91.0 mP
Fig. 4.4.1. Receiver operating characteristic (ROC) curve for the fluorescence polarization test for the detection of antibodies to Brucella spp. in cattle from smallholder farming areas in Zimbabwe (2004-2005). The cut off point was determined to be 91.0 mP.
123
Table 4.4.1. Sensitivity and specificity of the Rose Bengal plate test (RBPT) (n=789) and the fluorescence polarization using serum (FPASRM) (n= 776) or stored whole blood (FPABLD) (n=351) relative to the competitive enzyme immunosorbent assay (c-
ELISA) which was used as reference test.
Test Sensitivity (%) 95 % CI Specificity (%) 95% CI PIa AUCb
RBPT 78.3 66.9, 87.3 98.2 96.9, 99.0 176.5 0.88
FPASRMc 73.9 61.9, 83.8 98.4 97.2, 99.2 172.4 0.88
FPABLDd 61.1 43.5, 76.9 98.4 95.5, 99.1 159.5 0.85
aPI = Performance index (Sensitivity + specificity) bArea under a curve = an indication of the expected accuracy of the test cCut off point of 91.0 mP as determined by ROC curve dCut off point of 91.3 mP as determined by ROC curve
124
Fig. 4.4.2 represents a histogram of the distribution of the serum FPA mP results obtained using samples tested as positive and negative on the basis of the c-ELISA. By inspection, using a cut off point of 90mP resulted in picking 15 false positive and 9 false negative sera (Fig. 4.4.2).
800 687 700
600
500
400 Positive sera
300 Negative sera
200
Number of observations of Number 100 54 0 8 9 15 16 0 0 50 90 150 170 Milipolarization units (mP)
Fig. 4.4.2. Frequency distribution of the data obtained with the fluorescence polarisation assay for the detection of antibodies to Brucella spp. using sera from cattle from smallholder farms from Zimbabwe (2004-2005).
By raising the cut off point to 91.0 mP there were 12 false positive and 12 false negative sera were detected (data not shown). These findings were confirmed using
ROC analysis.
125
The results of the test agreement are shown in Table 4.4.2. On the basis of the
McNemar X2 test, there was no evidence that the tests detected different proportions positive (P>0.05). However, there were slightly different proportions of positives between the FPA using stored whole blood and the RBPT, but, the kappa statistic indicated substantial agreement (Table 4.16). It is noteworthy that the serum FPA showed an almost perfect agreement (kappa = 0.9) with the serial RBPT/c-ELISA test results (Table 4.16). For the c-ELISA, the level of agreement with all the other tests was moderate when a cut off point of 30% PI was used but improved remarkably with a cut off point of 35% PI.
Table 4.4.2. Test agreement for the Rose Bengal plate test (RBPT), competitive enzyme immunosorbent assay (c-ELISA) and the fluorescence polarisation assay using, serum
(FPASRM) or stored whole blood (FPABLD).
Test MacNemar’s X2 test Kappa statistic* SE
(P>Chi2) (Kappa)
RBPT agreement with c-ELISA 0.71 0.8 0.04
FPASRM agreement with c-ELISA 0.26 0.8 0.04
FPABLD agreement with c-ELISA 0.19 0.6 0.05
FPASRM agreement with RBPT 0.05 0.8 0.03
FPABLD agreement with RBPT 0.03 0.7 0.05
FPASRM agreement with RBPT/c-ELISAa 1.00 0.9 0.03
FPABLD agreement with RBPT/c-ELISA 0.31 0.7 0.05
* Common interpretation of kappa: <0.2 = slight agreement, 0.2 to 0.4 = fair agreement,
0.4 to 0.6 = moderate agreement, 0.6 to 0.8 = substantial agreement, >0.8 = almost perfect agreement. aSerial testing results using the RBPT as a screening test and c-ELISA as a confirmatory test.
126
4.5 Bacteriological investigations of individual cattle by culture and isolation,
and characterization of some Brucella spp. from Zimbabwe by biochemical
profiling and AMOS PCR
4.5.1 Culture and isolation of Brucella spp. from clinical specimens
Table 4.5.1 shows the details of the milk and other clinical specimens collected from smallholder cattle for culture and isolation of Brucella spp. B. abortus (biovar 1 and 2) were isolated from two milk samples originating from two herds from Gokwe smallholder cattle farms. These isolates (B14 and B20) (Table 3.5.1) were further characterised by biochemical tests and AMOS-PCR.
Table 4.5.1. Details of the culture and isolation of Brucella spp. from milk and other specimens
Sudy area Total milk Total Brucella spp. Vaginal Total samples milk isolated swabs and positive culutured samples hygroma samples positive fluid Gokwe 54 2 B. abortus biovar 1 2 0 (herd 1) B. abortus bivar 2 (herd 2) Mrirangwe 21 0 - 5 0 Mushagashe 10 0 - 0 0 Nharira 68 0 - 0 0 Rusitu 20 0 - 3 0 Wedza 20 0 - 0 0 Total 193 2 - 10 0
127
4.5.2 Characterisation of Brucella isolates using biochemical profiles and AMOS-CR
All 14 Brucella isolates characterized in this study yielded the following results that are typical of the genus; Gram-negative coccobacilli, non-motile, positive for modified
Ziehl-Neelsen staining, oxidase and catalase production, and negative for indole production (Table 4.5.2). Their growth was aerobic after 2-5 days at 37 ○C under 10%
CO2. Their growth on Mueller-Hinton agar produced colonies that were convex, with entire edges and a smooth shiny consistency. When viewed by transmitted light, the colonies produced a golden yellow color that exhibited the characteristic “honey drop- like” appearance. These results were similar to those of the reference strains (Table
4.5.3).
Of the tests used to differentiate species and biovars isolates belonging to the same biovars of B. abortus showed consistently similar results, except for their CO2- dependence for growth (Table 4.5.4). Regardless of the biovar type, seven of the 13 B. abortus isolates were CO2-independent, while the remaining six strains were CO2- dependent. The B. abortus isolates were lysed by phages Tb, Fi, Bk2, Iz and resistant to
R/C. Only one isolate was lysed by the R/C phage. The single B. melitensis isolate was resistant to all phages but showed partial lysis to Bk2 (Table 4.5.4). B. abortus isolates were agglutinated by A-antiserum and B. melitensis by the M-antiserum, but all were not agglutinated by the R antiserum. Further, all isolates were detected by the AMOS-
PCR showing DNA amplicons of sizes 498 bp and 731 bp for B. abortus (both biovars
1 and 2) and B. melitensis (biovar 1), respectively (Fig 4.5.1 and 4.5.2).
128
Table 4.5.2. Basic biochemical and metabolic profiles of field Brucella spp. from Zimbabwe
Brucella Biochemical properties Growth isolate characteristics in TSA reference in the presence of dyes a no. Cat Oxi Ure Mot Ind MZN T20 T40 BF20
B1 + + + - - + - - + B4 + + + - - + - - + B5 + + + - - + - - + B6 + + + - - + - - + B7 + + + - - + - - + B8 + + + - - + - - + B9 + + + - - + - - + B10 + + + - - + + + + B12 + + + - - + - - + B14 + + + - - + - - + B15 + + + - - + - - - B16 + + + - - + - - + B20 + + + - - + - - - B21 + + + - - + - - +
Cat, Catalase; Oxi, Oxidase, Mot, Motility; Ind, Indole; MZN, Modified Ziehl Neelsen stain;
TSA, Tryptone Soya agar; T20, 20µl/ml thionin; T40, 40 µl/ml thionin; BF, 20µl/ml basic fuchsin; + = positive reaction; - = negative reaction. aUrease enzyme production = All isolates positive within 2 hours of culture
Table 4.5.3. Biochemical and metabolic profiles of selected reference strains of Brucella spp.
Brucella reference Ref no. Ure H2S Ind Cat Oxi Growth in TSA plus species and biovar T20 T40 BF B. abortus 1 544 + + - + + - - + B. abortus 2 86/8/59 + +/- - + + - - - B. abortus 3 Tulya + +/- - + + + - + B. abortus 4 292 + + - + + - - + B. abortus S99 + + - + + + + + B. melitensis 1 16M + - - + + + - + B. melitensis 3 Ether + - - + + +/- - + B. suis 1 1330 + a + - + + + + - B. suis 4 40 +a ND - + + + - - B. canis RM-6/66 + - - + + + - - B. ovis 63/290 - - - + - + - - B. neotomae 5K-33 + ND - + - + - -
Ure, Urea; H2S, Hydrogen Sulphide production; Ind, Indole production; Cat, Catalase; Oxi,
Oxidase; TSA, Tryptone Soya agar; T20, 20µl/ml thionin; T40, 40 µl/ml thionin; BF, 20µl/ml basic fuchsin; + = positive reaction; +/- = Doubtful positive reaction; - = negative reaction; aRapidly positive, within 5 minutes;L; ND = no data
129
Table 4.5.4. Summary of phenotypic characteristics of the field Brucella spp. from Zimbabwea
Growth characteristics Monospecific Phages at RTD AMOS-PCR Interpretation
Sera
Isolate CO2 H2S BF TH A M Tb BK2 Fi Iz R/C Size of DNA
No Dependent detected
B1, 6, - + + - + - CL CL CL CL NL 498 bp B. abortus 1,
7, 14,
16, 21
B15 - + - - + - CL CL CL CL CL 498 bp B. abortus 2,
B10 - - + - + NL PL NL NL NL 731 bp B. melitensis 1
B4, 5, + + + - + - CL CL CL CL NL 498 bp B.abortus 1
8, 9,
12
B20 + + - - + - CL CL CL CL NL 498 bp B.abortus 2
a All tests carried out by the reference laboratory (VLA), Weybridge, UK. Isolate No. = Isolate identification DNA: Test by the AMOS PCR BF = Basic fuchsin at 2µl/ml (1/50,000 w/v) TH = Thionin at 20µl/ml (1/50,000 w/v) Phages: Tb = Tbilisi, BK2 = Berkeley type 2, Fi = Firenze, Iz = Izatnagar, R/C = phage lytic for non-smooth species of Brucella CL = Confluent Lysis PL = Partial lysis NL = No lysis RTD = Routine test dilution + = positive (yes) - negative (no) bp = base pairs
130
Fig. 4.5.1. AMOS-PCR results of the field Brucella strains from Zimbabwe. Lane 1, Molecular marker; Lane 2, B. ovis (976 bp); Lane 3, B. melitenisis bv1(731 bp); Lane 4, B. abortus bv 1(498 bp); Lane 5, B. suis bv4 (not amplified) (147 bp = primer dimers); Lanes 6-9, B. abortus bv1 (498 bp); Lane 10, B. melitensis bv1; Lane 11-14, B. abortus bv 1; Lane 15, B. abortus bv2; Lanes 16-17, B. abortus bv 1, Lane 18, B. abortus bv 2; Lane 19, B. abortus bv 3 (not amplified); Lane 20, DNA control (no DNA).
Fig. 4.5.2. AMOS-PCR agarose gel photograph of reference Brucella spp. Lane 2, B. ovis (976 bp), Lane
3, B. melitensis bv 1 (731 bp); Lane 4, B. abortus bv1 (498 bp); Lane 5, B suis bv 1 (285 bp)
131
CHAPTER V DISCUSSION
5.1 Prevalence of antibodies to Brucella species in individual cattle and herds from smallholder farms in Zimbabwe
The adaptation of epidemiological studies to estimate prevalence of brucellosis and establish risk factors for exposure, in smallholder cattle farming systems is beset with numerous challenges that include poor accessibility of some farms, small herd sizes and lack of facilities for proper animal restraint, thus making it difficult to utilise proper random sampling procedures. Therefore, prior knowledge of these limitations in the study populations is required to design sampling protocols that would allow accurate estimation of the outcome variables. In the present study, in order to estimate the seroprevalence of brucellosis, the cattle study populations were stratified according to the geographical location and agro-ecological region in order to control for variation due location of study area.
The mean individual cattle prevalence of antibodies to Brucella spp. obtained for all study areas (5.6%, 95% CI: 4.4%, 6.8%) was low and similar to the data from previous studies on brucellosis in smallholder farming areas in Zimbabwe (Madzima, 1987;
Mohan et al., 1996). This seems to suggest that for most areas, Brucella spp. infection rate in individual cattle is rather low and sporadic in nature. However, some areas like
Rusitu and Gokwe recorded higher prevalences than what has been previously reported for these areas (Bryant and Norval, 1985; Madzima, 1987; Madsen, 1989; Mohan et al.,
1996). This could reflect a significant increase in Brucella spp. infections in cattle in some smallholder farming areas, where the prevalence was previously low. The resultant high individual seroprevalence of Gokwe (12.6%; 95% CI: 3.9, 21.4) partly
132 accounted for the differences among the study areas because Wedza (2.3%; 95% CI:0.0,
5.3) and the other areas recored relatively low seroprevalence.
The mean prevalence rate reported in this study was lower than what was reported previously for commercial farms in Zimbabwe (Manley, 1969; Swanepoel et al., 1975;
1976) in spite of the disease control measures such as use of B. abortus S19 and test and slaughter policies that have been implemented in commercial farms (Madsen,
1989). Similarly, our results were lower than what has been reported in other similar smallholder crop-livestock farming systems in Africa ( McDermott et al., 1987; Jiwa et al., 1996; Kadohira et al., 1997; Muma et al., 2006), with the exception of some areas
(Bedard et al., 1993; Bishop, 1984).
At herd-level the study showed that Brucella infections are present in all the study areas
(mean = 25.0%, 95% CI: 18.1%; 31.9%) but in most of the herds the number of infected individual cattle is low. This is quite encouraging from a brucellosis eradication point of view. This would emphasise the need to implement disease control measures for these smallholder cattle farms in order to reduce the prevalence of brucellosis. Gokwe,
Marirangwe, Mushagashe and Nharira had significantly higher seroprevalence (P<0.05) compared to Rusitu and Wedza (Table 4.3). While there are no comprehensive figures from previous studies on herd level prevalence of antibodies to Brucella spp. in communal areas, these results are similar to those recorded for commercial farms but higher than what has been reported for the few communal areas (Manley, 1969;
Swanepoel et al., 1976; Madsen, 1989; Mohan et al., 1996). This would suggests a rise in the prevalence of brucellosis in the smallholder farming areas and this supports the observation that the disease is equally important in both commercial and smallholder
133 cattle farming areas (Swanepoel et al., 1975). In general, the results of this study showed lower prevalence rates that reported elsewhere on the African continent (Faye et al., 2005; Muma et al., 2006), but similar to other findings (Omer et al., 2000).
The reasons for the variation in the prevalence of antibodies to Brucella spp. both at individual animal- and herd-level among the study areas could not be fully explained.
A possible explanation could be related to differences in management styles (e.g. livestock intermix, source of replacement heifers), characteristics of animal population
(e.g. herd size, population density) and agro-ecological factors that promoted or restricted contact between herds (Salman and Meyer, 1984; McDermott and Arimi,
2002). For example, the high proportion of farms from Gokwe that shared facilities for grazing and watering of cattle compared to the other study areas which kept their herds as self-contained units and this could have increased chances of contact with infected cattle in the former (Data not shown). This mixing of naïve cattle is likely to increase the risk for infection with Brucella spp., especially following abortions (McDermott and
Arimi, 2002; Muma et al., 2006). In Rusitu valley, although many farms shared grazing facilities, physical barriers like mountains that are common in the area separated areas and this is likely to prevent co-mingling of animals from many herds. Previous studies have documented low prevalence of brucellosis in this area (Bryant and Norval, 1985;
Madsen, 1989).
Another possible determinant of the distribution of the infected herds among the study areas could be related to continual introduction of new animals into the smallholder farms for the purpose of restocking herds. The spread of the Brucella spp. from one herd to another and from one area to another is almost always due to the movement of
134 an infected animal into a non-infected susceptible herd (Blood and Radostits, 1989;
Crawford et al., 1990). When the smallholder dairy projects were started between 1981 and 1991, Bos taurus cattle were purchased from commercial dairy farms and translocated to the smallholder farms where they were mixed and cross-bred with the indigenous Bos indicus breeds. The Brucella status of these purchased cattle could not be verified prior to their translocation to the smallholder farms. In addition, the introduction of the land reform programme in Zimbabwe in 2000, brought about increased movement of Bos taurus cattle from commercial farms into communal farms and also Bos indicus cattle into previously commercial farms, which were later designated as A1 (small-scale commercial) and A2 (large-scale commercial) farms.
These farming practices brought about mixing of naïve cattle between the commercial and smallholder sectors in the country.
5.2 The risk factors for infection with Brucella spp. in individual cattle and
herds from smallholder farms in Zimbabwe
At individual animal level, the lack of difference of positive reactors between males and cows indicates that the risk of infection with Brucella spp. is likely to be independent of sex of cattle. This corroborates the results of McDermott and co-workers (McDermott et al., 1987). However, the relationship between sex and the risk for brucellosis has been shown to vary with different cattle subpopulations (Kadohira et al., 1997; Kubuafor et al., 2000; Muma et al., 2006; Omer et al., 2000).
The study showed that purchased cattle were not associated with increased risk of
Brucella seropositivity. It could be that the restocking of the smallholder herds may not did not constitute an increase in the risk of Brucella spp. infection. This could be either
135 due to the fact that the animals were truly negative or if truly infected, due to increased chances of latency in the predominantly young animals that were purchased. In young heifer calves, the odds of testing positive to antibodies to Brucella spp. is low but increase with increasing age towards sexual maturity (Cunningham, 1977). Further follow up studies would be required to investigate this relationship since some infected animals which test false negative to conventional serological tests may take years before they start shedding the bacterium and hence show seroconversion (Lapraik, 1982).
Although the observed relationship between age and Brucella seroprevalence generally agrees with other reports ( Omer et al., 2000; Muma et al., 2006) the findings in this study are at variance with what has been recorded (McDermott et al., 1987; Silva et al.,
2000).The preponderance of seropositive reactors in younger cattle is consistent with the biology of Brucella spp. The onset of sexual maturity is associated with a significant increase in the risk of infection with Brucella spp. (Walker, 1999) and such animals are likely to seroconvert. However, the age at which sexual maturity is attained varies with breeds of cattle and this is likely to influence the observed relationship between age and positive reactors in different sub-populations. The decrease in the risk of testing positive in older cattle could be related to latency, a phenomenon that is not uncommon in chronic brucellosis (Cunningham, 1977) and this leads to false negative serological results.
To minimise false positive reactions due to maternal antibodies in younger animals, only animals at least 2 years old were included in the study. Calves from seropositive dams are usually seropositive for up to 4-6 months due to colostral antibodies and later test negative (Blood and Radostits, 1989). In addition, only animals that were not
136 vaccinated against B. abortus S19 were tested. Moreover, c-ELISA used to confirm seropositive reactors can differentiate antibodies due to B. abortus S19 from those of natural infection (Nielsen et al., 2002).
Seropositivity due to cross-reacting antibodies caused by Yersinia enterocolitica
(Godfroid, 2002; Nielsen et al., 2004) are considered inconsequential since this pathogen is rare in the tropics. Therefore, due to high combined serial specificity of the tests used in this study (Nielsen et al., 1995; Samartino et al., 1999), it is unlikely that the classification of tested animals in the present study was biased towards false positive results.
At herd-level of aggregation, the multivariable regression analysis has revealed that the management practice of keeping mixed breeds and stocking density were independently associated with Brucella seropositivity of herds. Herds that kept mixed cattle breeds were approximately 7.5 times more likely to be seropositive compared to control herds(OR= 7.54; 95% CI: 2.55, 22.28). It is likely that this mixing of naïve cattle(Madsen, 1989) was an important risk factor for exposure to Brucella spp.
Brucellosis has been reported to be enzootic in some commercial dairy farms in
Zimbabwe, while others have eradicated it due to the implementation of the national eradication scheme that was gazetted for the commercial dairy farming sector (
Swanepoel et al., 1976; Mohan et al., 1996). At the same time, previous work showed low prevalence rates of the disease in communal areas (Bryant and Norval, 1985;
Madsen, 1989; Manley, 1969). Therefore, the movement of cattle from commercial farms could have resulted in the spread of brucellosis to smallholder farms. Elsewhere, it has been reported that the more often cattle are introduced into the herd, the greater
137 the risk of introducing Brucella spp. (Crawford et al., 1990). Frequent purchasing of cattle would increase chances of contact with infected herds (Omer et al., 2000;
Reviriego et al., 2000). In contrast farms that tend to keep closed herds are at a lower risk of infection with Brucella spp. (Bedard et al., 1993). One of the major reasons for introducing new animals into herds is for the purpose of upgrading the genetics of their stock, a feature which accounts for higher seroprevalence of brucellosis in dairy than in beef due to more interchange of cattle in the former (Christie, 1969). In the present study, there were no available records of brucellosis testing of cattle prior to their movement from commercial to smallholder areas. Therefore, in the absence of these data, the brucellosis status of these animals could not be ascertained.
Despite the variability in defining what a large herd is, it is generally accepted that an increase in herd size is usually accompanied by an increase in stocking density and increased risk of infection-espcially following abortion (Nicoletti, 1980; Salma and
Meyer, 1984). Thus, cateris peribus (having all variables fixed), it is likely that a larger herd size with increased stocking density tends to have conditions that favour maintenance of Brucella spp. infections than a small herd with low stocking density, because of the increased intensity of contact between infected and non-infected cattle.
This also partly explains why brucellosis prevalence tends to be higher in intensively managed large dairy herds ( Nicoletti, 1980; Bishop et al., 1994), while low stocking density is associated with reduced seropositivity (Bedard et al., 1993).
Although other management factors may be at play in influencing the observed relationship between stocking density and herd Brucella infection (Crawford et al.,
1990), the positive association between two factors, reported in this study is in
138 agreement with the general epidemiology of Brucella spp. as observed in other parts of the world (Nicoletti, 1980; Kadohira et al., 1997; McDermott and Arimi, 2002; Faye et al., 2005; Muma et al., 2007). This study is the first in Zimbabwe to investigate agro- ecological and management factors that are associated with increased risk of infection with Brucella spp. for smallholder cattle herds. Such epidemiological studies that seek to establish the the factors associated with exposure to Brucella spp. provide useful information that could used as the basis for designing brucellosis control programme in this farming sector. However, the fact that the logistic regression models were based mainly on statistical criteria not on relative biological importance, they could have missed other variables that are important risk factors for infection with Brucella spp. In general, the multivariable logistic regression model utilises herd-level variables to predict the occurrence of an outcome of interest (herd Brucella infection status), while the negative binomial regression model assesses risk at individual animal-level by taking into account of the herd effect ( Omer et al., 2000; Dohoo et al., 2003).
The use of communal grazing pastures, common water sources and communal dip tanks for control of ticks were not associated with increased odds of farms testing positive to
Brucella spp. antibodies. While this would appear to be contrary to the biology of the disease, but in reality, in typical smallholder dairy farms in Zimbabwe, only a few adjacent herds, with relatively low numbers of cattle share these communal facilities.
Moreover, most of these herds tend to be sedentary, and rarely do these facilities get shared by large numbers of animals at any given time.Therefore, under these conditions of extensive management systems with sedentary herds, there is a tendency of infection with Brucella spp. to be clustered by herds (Madsen, 1989; Kadohira et al., 1997).
139
Higher seroprevalence of brucellosis has been established in pastoral than in sedentary herds (Kadohira et al., 1997). In the former, herds usually aggregate to form super herds resulting in concentrated stock at any given time and space, as these animals share the same facilities for grazing and water (Oloya et al., 2006). This livestock intermix is an important risk factor for spread of infection with Brucella spp. (McDermott and Arimi,
2002). Thus brucellosis risk increases with a change from purely extensive, nomadic to more intensive forms of cattle management (Thim and Wundt, 1976). Indeed, this phenomenon is similarly observed for other infectious diseases such as bovine tuberculosis ( Neil et al., 1989; Munroe et al., 1999).
The most important method of spread of B. abortus is from infected cows that contaminate pastures and drinking water such that susceptible animals pick up infection when grazing or drinking water (Blood and Radostits, 1989). The survival of Brucella spp. in the environment is very much dependent on temperature, humidity and the presence of nutrients. In temperate regions, the organisms may survive for up to 100 days in soil in winter and 30 days in summer (Blood and Radostits, 1989). Since these organisms are very susceptible to sunlight and heat (Blood and Radostits, 1989), their survival under tropical conditions is likely to be significantly reduced. In the different areas in this study, the sururvival of Brucella spp is likely to be only a few days on hot and dry conditions, while the survival will be longer in warm and wet conditions in summer which will potentiate spread of brucellosis.
One source of bias in the present study is that cattle from some small herds (<10 cattle) were not included in the study and this could have missed important risk factors related to inherent management styles. In addition, the number of animals sampled from each
140 farm could have had an influence on classification of farms as being positive or negative. Classification of farms as positive on the basis of the presence of at least a single positive reactor animal resulted in higher herd sensitivity, but with some slight compromise on herd specificity. However, it was not expected that this would bias the results because under the current test regime, herd specificity was estimated at 100% when the cut-off point for classifying a herd as positive when at least single positive animal was detected, according to Herdacc™(Jordan, 1995).
Similarly, the problem of misclassification of positive cattle due to antibodies produced by B. abortus S19 was negligent since the farms included in the study were not using the vaccine. In addition, the use of c-ELISA test is able to differentiate antibodies produced by B. abortus from those of field strains of Brucella spp. (Nielsen et al.,
1989). Serological cross reactions due to antibodies caused by Yersinia enterocolitica
(Godfroid, 2002; Nielsen et al., 2004) were unlikely to be a problem since this pathogen is rare in the tropics (Murray et al., 1999). Further, due to the high combined serial specificity of the RBT and c-ELISA used in this study (Nielsen et al., 1995; Samartino et al., 1999), it is unlikely that the classification of tested cattle was biased towards false positive results.
5.3 Prevalence and risk factors for abortions in cows from smallholder farms
naturally infected with Brucella species
This study was conducted to investigate the contribution of Brucella spp. infections on abortions and to identify the risk factors for abortion in individual cattle and herds from smallholder dairy farms in Zimbabwe. The study established that Brucella seropositive
141 cows were approximately 10 times (OR= 9.9, 95% CI: 4.5, 19.8) more likely to have aborted their foetuses compared to seronegative cows. The preponderance of abortions in Brucella sero-positive cows and herds suggests that exposure to Brucella spp. is an important risk factor for abortion. Further, high antibody titres exhibited by some of the seropositive animals suggested the presence of active infection. This is consisitent with the biology of B. abortus in pregnant cows and concurs with what has been reported in other parts of the world (Swanepoel et al., 1976; McDermott et al., 1987; Kubuafor et al., 2000; Silva et al., 2000; McDermott and Arimi, 2002; Schelling et al., 2003; Muma et al., 2007). Cows are particularly susceptible to infection during early pregnancy and this has been explained by the presence of erythritol which is a growth stimulant for B. abortus (Quinn et al., 1999). Such infections could result in late term abortion
(Cunningham, 1977). However, in infected herds, a certain proportion of infected cows may not abort (Brinley Morgan, 1977) and this could distort association between history of abortion and seropositivity.
The observed relationship between Brucella seropositivity and age was at variance with what has been recorded elsewhere (McDermott et al., 1987; Silva et al., 2000; Muma et al., 2006), but concurred with those of Omer and co-workers (Omer et al., 2000).
However, it is likely that other factors may be at play here to account for the variance observed in our study. It is generally accepted that the onset of sexual maturity is likely to be associated with increased risk of seropositivity (Gul and Khan, 2007). Cows that are pregnant are particularly susceptible since the gravid uterus of ungulates produces erythritol and other allantoic factors, which stimulate the growth of Brucella spp (Smith et al., 1962; Walker, 1999). But in herds where B. abortus is endemic, the median titres of antibodies ultimately decline (Muma et al., 2006) because a proportion of older cows
142 that was previously exposed to infection may develop latent infection and these would be negative to conventional serological tests (Cunningham, 1977; Lapraik, 1982). This degree of “tolerance” to B. abortus infection would result in a decline in the numbers of seropositive reactors with increasing age of the cows.
The multiple logistic regression analysis revealed that the highest odds (OR=4.7, 95%
CI: 2.0, 11.1) of abortion was observed in the age group 5.5-7 years when compared to age group 2-4 years. The risk of abortion declined in older cows. Since the actual age at which the cows aborted was not established, it is likely that most of the cows in the 5.5 to 7 year-old age group experienced these abortions during the first pregnancy. It could be that such young cows were particularly susceptible due to lack of solid immunity
(Cunningham, 1977). As herd immunity builds, abortions become infrequent since most animals abort only once, even though some may still be infected and shedding B. abortus during apparently normal parturitions (Cunningham, 1977). Our results corroborate those of Muma and co-workers (Muma et al., 2007).
Of the herd level factors, area of origin, managememt practice of purchasing cattle from other herds (OR = 2.3, 95% CI: 1.4, 4.9) and Brucella seropositive herds (OR = 3.0,
95%CI: 1.4, 6.6) were independently associated with increased odds of abortions in smallholder cattle farms. Although the causes of abortion in cattle are multifactorial, these results showed that Brucella spp. infections contribute significantly to these abortions since seropositive herds are positively associated with the outcome. It is noteworthy that the distribution of Brucella seropositve herds is dependent on study area. A possible explanation for this area-depenednt distribution of Brucella seropositive herds could be related to management styles and agro-ecological factors
143 that promoted or restricted contact between herds (McDermott and Arimi, 2002). The fact that areas with larger herd sizes had increased risk of being Brucella seropositive is consistent with what is generally known about brucellosis (Nicoletti, 1980; Salman and
Meyer, 1984; Crawford et al., 1990). However, other factors could influence the distribution of seropostive herds. For example, Gokwe which had a lower median herd size than Nharira and Marirangwe had a higher herd-level Brucella seroprevalence.
There was a higher proportion of herds from Gokwe that shared facilities for grazing and watering of cattle compared to the Nharira and Marirangwe which kept their herds as self-contained units (Data not shown). The explanation for this could be that these cattle management strategies could have a stronger influence on seroprevalence than herd size. The mixing of cattle especially where large herds are likely to aggregate at watering points, similar to what has been observed with pastoralists’ cattle, is associated with higher Brucella seroprevalence (Kadohira et al., 1997).
This study has revealed that the risk of abortion increased in herds that had purchased cattle from other farms. These smallholder herds continually purchase cattle from other herds, especially commercial dairy farms, for the purpose of improving the genetics or to restock their herds. While the cause of abortions in these herds could not be explicitly identified, they could be linked to either infectious causes or the stress of translocating the animals. However, it is likely that Brucella infection contributed a significant proportion of these abortions. It is possible that the purchase of animals could have introduced Brucella spp. infections in these herds. It is generally accepted that the spread of the Brucella spp. from one herd to another is almost always due to the movement of an infected animal into a non-infected susceptible herd (Crawford et al.,
1990; Kabagambe et al., 2001). The practice of keeping mixed breeds of cattle has been
144 shown to be an important risk factor for infection with Brucella spp. (Omer et al.,
2000). In these herds, the risk of Brucella abortions is high due to increased chances of
“importation” of brucellosis from infected herds. Therefore, in these Brucella infected smallholder cattle farms, the disease is likely to be a major contributor to loss of the calf-crop causing a significant economic loss to their livelihood. The livelihood of most people in smallholder farms is dependent on cattle production. Cattle are a vital source of milk, meat, income, drought power and manure, in addition to their use in numerous social and cultural roles. In addition, the economy of Zimbabwe is heavily dependent on agriculture, which contributes to about 11-14% of the national gross domestic product
(GDP), with the Livestock industry contributing about 4% of the national GDP.
Therefore the important inefectious diseases such as brucellosis could cause significant economic loss if uncontrolled. Hence the need to control bovine brucellosis in the smallholder cattle sector can not be over-emphasized.
5.4 Comparison of the Fluorescence polarization assay with the Rose Bengal
test and the competitive ELISA for the serological diagnosis of bovine
brucellosis in smallholder cattle farms in Zimbabwe
The purpose of this study was to evaluate the diagnostic performance of fluorescence polarization assay (FPA) relative to the conventional tests; the c-ELISA and RBT, in order to introduce it as an alternate, simple and rapid diagnostic test for bovine brucellosis in smallholder cattle farms naturally inefected with Brucella spp. In addition to using sera, the FPA was also conducted on stored whole blood samples to detect antibodies to Brucella spp. In the absence of a “gold standard tests” to classify animals as truly infected and non-infected, the diagnostic performance of the FPA has
145 been evaluated relative to the c-ELISA and and the serial c-ELISA/RBT testing, in herds naturally infected with Brucella spp. The values of sensitivity, specificity and the performance index (sum of percent sensitivity and specificity) of FPA relative to the c-
ELISA, reported in this communication are similar to those of the RPT, which is routinely used as a screening test for the diagnosis of bovine brucellosis in Zimbabwe.
The cut off value of 91.0 mP that was determined by ROC curves gave the maximum value of the performance index that is considered to be the optimum value for a particular assay. The cut off points may be altered in order to increase or decrease the values of sensitivity and specificity depending on the relative consequences of getting more false positive or false negative results, depending on the test situation (McGiven et al., 2003). However, these findings suggested that the FPA could be used alternately to the RBT without seriously compromising the sensitivity but with an added advantage of improving on specificity. In previous studies, the c-ELISA has been reported to have superior or similar values of sensitivity and specificity in comparison with the RBT and FPA (McGiven et al., 2003; Nielsen et al., 1996a). Although the sensitivity of the test using whole blood was found to be low, the specificity was similar to that obtained using serum and marginally higher than that of the RBT.
Despite the relatively lower sensitivity values reported in this communication, the results for specificity are similar to what has been reported in literature for FPA
(Nielsen et al., 1996a ; Dajer et al., 1999; McGiven et al., 2003; Minas et al., 2005).
Hence, these results have shown that the FPA is a promising test which could be adopted for the diagnosis of brucellosis based on its superior specificity. A test with an increased specificity such as has been observed for FPA in this study, would be suitable for use in areas where the prevalence of brucellosis is low or towards the end of an eradication programme where non-specific serological reactions are likely to be
146 problematic (Godfroid et al., 2002). Furthermore, the specificity of CFT, which is currently used as a confirmatory test in Zimbabwe, has been questioned due to its failure to detect persistently infected animals (Godfroid et al., 2002) or differentiate B. abortus S19 vaccinal antibodies from those produced against natural infection (Nielsen,
2002). The FPA was reported to be able to distinguish vaccinal antibodies from those of natural infection (Nielsen et al., 1996a). In addition serological cross reactions in cattle known to be infected with E. coli O:157: H7 and Y. enterocolitica O:9 were reported not to be a major problem with FPA (Nielsen et al., 2004). Sereological cross reactions were unlikely to influence the results of this study, because these pathogens are relatively rare in cattle, and Y. enterocoloitica has been reported to be rare or absbent in the tropical region (Murray et al., 1999).
By assuming that the two tests, c-ELISA and RBT are conditionally dependent (Muma et al., 2007), it has been possible to evaluate the FPA against the serial test results of these two tests. The results of the kappa statistic indicated an almost perfect agreement between the two tests. In spite of a relatively lower sensitivity, these results are similar to those of Nieslen and Gall (Nielsen and Gall, 2001). Thus the FPA could potentially circumvent the need to use the serial RBT/c-ELISA tests in the detection of antibodies to Brucella spp. This will not only improve the turn around time, but also reduce the cost of the serological testing. The results of this study ssupport the findings of Muma and co-workers where a more sophisticated method which used a Bayesian formulation of the latent class model to evaluate these serological tests in the absence of a gold standard test was available (Muma et al., 2007).
147
However, further validation studies using sera and whole blood from animals with known Brucella spp. infection status are required to determine proper estimates of test sensitivity and specificity as has been reported elsewhere (Nielsen and Gall, 2001).
5.5 Bacteriological investigations of individual cattle by culture and isolation,
and characterization of some Brucella spp. from Zimbabwe by biochemical
profiling and AMOS PCR
The bacteriological investigations of individual cattle using culture of milk and other clinical specimens showed a very low sensitivity. Only two milk samples from cows originating from separate herds in Gokwe tested positive for B. abortus, confirming the existence of actice infection in these herds which also tested positive by serological tests. The low isolation rate could be reflective of the low level of Brucella spp. infection in individual cattle from smallholder farms in the study areas. Similarly,
Brucella spp. are fastidious micro-organisms snd their isolation is difficult as they can easily be overgrown by contaminants (Farrell, 1972). More over, the shedding of B. abortus in the infected cow tends to occur intermittently, usually in low numbers and chances of isolating them tend to increase at towards the end of lactation when volumes of milk are reduced (Cunningham, 1977; OIE, 2004). Isolation of these microorganisms may better be done using specimens from the aborted foetus and placenta (Alton et al.,
1988) (Blood and Radostits, 1989), but the limitations of the current project precluded the use of such clinical specimens. Nevertheless, the isolation of two strains of B. abortus from two different seropositive herds confirmed their infection status. These results could suggest that, brucellosis was endemic in some of the smallholder farming areas, a situation that is similar to that in commercial herds. It is likely that in the
148 absence of control measures for brucellosis in communal areas, the continual movement of cattle from commercial to communal areas is likely to be concomitantly associated with an increase in the prevalence of brucellosis in the latter, thereby establishing a state of endemicity in some areas.
This study provided the the first detailed biochemical profiling and AMOS PCR characterization of some Brucella isolates from Zimbabwe. Generally speaking, the phage sensitivity patterns of all the Brucella isolates investigated in this study were consistent with what has been reported for their respective species and biovars (Garritty et al., 2005). However, a single isolate of B. abortus (B15) was lysed by R/C phage, while B. melitensis (B10) was partially lysed by Bk2, which indicated a possibility of dissociation of the colonies to the non-smooth phases (Garritty et al., 2005). In contrast, all the B. abortus isolates and the single B. melitensis were agglutinated by the A (A- specific epitope dominant) M (M-specific epitope dominant) monospecific antiserum respectively which indicated absence of dissociation. All smooth strains of Brucella may possess either the A, M or both A and M antigenic epitopes on the O chains of the lipopolysaccharides (Garritty et al., 2005). None of our isolates were agglutinated by the R antiserum which could indicate dissociation. The use of phage typing as a means of differentiating Brucella spp. has become less discriminatory as a typing tool because of the discovery of new strains with atypical sensitivity patterns (Jahans et al., 1997).
The growth characteristics and the biochemical profiles reported in this study are consistent with what is reported for Brucella spp. and biovars (Alton et al.,1988; Quinn et al., 1999; Garritty et al., 2005). However, the requirement for CO2 for growth was at variance with reports from literature (Alton et al., 1988). Although most strains of B.
149
abortus biovars 1-4 require CO2 for primary isolation (Garritty et al., 2005), this attribute is quickly lost on repeated subcultures (Alton et al., 1988).
The use of the AMOS-PCR assay confirmed the identity of the Brucella spp. that was attained by phenotypic tests. The IS711 analysis using AMOS-PCR can identify only three B. abortus biovars; 1, 2 and 4, all three biovars of B. melitensis, biovar 1 of B. suis and B. ovis, but individual biovars within a species are not differentiated (Bricker and
Halling, 1994). However, using the current tests, it was not possible to differentiate strains of the same biovar. Therefore, further DNA fingerprinting methods such as the variable number of tandem repeat analysis (VNTR) (Bricker et al., 2003) could be used to investigate further molecular epidemiology of these Brucella isolates.
The preponderance of B. abortus biovar 1 could suggest that it is the predominant cause of bovine brucellosis in both commercial and communal cattle in Zimbabwe. This result is not surprising because infections with Brucella spp. tend to colonize a given geographical area. Our studies corroborate those of Mohan and co-workers who reported similar findings (Mohan et al., 1996). In addition, this study documented the occurrence of B. abortus biovar 2 in cattle from both farming sectors, albeit at a lower frequency than B. abortus biovar 1. Similar frequencies where 90% of the B. abortus isolates were biovar1 and 10% biovar 2 have been documented for South Africa (Herr et al., 1991; Bishop et al., 1994). South Africa, to a large extent, shares similar geographic, climatic and livestock husbandry systems with Zimbabwe. While it is difficult to explain the reasons for the distribution of these B. abortus biovars in the cattle farming sectors, this could be largely influenced by the purchase of animals from commercial farms practiced by communal farmers for the purpose of improving the
150 genetics of their herds (Christie, 1969; Madsen, 1989). Despite that relatively fewer isolates were characterized these results suggest that B. abortus biovar 3, and indeed other biovars are rare or non-existent in Zimbabwe. However, study of a large number of isolates could reveal a more accurate distribution of B. abortus biovars in Zimbabwe.
Elsewhere, B. abortus biovar 3 has been infrequently reported in South Africa, East and
North Africa, while there seems to be no reports of isolation of the other biovars (Herr et al., 1991; Quinn et al., 1999). World wide, in countries where bovine brucellosis is endemic, B. abortus biovar 1 is predominant and B. abortus biovar 2 occurs less frequently while the other biovars are rare (Quinn et al., 1999; Bishop et al., 1994).
151
CHAPTER VI
CONCLUSIONS AND RECOMMENDATIONS
6.1 Conclusions
6.1.1 Prevalence of antibodies to Brucella spp. in individual cattle and herds from
smallholder farms in Zimbabwe
This study showed that, despite the lack of control measures, the prevalence of antibodies to Brucella spp. in individual cattle is low in five of the six study areas of
Marirangwe, Mushagashe, Nharira, Rusitu Valley and Wedza except in Gokwe where the disease is likely to be endemic. The observed pattern of distribution of Brucella seropositive reactor cattle is likely to be influenced by cattle management differences.
At herd-level, the prevalence was high in four of the six study areas of Gokwe,
Marirangwe, Mushagashe, and Nharira, and low in Wedza and Rusitu smallholder farming areas. While the herd-level seroprevalnce is high, most of the herds had low numbers of positive reactor cattle. Despite the limited sample sizes studied, it is likely that bovine brucellosis is widespread smallholder cattle farms in Zimbabwe. It is envisaged that the herd-level Brucella seroprevalence (mean 25.0%) reported in this study could be reduced, if rigorous measures to control animal movement and mixing are implemented.
152
6.1.2 The risk factors for infection with Brucella spp. in individual cattle and
herds from smallholder farms in Zimbabwe
The study revealed that in individual cattle, the prevalence of antibodies to Brucella spp. was independent of sex. The risk of infection with Brucella spp. increased with age from 2-4 years age group to the 4.5 to 5 years age group and subsequently decreased with increasing age of cattle.
At herd level stocking density and the management practice of keeping mixed breeds of cattle were independently associated with Brucella seroprevalence. The mixing of herds, created through purchase of cattle from commercial farms was an important risk risk factor for brucellosis in smallholder cattle.
The practice of communal grazing, use of communal dip tanks and common water sources were not associated with increased risk for Brucella spp. infection probably due to the extensive nature of cattle management. While extensive cattle management systems in smallholder farms coupled with low stocking density are likely to reduce the risk of transmission of Brucella spp., the purchase of cattle increases the chance of contact with other infected herds. Thus, the introduction of brucellosis control measures could be beneficial to these smallholder dairies.
153
6.1.3 Prevalence and risk factors for abortions in cows from smallholder farms
naturally infected with Brucella species
Infections with Brucella spp. significantly contribute to abortions in smallholder cattle in Zimbabwe. Exposure to Brucella spp. and age of cows were identified as important risk factors for abortion. The risk of abortion increased from the 2-4 years age group and peaked at 5.5-7 years age group and further declined with increasing age.
The distribution of aborting herds was influenced by the study area and could be related to cattle management differences. Herds that were Brucella seropositive and those that purchased cattle from other farms were at increased risk of recording abortions.
Therefore serological testing of animals prior to movement could reduce the risk of introducing infected animals. The use of B. abortus S19 vaccination in areas with high prevalence of Brucella spp. could improve herd immunity and reduce the incidence of abortions.
6.1.4 Comparison of the Fluorescence polarisation assay with the Rose Bengal
test and the competitive ELISA for the serological diagnosis of bovine
brucellosis in smallholder cattle farms in Zimbabwe
This study has shown that there is good level of agreement among the serological tests; the RBT, c-ELISA and the FPA. Relative to the c-ELISA, the FPA was shown to be highly specific and moderately sensitive. The performance index for serum FPA similar to that of RBT but the former showed a superior specificity than the latter. The specificity of the whole blood FPA was similar to that of the serum FPA. Thus, based on the fact that the FPA is relatively cheap, and because of its ease of use and rapidity, the procedure, the test could be adopted as a confirmatory test for brucellosis due to its
154 superior specificity. The test could be adapted for use in the field. But, the performance of both the FPA and c-ELISA under Zimbabwean conditions would be improved by adapting the test cut-off points to local conditions rather than using universally recommended values.
6.1.5 Bacteriological investigations of individual cattle by culture and isolation,
and characterization of some Brucella spp. from Zimbabwe by biochemical
profiling and AMOS PCR
The bacteriological investigation of herds using culture of milk samples showed a very low sensitivity and this could be reflective of the low level of Brucella spp. infection in individual cattle from smallholder farms in the study areas.
The biochemical profiles of Brucella abortus isolates (collected from both commercial and smallholder cattle farms) typed in this study were consistent with those of reference strains, except one B. abortus isolate that showed some atypical results for sensitivity to phages. The study has shown that B .abortus biovars 1 and 2 are present in both commercial and communal cattle farming sectors. Despite the limited number of
Brucella isolates investigated, it is likely that B. abortus biovar 1 is the predominant cause of bovine brucellosis in both commercial and communal cattle farming sectors in
Zimbabwe, while biovar 2 is infrequent. There is likely sharing of B. abortus strains between the commercial and smallholder cattle farms. However, further bacteriological investigations are required to confirm their infection status of individual cattle and herds that were identified as positive on the basis of serology. The study of more isolates using tests such as the variable number of tandem repeat analysis (VNTR) which can
155 identify within-biovar differences would reveal more information on the molecular biology of B. abortus in smallholder cattle farms.
6.2 Recommendations
6.2.1 Diagnosis of bovine brucellosis
It is recommended that the diagnosis of bovine brucellosis be accomplished serologically using a combination of RBT and c-ELISA because they have been shown to be reliable, accurate and are less cumbersome than the CFT that is currently in use as a confirmatory test in Zimbabwe. The FPA could be adopted as a field test because it is simple, reliable, inexpensive and can be applied on fresh whole blood immediately collected from the animal.
The culture of Brucella spp. for the purpose of confirming diagnosis should be avoided as much as possible because as it is not only difficult and time consuming, but also potentially hazardous because the microorganisms are important zoonotic pathogens.
Exceptionally, culture of Brucella spp. may be carried out provided there are experienced personnel and that there are special laboratories designed to handle such dangerous pathogens.
6.2.2 Control of bovine brucellosis in smallholder dairy farming areas
The importance of brucellosis to animal production and public health justifies control measures such as have been adopted in some countries where the disease occurs
(Chukwu, 1987). In Zimbabwe, previous brucellosis control measures have mainly focused on the commercial cattle farming sector, these should also be extended to the smallholder cattle farming areas where the livelihood of people is heavily dependent on
156 cattle production. While eradication may not be possible in the near future, it is not an impossible long-term objective. It is therefore recommended that cattle movement control be strictly enforced. Serological testing of animals for antibodies to Brucella spp. should be carried out prior to translocation of animals to other farms. If possible purchase of replacement heifers for smallholder dairy schemes should be done from brucellosis-free commercial herds. The introduction of a brucellosis accreditation scheme in the smallholder cattle farms especially where there are newly introduced dairy co-operatives would bring long term benefits. Therefore the test and slaughter policy should be stringently implemented and monitored by State authorities.
In farms where the prevalence of bovine brucellosis is relatively high, it is recommended that vaccination of calves with B. abortus S19 vaccine be done. This would significantly reduce the incidence of the disease due to an increase in the level of solid herd immunity and would ultimately improve cattle productivity in smallholder farming areas. In addition, this would go a long way towards improving the safety of milk from smallholder farmers whose livelihood depends on livestock production.
157
REFERENCES
Akova, M., Uzun, O., Akalin, H.E., Hayran, M., Unal, S., Gur, D., 1993. Quinolones in treatment of human brucellosis: comparative trial of ofloxacin-rifampin versus doxycycline-rifampin. Antimicrobial Agents and Chemotherapy 37, 1831- 1834. Alton, G.G., Jones, L.M., Pietz, D., 1975. Laboratory Techniques in Brucellosis, Geneva, 63-34 pp. Alton, G.G., 1977. Development and evaluation of serological tests. In: Brucellosis: An International Symposium, London, pp. 161-171. Alton, G.G., Corner, L.A., 1981. Vaccination of heifers with a reduced dose of Brucella abortus Strain 19 vaccine before 1st mating. Australian Veterinary Journal 57, 548-550. Alton, G.G., 1987. Control of Brucella melitensis infection in sheep and goats - a review. Tropical Animal Health and Production 19, 65-74. Alton, G., Jones, L.M., Angus, R.D., Verger, J.M., 1988, Techniques for the brucellosis laboratory, Institut National de la Recherche Agronomique, Paris, France, pp. 81-134. Anon., 1957. An orbituary; Mr L.E. Bevan. The Veterinary Record 69, 421-422. Anon., 1986. Fistulous withers and poll evil. The Merck Veterinary Manual, 6th Ed., Merck and Company, Inc., Rahway, N.J., USA, pp 455. Anon. 1995. Animal Health (Brucellosis) Regulations, 1995. Statutory Instrument 104 of 1995, of section 5 of the Animal Health Act. Department of Veterinary Services, P. O. Box CY66, Causeway, Harare, Zimbabwe. Anon. 2001. Annual Census Report for 2001. Department of Veterinary Services, Ministry of Lands, Agriculture and Ruaral Resettlement, P. O. Box CY66, Causeway, Harare, Zimbabwe. Anon. 2005. Modified Brucella Selective Supplement, Oxoid Limited, Wade Road, Basingstoke, Hampshire, England. Anon 2006. Bovine brucellosis (http://www.fao.org/ag/againfo/subjects/en/ health/diseases-cards/brucellosi-bo.html). Ariza, J., Gudiol, F., Pallarés, R., Rufí, G., Fernández-Viladrich, P., 1985. Comparative trial of rifampin-doxycycline versus tetracycline-streptomycin in the therapy of human brucellosis. Antimicrobial Agents and Chemotherapy 28, 548-551.
158
Baggley, C.V., Burrell, W.C., Esplin, G.M., Walters, J.L., 1984. Effect of epididymitis on semen quality of rams. Journal of the American Veterinary Medical Association 185, 876-877. Baggley, C.V., Paskett, M.E., Mathews, N.J., Stenquist, N.J., 1985. Prevalence and causes of ram epididymitis in Utah. Journal of the American Veterinary Medical Association 186, 798-801. Baldi, P.C., Wanke, M.M., M.E., L., Fossati, C.A., 1994. Brucella abortus cytoplasmic proteins used as antigens in an ELISA for the diagnosis of canine brucellosis. Veterinary Microbiology 41, 127-134. Banai, M., Adams, L.G., Frey, M., Pugh, R., Ficht, T.A., 2002. The myth of Brucella L- forms and possible involvement of Brucella penicillin binding proteins (PBPs) in pathogenicity. Veterinary Microbiology 90, 263-279. Barrow, G.I., Felthman, R.K.A., 1993. Cowan and Steele’s Manual for the identification of medical bacteria, 3 Edition. Cambridge Press, 94-150 pp. Beckett, F.W., MacDiarmid, S.C., 1985. The effect of reduced dose Brucella abortus strain 19 vaccination in accredited dairy herds. British Veterinary Journal 141, 507-514. Bedard, B.G., Martin, S.W., Chinombo, D., 1993. A prevalence study of bovine tuberculosis and brucellosis in Malawi. Preventive Veterinary Medicine 16, 193- 205. Bekele, T., Kasali, O.B., Mukasa-Mugerwa, E., Scholtens, R.G., Yigzaw, T., 1989. The Prevalence of Brucellosis in Central Ethiopia. Bulletin of Animal Health and Production in Africa 37, 97-98. Berman, D.T., Irwin, M.R., 1952. Studies on repeated vaccination of cattle with B. abortus S19. The response of vaccinated and revaccinated cattle to conjuncticval exposure with a virulent strain of B. abortus during the third gestation period. American Journal of Veterinary Research 13, 351. Bevan, L.E.W., 1914. Annual Report of the Veterinary Bacteriologist, Southern Rhodesia. Bevan, L.E.W., 1931. Notes on a case of Rhodesian undulant fever. Transactions of the Royal Society for Tropical Medicine and Hygiene 24, 93-95. Bishop, G.C., Bosman, P.P., 1994. Brucella suis infection. In: Coetzer, J.A.W., Thomson, G.R., Tustin, R.C. (Eds.), Infectious Diseases of Livestock with
159
special reference to Southern Africa II. Oxford University Press, London, London, 1076-1077 pp. Bishop, G.C., Bosman, P.P., Herr, S., 1994. Bovine Brucellosis. In: Coetzer,J.A.W, Thomson, G.R., Tustin, R.C. (Eds.), Infectious Diseases of Livestock with special reference to Southern Africa II. Oxford University Press, Cape Town, pp. 1053-1066. Bishop, G.C., Bosman, P.P., Herr, S., 1994. Brucella melitensis. In: Coetzer, J.A.W., Thomson, G.R., Tustin, R.C. (Eds.), Infectious diseases of livestock with special reference to Southern Africa II. Oxford University Press, Cape Town, pp. 1067- 1075. Blasco, J.M., Garin-Bastuji, B., Marin, C.M., Gerbier, G., Finlo, J., Jimenez De Bagues, M.P. and Cau, C. 1994. Efficacy of different rose Bengal and complement fixation antigens for the diagnosis of Bruella melitensis infection in sheep and goats. Veterinary Record, 134:415-420. Blood, D.C., Radostits, O.M., 1989. Veterinary Medicine. A textbook of the Diseases of Cattle, Sheep, Pigs, Goats and Horses, 7th Edition, Bailliere Tindall, London, 677-690 pp. Bricker, B.J., 2002. PCR as a diagnostic tool for brucellosis. Veterinary Microbiology 90, 435-446. Bricker, B.J., Ewalt, D.R., Halling, S.M., 2003. Brucella 'HOOF-Prints': strain typing
by multi-locus analysis of variable number tandem repeats (VNTRs). BMC
Microbiology, 3: 1-13.
Bricker, B.J., Ewalt, D.R., MacMillan, A.P., Foster, G., Brew, S., 2000. Molecular characterization of Brucella strains isolated from marine mammals. Journal of Clionical Microbiology 38, 1258-1262. Bricker, B.J., Halling, S.M., 1994. Differentiation of Brucella abortus bv. 1, 2, and 4, Brucella melitensis, Brucella ovis, and Brucella suis bv. 1 by PCR. Journal of Clinical Microbiology 32, 2660-2666. Brinley Morgan, W.J., 1967. The serological diagnosis of bovine brucellosis. The Veterinary Record 80, 612-620. Brinley Morgan, W.J., 1977. The diagnosis of Brucella abortus Infection in Britain. . In: Bovine Brucellosis: An International Symposium, Texas A & M University Press, College Station, London, pp. 21-39.
160
Bryant, B.A., Norval, R.A.I., 1985. Diseases affecting domestic animals in communal lands in Manicaland. Zimbabwe Veterinary Journal 16, 9-17. Buddle, M.B., 1953. A Brucella mutant causing genital diseases of sheep in New Zealand. Australian Veterinary Journal 29, 145-153. Buddle, M.B., 1956. Studies on Brucella ovis (N.SP.), a cause of genital disease of sheep in New Zealand. Journal of Hygiene 54, 351-364. Burgess, G.W., Norris, M.J., 1982. Evaluation of the cold complement fixation test for the diagnosis of ovine brucellosis. Australian Veterinary Journal 59, 23-25. Burgess, G.W., Spencer, T.L., Norris, M.J., 1985. Experimental infection of goats with Brucella ovis. Australian Veterinary Journal 62, 262-264. Carmichael, L.E., 1966. Abortion in 200 Beagles. Journal of the American Veterinary Medical Association 149, 1126. Carmichael, L.E., Kenney, R.M., 1968. Canine abortions caused by Brucella canis. Journal of the American Veterinary Medical Association 152, 605-616. Carter, G.R., Chengappa, M.M., 1991. Brucella. In: Essentials of Veterinary Bacteriology and Mycology, 4 Edition. Lea and Febiger (UK) Ltd, UK, 196-201 pp. Celli, J., Gorvel, J.-P., 2004. Organelle robbery: Brucella interactions with the endoplasmic reticulum. Current Opinion in Veterinary Microbiology 7, 93-97. Chenje, M., Sola, L., Paleczny, D., 1998. The State of Zimbabwe's Environment. Ministry of Mines, Environment and Tourism, Karigamombe Centre, P.O. Box CY7753, Causeway, Harare. Christie, T.E., 1969. Eradication of brucellosis in Northern Ireland: field problem and experience. The Veterinary Record 85, 268-269. Chukwu, C.C., 1985. Brucellosis in Africa Part I: The Prevalence. Bulletin of Animal Health and Production in Africa 33, 193-198. Chukwu, C.C., 1987. Brucellosis in Africa, Part II: The importance. Bulletin of Animal Health and Production in Africa 35, 92-98. Clavareau, C., Wellemans, V., Walravens, K., Tryland, M., Verger, J.M., Grayon, M., Cloeckaert, A., Letesson, J.J., Godfroid, J., 1998. Phenotypic and molecular characterization of a Brucella strain isolated from a minke whale (Balaenoptera acutorostrata). Veterinary Microbiology, 144, 3267-3273. Claxton, P.D., 1968. A comparison of two commercial vaccines and two methods of vaccination. Australian Veterinary Journal 44, 48-54.
161
Cloeckaert, A., Verger, J.M., Grayon, M., Paquet, J.Y., Garin-Bastuji, B., Foster, G., Godfroid, J., 2001. Classification of Brucella spp. isolated from marine mammals by DNA polymorphism at the omp2 locus. Microbes and Infection, 3, 729-738. Coetzer, J.A.W., de Vos, V., Kriek, N.P., Tustin, R.C., Swanepoel, R., Picard, J.A. 2000. Bovine Brucellosis. In: Selected Animal Infectious Diseases (CD-ROM). Department of Veterinary Tropical Diseases, University of Pretoria, South Africa. Collier, J.R., Molello, J.A., 1964. Comparative Distribution of Brucella abortus, Brucella melitensis and Brucella ovis in experimentally infected pregnant ewes. American Journal of Veterinary Research 25, 930-934. Condy, J.B., Vickers, D.B., 1972. Brucellosis in Rhodesian Wildlife. Journal of the South African Veterinary Association 43, 175-179. Corbel, M.J., 1997. Brucellosis: an overview. Emerging Infectious Diseases 3, 213-221. Corbel, M.J., Brinley-Morgan, W.J., 1984. Genus Brucella Meyer and Shaw 1920, 173AL, In: Bergey's Manual for Classification of Systematic Bacteriology, Vol. 1, Williams and Wilkins, London, , 377-388 pp. Corbel, M.J., Stuart, F.A., Brewer, R.A., Jeffrey, M., Bradley, R., 1989. Arthropathy associated with Brucella abortus strain 19 vaccination in cattle. Examination of field cases. British Veterinary Journal 145, 337. Crawford, R.P., Huber, J.D., Adams, B.C., 1990. Epidemiology and surveillance. In: Nelson, K.E., Ducan, J.R. (Eds.) Animal brucellosis. CRC press, Florida, pp. 131-151. Cunningham, B., 1977. A difficult disease called Brucellosis In: Bovine Brucellosis: An International Symposium, Texas A & M University Press, College Station, London, pp. 11-20. Cutler, S. Whatmore, A.M., A.J., C., Commander, N.J. 2005. Brucellosis- a new aspect of an old disease. Journal of Applied Microbiology 98, 1270-1281. Dajer, A., Luna-Martinez, E., Zapata, D., Villegas, S., Gutierrez, E., Pena, G., Gurria, F., Nielsen, K., Gall, D., 1999. Evaluation of a fluorescence-polarization assay for the diagnosis of bovine brucellosis in Mexico. Preventive Veterinary Medicine 40, 67-73.
162
Dargatz, D.A., Smith, J.A., Knight, A.P., Farin, P.W., Kimberling, C.V., 1990. Antimicrobial therapy for rams with Brucella ovis infection of the urogenital tract. Journal of the American Veterinary Medical Association 196, 605-610. Davis, D.S., Elzer, P.H., 2002. Brucella vaccines in wildlife. Veterinary Microbiology 90, 533-544. De Ley, J., Mannheim, W., Segers, P., Lievens, A., Deninj, M., Vanhoucke, M., Gillis, M., 1987. Ribosomal nucleic acid cistron similarlities and taxonomic neighbourhood of Brucella and CDC group Vd. International Journal of Systematic Microbiology 37, 35-42. Delrue, R.M., Lestrate, P., Tibor, A., Letesson, J.-J., De Bolle, X., 2004. Brucella pathogenesis, genes identified from random large-scale screens. FEMS Microbiology Letters 231, 1-12. Deqiu, S., Donglou, X., Jiming, Y., 2002. Epidemiology and control of brucellosis in China. Veterinary Microbiology 90, 165-182. Diaz, R., Moriyon, I., 1989. Laboratory Techniques in the Diagnosis of Human Brucellosis. In: Brucellosis: Clinical and Laboratory Aspects, Young, E.J. and Corbel, M.J. (Eds), CRC Press Inc., 73-84 pp. Doganay, M., Aygen, B., 2003. Human brucellosis: an overview. International Journal of Infectious Diseases 7, 173-182. Dohoo, I., Martin, W., Stryhn, H., 2003. Veterinary Epidemiologic Research. AVC Inc., Charlottetown. Duncan, J.R., Wilkie, B.N., Hiestand, F., Winter, A.J., 1972. The serum and secretory immunoglubulins of cattle:characterisation and quantitation. Journal of Immunology 108, 965-976. Elber, S., 1981. Rev. 1 Brucella melitensis vaccine, Part II: 1968-1980. The Veterinary Bulletin 51, 67-73. Enright, F.M., Walker, J.V., Jeffers, G., Deyoe, B.L., 1984. Cellular and humoral responses of Brucella abortus infected bovine fetuses. American Journal of Veterinary Research 45, 424-430. Ewalt, D.R., Payeur, J.B., Rhyan, J.C., Geer, P.L., 1997. Brucella suis biovar 1 in naturally infected cattle: a bacteriological, serological, and histological study. Journal of Veterinary Diagnostic Investigation 10, 417-420.
163
Farrel, I.D., 1974. The Development of a new selective medium for the isolation of Brucella abortus from contaminated sources. Research in Veterinary Science 16, 280-286. Fayazi, Z., Ghadersohi, A., Hirst, R.G., 2002. Development of a Brucella suis specific hybridisation probe and PCR which distinguishes B. suis from B. abortus. Veterinary Microbiology 84, 253-261. Faye, B., Castel, V., Lesnoff, M., Rutabinda, D., Dhalwa, J., 2005. Tuberculosis and brucellosis prevalence survey on dairy cattle in Mbarara milk basin (Uganda). Preventive Veterinary Medicine 67, 267-281. Fekete, A., Bantle, J.A., Halling, S.M., Sanborn, S.R., 1992. Amplification fragment length polymorphism in Brucella strains by use of polymerase chain reaction with arbitrary primers. Journal of Bacteriology 174, 7778-7783. Fernando, S., Sportsman, J., Wilson, G., 1992. Studies of the low dose hook effect in a competitive immunosorbent assay. Journal of Immunological Methods 151, 27- 27. Ficht, T.A., 2003. Intracellular survival of Brucella: defining the link with persistence. Veterinary Microbiology 92, 213-223. Foster, G., Jahans, K.L., Reid, R.J., Ross, H.M., 1996. Isolation of Brucella species from cetaceans, seals and an otter. The Veterinary Record 138, 583-586. Foster, G., MacMillan, A.P., Godfroid, J., Howie, F., Ross, H.M., Cloeckaert, A., Reid, R.J., Brew, S., Patterson, I.A.P., 2002. A review of Brucella spp. infection of sea mammals with particular emphasis on isolates from Scotland. Veterinary Microbiology 90, 563-580. Foster, G., Osterman, B.S., Godfroid, J., Jacques, I., Cloeckaert, A., 2007. Brucella ceti sp. nov. and Brucella pinnipedialis sp. nov. for Brucella strains with cetaceans and seals as their preferred hosts. International Journal of Systematic and Evolutionary Microbiology 57, 2688-2693. Franklin, J.E.F., 1965. Brucella abortus in bulls: A study of twelve naturally infected cases. The Veterinary Record 77, 132-135. Gall, D., Nielsen, K., Forbes, L., Davis, D., Elzer, P., Olsen, S., Balsevicius, S., Kelly, L., Smith, P., Tan, S., Joly, D. 2000. Validation of the fluorescence polarization assay and comparison to the other serological assays for the detection of serum antibodies to Brucella abortus in bison. Journal of Wildlife Diseases, 36:469- 476.
164
Gallangher, J., 1973. The Rose Bengal Plate Agglutination test in dairy cattle in Zambia vaccinated over age with strain 19 Brucella abortus. Tropical Animal Health and Production 5, 253-258. Garritty, G.M., Bell, J.A., Lilburn, T., 2005. Family III, Brucellaceae Breed, Murray and Smith 1957, 394AL. In: Bergey’s Manual of Systematic Bacteriology, Volume II, (2nd Ed.). Brenner, D.J., Krieg, N.R. and Staley, J.T.(Ed.), Springer Science +Business Media, Inc., New York, NY 10013, USA, 370- 392. pp. Godfroid, J., 2002. Brucellosis in wildlife. Revue Scientifique Et Technique De Office International Des Epizooties 21, 277-286. Godfroid, J., Michel, P., Uytterhaegen, L., Desmedt, C., Rasseneur, F., Boelaert, F., Saegerman, C., Patigny, X., 1994. Brucella suis biotype 2 infection of wild boars (Sus scrofa) in Belgium. Annales De Medecine Veterinaire 138, 263-268. Godfroid, J., Saegerman, C., Wellemans, V., Walravens, K., Letesson, J.J., Tibor, A., Mc Millan, A., Spencer, S., Sanna, M., Bakker, D., Pouillot, R., Garin-Bastuji, B., 2002. How to substantiate eradication of bovine brucellosis when aspecific serological reactions occur in the course of brucellosis testing. Veterinary Microbiology 90, 461-477. Golding, B., Scott, D.E., Scharf, O., Huang, L.-Y., Zaitseva, M., Lapham, C., Eller, N., Golding, H., 2001. Immunity and protection against Brucella abortus. Microbes and Infection 3, 43-48. Gordon, J.C., Pue, H.L., Rutgers, H.C., 1985. Canine brucellosis in a household. Journal of the American Veterinary Medical Association 186, 695-698. Gous, T.A., van Rensburg, W.J., Gray, M., Perrett, L.L., Brew, S.D., Young, E.J., Whatmore, A.M., Gers, S., Picard, J., 2005. Brucella canis in South Africa. The Veterinary Record 157, 668. Gul, S.T., Khan, A., 2007. Epidemiology and Epizootology of Brucellosis: A review. Pakistan Veterinary Journal 27, 145-151. Hall, W.H., 1989. History of Brucella as a human pathogen. In: Brucellosis: Clinical and Laboratory Aspects, Young, E.J. and Corbel, M.J. (Eds), CRC Press, 1-10 pp. Hasoda, K., Yasuda, T., 1989. Homogenous immunoassay for alpha2 plasmin inhibitor (alpha2 PI) and alpha2 plasmin complex. Application of a sandwich liposome immune lysis assay (LILA) technique. Journal of immunological methods 121, 121-121.
165
Herr, S., 1994. Brucella melitensis infection. In: Coetzer, J.A.W., Thomson, G.R., Tustin, R.C. (Eds.), Infectious diseases of livestock with special reference to Southern Africa II. Oxford University Press, Cape Town,1073-1075 pp. Herr, S., Lawrence, J.V., Brett, O.L., Ribeiro, L.M., 1991. A serological comparison of complement fixation reactions using Brucella abortus and B. melitensis antigens in B. abortus infected cattle. Onderstepoort Journal of Veterinary Research 58, 111-114. Hill, W.K. 1963. Standardisation of the complement fixation test for brucellosis. Bulletin of the Organization of International Epizootics ( O.lE), 60:401. Holt, J.G., Krieg, N.R., Sneath, P.H., Williams, S.T., 1994. Gram-Negative Aerobic/Microaerophilic Rods and Cocci. In: Bergey’s Manual of Determinative Bacteriologyt, 9 Edition, Wiliams and Wilkins, 428 East Preston Street, Baltimore, Maryland, 21202, USA, 71-174 pp. Hornsby, R.L., Jensen, A.E., Olsen, S.C., Thoen, C.O., 2000. Selective media for isolation of Brucella abortus strain RB51. Veterinary Microbiology 73, 51-60. Huddleston, F., 1946. The mucoid phase of the genus Brucella. American Journal of Veterinary Research 7, 5-10. Hughes, K.L., 1972. Experimental Brucella ovis infection in ewes. Australian Veterinary Journal 48, 12-17. Hughes, K.L., Claxton, P.D., 1968. Brucella ovis infection: An evaluation of microbiological, serological and clinical methods of diagnosis in the ram. Australian Veterinary Journal 44, 41-47. Hunter, D., Allan, J., 1972. An evaluation of milk and blood tests used to diagnose brucellosis. The Veterinary Record 91, 310-312.
Jacques, I., Olivier-Bernardin, V. and Dubray, G. 1998. Efficacy of ELISA compared to conventional tests (RBPT and CFT) for the diagnosis of Brucella melitensis infection in sheep. Veterinary Microbiology, 64:61-73. Jahans, K.L., Foster, G., Broughton, E.S., 1997. The characterisation of Brucella strains isolated from marine mammals. Veterinary Microbiology 57, 373-382. Jiwa, S.F.H., Kazwala, R.R., Tungaraza, R., Kimera, S.I., Kalaye, W.J., 1996. Bovine brucellosis serum agglutination test prevalence and breed disposition according to prevalent management systems in the Lake Victoria zone of Tanzania. Preventive Veterinary Medicine 26, 341-346.
166
Jordan, D. 1995. Herdacc: A programme for calculating herd level (aggregate) sensitivity and specificity. (Guelph, Canada, Department of Population Medicine, University of Guelph). Kabagambe, E.K., Elzer, P.H., Geaghan, J.P., Opuda-Asibo, J., Scholl, D.T., Miller, J.E., 2001. Risk factors for Brucella seropositivity in goat herds in eastern and western Uganda. Preventive Veterinary Medicine 52, 91-108. Kadohira, M., McDermott, J.J., Shoukri, M.M., Kyule, M.N., 1997. Variations in the prevalence of antibody to brucella infection in cattle by farm, area and district in Kenya. Epidemiology and Infection 118, 35-41. Kagumba, M., Nandokha, E., 1978. A survey of the prevalence of bovine brucellosis in East Africa. Bulletin of Animal Health and Production in Africa 26, 224-229. Kohler, S., Michaux-Chalachon, S., Porte, F., Ramuz, M., Liautard, J.-P., 2003. What is the nature of the replicative niche of the stealthy bug named Brucella?. Trends in Microbiology 11, 215-219. Kubuafor, D.K., Awumbila, B., Akanmori, B.D., 2000. Seroprevalence of brucellosis in cattle and humans in Akwapim-South district of Ghana: public health implications. Acta Tropica 76, 45-48. Kumi-Diaka, J., Bale, O.O.J., Ogwu, D., Osori, D., 1980. The effect of Brucella abortus on spermatogeneis in the Zebu bulls (Bos indicus): A case report. . Theriogenelogy 14 167-171. Kuplulu, O., Sarimehmetoglu, B., 2004. Isolation and identification of Brucella spp. in ice cream. Food Control 15, 511-514. Lambert, G., Manthei, C.A., Deyoe, B.L., 1963. Studies on Brucella abortus infection in bulls. American Journal of Veterinary Research 24, 1152-1157. Lapraik, R.D., 1975. Brucellosis: A study of five calves from reactor dams. The Veterinary Record 97, 52-54. Lapraik, R.D., 1982. Latent bovine brucellosis. The Veterinary Record 111, 578-579. Leclerc, V., Dufour, B., Lombard, B., Gauchard, F., Garin-Bastuji, B., Salvat, G., Brisabois, A., Poumeyrol, M., De Buyser, M.L., Gnanou-Besse, N., Lahellec, C., 2002. Pathogens in meat and milk products: surveillance and impact on human health in France. Livestock Production Science 76, 195–202. Lee, K., Cargill, C., Atkinson, H., 1985. Evaluation of an enzyme-linked immunosorbent assay for the diagnosis of Brucella ovis infection in rams. Australian Veterinary Journal 62, 91-93.
167
Leyla, G., Kadri, G., Umran, O., 2003. Comparison of polymerase chain reaction and bacteriological culture for the diagnosis of sheep brucellosis using aborted fetus samples. Veterinary Microbiology 93, 53-61. Libal, M.C., Kirkbride, C.A., 1983. Brucella ovis-induced abortion in ewes. Journal of the American Veterinary Medical Association 183, 553-554. Lucero, N.E., Escobar, G.I., Ayala, S.M., Silva Paulo, P., Nielsen, K., 2003. Fluorescence polarization assay for diagnosis of human brucellosis. Journal of Medical Microbiology 52, 883-887. Lucero, N.E., Jacob, N.O., Ayala, S.M., Escobar, G.I., Tuccillo, P., Jacques, I., 2005. Unusual clinical presentation of brucellosis caused by Brucella canis. Journal of Medical Microbiology 54, 505-508. MacDiarmid, S.C., Hellstrom, J.S., 1987. An intradermal test for the diagnosis of Brucellosis in extensively managed cattle herds. Preventive Veterinary Medicine 4, 361-369. Madsen, M., 1989. The current status of brucellosis in Zimbabwe. Zimbabwe Veterinary Journal 20, 133-145. Madsen, M., Anderson, E.C., 1995. Serologic survey of Zimbabwean wildlife for brucellosis. Journal of Zoo and Wildlife Medicine 26, 240-245. Madzima, W.M. 1987. Zimbabwe:Bovine brucellosis and brucellosis of small ruminants: Diagnosis, control and vaccination. In: Technical series , Office International des Epizooties (OIE)), (Paris, France, pp. 80-82. Manley, F.H. 1969. Brucellosis in Rhodesia. A report to the Director of Veterinary Services (Salisbury). Marin, C.M., Alabart, J.L., Blasco, J.M., 1996. Effect of antibiotics contained in two Brucella selective media on growth of Brucella abortus, B.melitensis, and B.ovis. Journal of Clinical Microbiology 34, 426-428. Marin, C.M., Debagues, M.P.J., Blasco, J.M., Gamazo, C., Moriyon, I., Diaz, R., 1989. Comparison of 3 serological tests for Brucella ovis infection of rams using different antigenic extracts. TheVeterinary Record 125, 504-508. McDermott, J.J., Arimi, S.M., 2002. Brucellosis in sub-Saharan Africa: epidemiology, control and impact. Veterinary Microbiology 90, 111-134. McDermott, J.J., Deng, K.A., Jayatileka, T.N., El Jack, M.A., 1987. A cross-sectional cattle disease study in Kongor rural council, southern Sudan. I. prevalence
168
estimates and age, sex and breed associations for brucellosis and contagious bovine pleuropneumonia. Preventive Veterinary Medicine 5, 111-123. McEwen, A.D., Priestley, F.W., 1938. Experiments on contagious abortion: Immunisation studies with vaccines of graded virulence. The Veterinary Record 50, 1097-1106. McEwen, A.D., Priestley, F.W., 1940. The vaccination of Guinea pigs against Brucella abortus infection with living and heat-killed suspensions. The Veterinary Record 52, 743-744. McGiven, J.A., Tucker, J.D., Perrett, L.L., Stack, J.A., Brew, S.D., MacMillan, A.P., 2003. Validation of FPA and cELISA for the detection of antibodies to Brucella abortus in cattle sera and comparison to SAT, CFT, and iELISA. Journal of Immunological Methods 278, 171-178. Meinershagen, W., Frank, F.W., Waldhalin, D.G., 1974. Brucella ovis as a cause of abortion in ewes. American Journal of Veterinary Research 35, 723-724. Mercier, E., JumasBilak, E., AllardetServent, A., Ocallahan, D., Ramuz, M., 1996. Polymorphism in Brucella strains detected by studying distribution of two short repetitive DNA elements. Journal of Clinical Microbiology 34, 1299-1302. Meyer, M.E., Cameron, H.S., 1961. Metabolic characterisation of the genus Brucella. II. Oxidative metabolic patterns of the described species. Journal of Bacteriology 82, 396. Mikolon, A.B., Gardner, I.A., Hietala, S.K., De Anda, J.H., Pestana, E.C., Hennager, S.G., Edmondson, A.J., 1998. Evaluation of North American antibody detection tests for diagnosis of brucellosis in goats. Journal of Clinical Microbiology 36, 1716-1722. Milward, F.W., Nicoletti, P., Hoffman, E., 1984. Effectiveness of various therapeutic regimens for bovine brucellosis. American Journal of Veterinary Research 45, 1825-1828. Minas, A., Stournara, A., Minas, M., Papaioannou, A., Krikelis, V., Tselepidis, S., 2005. Validation of fluorescence polarization assay (FPA) and comparison with other tests used for diagnosis of B. melitensis infection in sheep. Veterinary Microbiology 111, 211-221. Mohan, K., Makaya, P.V., Muvavarirwa, P., Matope, G., Mahembe, E., Pawandiwa, A., 1996. Brucellosis surveillance and control in Zimbabwe: bacteriological and
169
serological investigation in dairy herds. Onderstepoort Journal of Veterinary Research 63, 47-51. Moreno, E., Cloeckaert, A., Moriyon, I., 2002. Brucella evolution and taxonomy. Veterinary Microbiology 90, 209-227. Moreno, E., Stackebrandt, E., Dorsch, M., Wolters, J., Busch, M., Mayer, H., 1990. Brucella abortus 16S rRNA and lipid A reveal a phylogenetic relationship with members of the alpha-2 subdivision of the class Proteobacteria. Journal of Bacteriology 172, 3569-3576. Moyer, N.P., Holocomb, L.A., 2005. Brucella. In: Manual of Clinical Microbiology, 6th Ed., Murray, P.R., Jo Barron, E., Pfaller, M.A., Tenover, F.C. and Yolken, R.H. (Eds), ASM Press, Washington DC 20005, USA, 549-555 pp. Msanga, J.F., Mukangi, D.J.A., Tungaraza, R., 1986. Bovine brucellosis in the Lake Zone of Tanzania: The present situation. Bulletin of Animal Health and Production in Africa 34, 230-234. Muma, J.B., Samui, K.L., Siamudaala, V.M., Oloya, J., Matope, G., Omer, M.K., Munyeme, M., Mubita, C., Skjerve, E., 2006. Prevalence of antibodies to Brucella spp. and individual risk factors of infection in traditional cattle, goats and sheep reared in livestock-wildlife interface areas of Zambia. Tropical Animal Health and Production 38, 195-206. Muma, J.B., Toft, N., Oloya, J., Lund, A., Nielsen, K., Skjerve, E., 2007. Evaluation of three serological tests for brucellosis in naturally infected cattle using latent class analysis. Veterinary Microbiology 125, 187-192. Munroe, F.A., Dohoo, I.R., McNab, W.B., Sprangler, L., 1999. Risk factors for the between herd Mycobacterium in Canadian cattle and cervids between 1985 and 1994. Preventive Veterinary Medicine 41, 119-133. Murray, R.M., 1969. Scrotal Abnormalities in rams in Tropical Queensland with particular reference to ovine Brucellosis and its control. Australian Veterinary Journal 45, 63-67. Murray, P.R., Baron, E.J., Pfaller, M. A., Tenover, F.C. and Yolken, R.H. 1999. Manual of Clinical Microbiology, 7th Ed. (ASM Press, Washington DC), 625-631. Neil, S.D., Hanna, J., O’Brien, J.J., McCracken, R.M., 1989. Transmission of tuberculosis from experimentally infected cattle to in-contact calves. Veterinary Record 124, 269-271.
170
Nelson, C., 1977. Immunity to Brucella abortus. In: Bovine Brucellosis: An International Symposium, Texas A & M University Press, College Station, London, pp. 177-188. Nicoletti, P., 1967. Utilisation of the card test in brucellosis eradication. Journal of the American Veterinary Medical Association 151, 1778-1781. Nicoletti, P., 1980. The epidemiology of bovine brucellosis. Advances in veterinary
science and Comparative Medicine, 24:69-98.
Nicoletti, P., 1984. The control of bovine brucellosis in tropical and subtropical regions. Preventive Veterinary Medicine 2, 193-196. Nicoletti, P., 2002. A short history of brucellosis. Veterinary Microbiology 90, 5-9. Nicoletti, P., Jones, L.M., Berman, D.T., 1978. Comparison of the subcutaneous and conjunctival route of vaccination with Brucella abortus S19 vaccine in adult cattle. Journal of the American Veterinary Medical Association 173, 1450-1456. Nielsen, K., 2002. Diagnosis of brucellosis by serology. Veterinary Microbiology 90, 447-459. Nielsen, K., Cherwonogrodzky, J.W., Duncan, J.R., Bundle, D.R., 1989. Enzyme-linked immunosorbent assay for differentiation of the antibody response of cattle naturally infected with Brucella abortus or vaccinated with strain 19. American Journal of Veterinary Research 50, 5-9. Nielsen, K., Gall, D., 2001. Fluorescence polarization assay for the diagnosis of brucellosis: A review. Journal of Immunoassay and Immunochemistry 22, 183- 201. Nielsen, K., Gall, D., Bermudez, R., Renteria, T., Moreno, F., Corral, A., Monroy, O., Monge, F., Smith, P., Widdison, J., Mardrueno, M., Calderon, N., Guerrero, R., Tinoco, R., Osuna, J., Kelly, W., 2002. Field trial of the brucellosis fluorescence polarization assay. Journal of Immunoassay and Immunochemistry 23, 307-316. Nielsen, K., Gall, D., Jolley, M., Leishman, G., Balsevicius, S., Smith, P., Nicoletti, P., Thomas, F., 1996a. A homogeneous fluorescence polarization assay for detection of antibody to Brucella abortus. Journal of Immunological Methods 195, 161-168. Nielsen, K., Gall, D., Lin, M., Massangill, C., Samartino, L., Perez, B., Coats, M., Hennager, S., Dajer, A., Nicoletti, P., Thomas, F., 1998. Diagnosis of bovine
171
brucellosis using a homogeneous fluorescence polarization assay. Veterinary Immunology and Immunopathology 66, 321-329. Nielsen, K., Gall, D., Smith, P., Vigliocco, A., Perez, B., Samartino, L., Nicoletti, P., Dajer, A., Elzer, P., Enright, F., 1999. Validation of the fluorescence polarization assay as a serological test for the presumptive diagnosis of porcine brucellosis. Veterinary Microbiology 68, 245-253. Nielsen, K., Smith, P., Widdison, J., Gall, D., Kelly, L., Kelly, W., Nicoletti, P., 2004. Serological relationship between cattle exposed to Brucella abortus, Yersinia enterocolitica O : 9 and Escherichia coli O157 : H7. Veterinary Microbiology 100, 25-30. Nielsen, K.H., Kelly, L., Gall, D., Balsevicius, S., Bosse, J., Nicoletti, P., Kelly, W., 1996b. Comparison of enzyme immunoassays for the diagnosis of bovine brucellosis. Preventive Veterinary Medicine 26, 17-32. Nielsen, K.H., Kelly, L., Gall, D., Nicoletti, P., Kelly, W., 1995. Improved competitive enzyme immunoassay for the diagnosis of bovine brucellosis. Veterinary Immunology and Immunopathology 46, 285-291. Nielsen, O., Stewart, R.E., Nielsen, K., Measures, L., Duignan, P., 2001. Serologic survey of Brucella spp. antibodies in some marine mammals of North America. Journal of Wildlife Diseases 37, 89-100. Nielsen, K., Gall, D., Smith, P., Bermudes, R., Moreno, F., Renteria, T., Ruiz, A., Aparicio, L., Vazquez, S., Dajer, A., Luna, E., Samartino, L., Halbert, G. 2005. Evaluation of serological tests for detection of caprine antibody to Brucella melitensis. Small Ruminants Research, 56: 256-258.
Noordhuizen, J.P.T., Frankena, K., van der Hoofd, C.M., and Graat, E.A.M. 1997.
Application of Quantitative Methods in Veterinary Epidemiology. Wageningen
Pers, Wageningen, 31-62.
OIE, 2004. Manual of the Diagnostic Tests and vaccines for Terrestial animals, Vol 1, 5 Edition. Office International Des Epizooties, Paris, France, 409-438 pp. Oliveira, S.C., Soeurt, N., Splitter, G., 2002. Molecular and cellular interactions between Brucella abortus antigen and host immune responses. Veterinary Microbiology 90, 417-424.
172
Oloya, J., Opuda-Asibo, J., Djonne, B., Muma, J.B., Matope, G., Kazwala, R., Skjerve, E., 2006. Responses to tuberculin among Zebu cattle in the transhumance regions of Karamoja and Nakasongola district of Uganda. Tropical Animal Health and Prodroduction 38, 275-283. Olsen, S.C., Stoffregen, W.S., 2005. Essential role of vaccines in brucellosis control and eradication programs for livestock. Expert Review on Vaccines 4, 915-928. Omer, M.K., Skjerve, E., Woldehiwet, Z., Holstad, G., 2000. Risk factors for Brucella spp. infection in dairy cattle farms in Asmara, State of Eritrea. Preventive Veterinary Medicine 46, 257-265. Oncel, T., 2005. Seroprevalence of Brucella canis infection in two provinces in Turkey. Turkish Veterinary Journal of Veterinary and Animal Science 29, 779-783. Osterman, B., Moriyon, I., 2003. International Committee on Systematics of Prokaryotes- Subcommitte on the taxonomy of Brucella. International Journal of Systematic and Evolutionary Microbiology 56, 1173-1175. Palmer, M.V., Cheville, N., Jensen, A., 1996. Experimental infection of pregnant cattle with vaccine candidate Brucella abortus RB51: Pathologic, bacteriologic and serologic findings. Veterinary Pathology 33, 682-691. Pandey, G. S., Kobayashi, K. , Nomura, Y., Nambota, A., Mwima, H.K., Suzuki, K.
1999. Studies on sero-prevalence of brucellosis in Kafue lechwe (Kobus leche
kafuensis in Zambia). Indian Veterinary Journal 76: 275-278.
Pappas, G., Papadimitriou, P., Akritidis, N., Christou, L., Tsianos, E.V., 2006. The new global map of human brucellosis. Lancet of Infectious Diseases 6, 91-99. Plant, J.W., Eamens, G.J., Seaman, J.T., 1986. Serological, bacteriological and pathological changes in rams following different routes of exposure to Brucella ovis. Australian Veterinary Journal 63, 409-412. Quinn, P.J., Carter, M.E., Markey, B., Carter, G.R., 1999. Clinical Veterinary Microbiology. Mosby International Limited, Edinburgh, 261-267 pp. Raad, I., Rand, K., Gaskins, D., 1990. Buffered charcoal-yeast extract medium for isolation of brucellae. Journal of Clinical Microbiology 28, 1671-1672. Rahaley, R.S., Dennis, S.M., 1984. Histopathology of experimental brucellosis in rams following vaccination with Brucella ovis. Australian Veterinary Journal 61, 353-356.
173
Ramirez-Pfeiffer, C., Nielsen, K., Marin-Ricalde, F., Rodriguez-Padilla, C., Gomez- Flores, R., 2006. Comparison of fluorescence polarisation assay with card and complement fixation tests for the diagnosis of goat brucellosis in a high- prevalence area. Veterinary Immunology and Immunopathology 121-127, 121- 127. Reviriego, F.J., Moreno, M.A., Dominguez, L., 2000. Risk factors for brucellosis seroprevalence of sheep and goat flocks in Spain. Preventive Veterinary Medicine 44, 167-173. Rogers, R.J., Cook, D.R., Ketterer, P.J., Baldock, F.C., Blackall, P.J., Stewart, R.W., 1989. An evaluation of three serological tests for antibody to Brucella suis in pigs. Australian Veterinary Journal 66, 77-80. Ross, H.M., Foster, G., Reid, R.J., 1994. Brucella species infection in sea mammals. The Veterinary Record 134, 359. Ross, H.M., Jahans, K.L., MacMillan, A.P., Reid, R.J., Thompson, P.M., Foster, G., 1996. Brucella species infection in North Sea seal and cetacean populations. The Veterinary Record 138, 647-648. Sachs, R., Staak, C., Croocock, C.M. 1968. Serological Investigation of brucellosis in
game animals in Tanzania. Bulletin of Epizootic Diseases in Africa 16: 93-100.
Salman, M.D., Meyer, M.E., 1984. Epidemiology of bovine brucellosis in the Mexicali Valley: literature review of disease-associated factors. American Journal of Veterinary Research 45, 1557-1560. Samartino, L., Gregoret, R., Gall, D., Nielsen, K., 1999. Fluorescence polarization assay: Application to the diagnosis of bovine brucellosis in Argentina. Journal of Immunoassay 20, 115-126. Saunders, C.N., 1958. Abortion and Stillbirths in pigs-An analysis of 67 outbreaks. The Veterinary Record 70, 965-970. Schelling, E., Diguimbaye, C., Daoud, S., Nicolet, J., Boerlin, P., Tanner, M., Zinsstag, J., 2003. Brucellosis and Q-fever seroprevalences of nomadic pastoralists and their livestock in Chad. Preventive Veterinary Medicine 61, 279-293. Schurig, G.G., Roop, R.M.I., Bagchi, T., Boyle, S., Buhrman, S., Sriranganathan, N., 1991. Biological properties of RB51; a stable rough strain of Brucella abortus. Veterinary Microbiology 28, 171-188.
174
Schurig, G.G., Sriranganathan, N., Corbel, M.J., 2002. Brucellosis vaccines: past, present and future. Veterinary Microbiology 90, 479-496. Searson, J.E., 1986. Distribution of Brucella ovis in the tissues of rams reacting in a complement fixation test for ovine brucellosis. Australian Veterinary Journal 63, 30-31. Shin, S., Carmichael, L.E. 1999. Canine brucellosis caused by Brucella canis. In: Recent Advances in Canine Infectious Diseases, Carmichael, L.E., ed. (International Veterinary Information Service (www.ivis.org)). Scholz HC, Hubalek Z, Sedlácek I, Vergnaud G, Tomaso H, Al Dahouk S, Melzer F, Kämpfer P, Neubauer H, Cloeckaert A et al: Brucella microti sp. nov., isolated from the common vole Microtus arvalis
Silva, I., Dangolla, A., Kulachelvy, K., 2000. Seroepidemiology of Brucella abortus infection in bovids in Sri Lanka. Preventive Veterinary Medicine 46, 51-59. Smith, H., Williams, A.E., Pearce, J.H., Keppie, J., Harris-Smith, P.W., Fitz-George, R.B., Witt, K. 1962. Foeta-erythritol: a cause of the localisation of Brucella abortus in bovine contagious abortion. Nature, 193:47-49. Sohn, A.H., Probert, W.S., Glaser, C.A., Gupta, N., Bollen, A.W., Wong, J.D., Grace, E.M., McDonald, W.C., 2003. Human neurobrucellosis with intracerebral granuloma caused by a marine mammal Brucella spp. Emerging Infectious Diseases 9, 485-488. Subcommittee, 1988. Subcommittee onTaxonomy of Brucella. International Journal of Systematic Bacteriology 38, 450-452. Swanepoel, R., Blackburn, N.K., Lander, K., 1975. Investigation of infectious infertility and abortion of cattle. Rhodesia (Zimbabwe) Veterinary Journal 6, 42-55. Swanepoel, R., Blackburn, N.K., Lander, K., 1976. The occurrence, diagnosis and control of Brucellosis in cattle in Rhodesia. Rhodesia Veterinary Journal 7, 24- 31. Taleski, V., Zerva, L., Kantardjiev, T., Cvetnic, Z., Erski-Biljic, M., Nikolvski, B., Bosnjakosvki, J., Katalinic-Jankovic, V., Panteliadou, A., Stojkoski, S., Kirandziski, T., 2002. An overview of the epidemiology and epizootiology of brucellosis in selected countries of Central and Southeast Europe. Veterinary Microbiology 90, 147-156.
175
Terzolo, H.R., Paolicchi, F.A., Moreira, A.R., Home, A., 1991. Skirrow agar for simultaneous isolation of Brucella and Campylobacter species. Veterinary Record 129, 531-532. Thim, B., Wundt, W., 1976. The epidemiological situation of brucellosis in Africa. In:
International Symposium on Brucellosis (II), Rabat 1975. Development of
Biological Standards, 31:201-217.
Thomsen, A., 1950. Experimental studies on the incubation period of infectious abortion in cattle. The British Veterinary Journal 106, 41-54. Timoney, J.F., Gillespie, J.H., Scott, F.W., Barlough, J.E., 1988. Hagan Bruner’s Microbiology and Infectious Diseases of Domestic Animals with Reference to aetiology, epizootiology, pathogenesis, immunity, diagnosis and antimicrobial susceptibility. Comstock Publishing Associates, London, 135-152 pp. Van Heerden, K.M., van Ransburg, S.W.J., 1962. The immunisation of rams against ovine Brucellosis. Journal of the South African Veterinary Medical Association 33, 43-148. Van Tonder, E.M., Herr, S., Bishop, G.C., Bosman, P.P., 1994. Brucella ovis infection. In: Infectious Diseases of Livestock with special reference to Southern Africa Vol 2. Oxford University Press, London, 1067-1072 pp. Verger, J.M., Grayon, M., Zundel, E., Lechopier, P., Olivier-Bernadin, V., 1995. Comparison of the efficacy of Brucella suis strain 2 and Brucella melitensis Rev. 1 live vaccines against a Brucella melitensis experimental infection in pregnant ewes. Vaccine 13, 191-196. Verger, J.M., Grimont, F., Grimont, P.A.D., Grayson, M., 1985. Brucella, a monospecific genus as shown by deoxyribonucleic acid hybridization. International Journal of Systematic Bacteriology 35, 292-295. Waghela, S., Wandera, J.G., Wagner, G.G. (1980). Comparison of four serological tests in the diagnosis of caprine brucellosis. Research in Veterinary Science, 28:168- 171. Walker, R.L., 1999, Brucella. In: Veterinary Microbiology, Hirsh, D.C. and Zee,
Y.C. (Eds.), Blackwell Science Inc., Massachusetts, USA, 196-203 pp. Wanke, M.M., 2004. Canine brucellosis. Animal Reproduction Science 82-83, 195-207.
176
Webb, R.F., Quinn, C.A., Cockram, F.A., Husband, A.J., 1980. Evaluation of procedures for the diagnosis of Brucella ovis infection in rams. Australian Veterinary Journal 56, 172-175. Werney, U., Kerani, A.A., Viertel, P., 1979. Bovine brucellosis in Southern Region of the Somali Democratic Republic. Tropical Animal Health and Production 11, 31-35. Wrathall, A.E., Broughton, E.S., Gill, K.P.W., Goldsmith, G.P., 1983. Serological reactions to Brucella species in British pigs. The Veterinary Record 132, 449- 454. Wright, A.E., Smith, F., 1897. Application of a serum test to the differential diagnosis of typhoid and Malta fever. Lancet 1, 656. Zhan, Y., Yang, J.A., Cheers, C., 1993. Cytokines response of T-cell subsets from Brucella abortus infected mice to soluble Brucella proteins. Infection and Immunity 61, 2841-2847.
177
APPENDIX
The structured questionnaire used to collect Brucella epizootological data
A SURVEY TO DETERMINE THE RISK FACTORS FOR BRUCELLOSIS IN
SMALLHOLDER CATTLE FARMS IN ZIMBABWE
Department of Paraclinical Veterinary Studies, University of Zimbabwe, P.O. Box MP
167, Mount Pleasant, Harare. Tel. (04) 303211 Extension 1753.
SECTION ONE: FARM IDENTIFICATION
1. Date of visit and interview:______
2. Farm Sampling No.:______
3. Farm name:______
4. Name of owner:______
5. Village Name:______
6. District of origin:______
7. Farm size (ha):______
8. Year of establishment:______
178
SECTION TWO: FARM STRUCTURE
9. Type of farm, eg dairy or mixed dairy and beef:______
10. Herd size:______
11. Cattle management types
0. breeds not mixed (indigenous Bos indicus only):______
1. mixed breed types (Bos indicus/Bos Taurus):______
12. Are cattle kept together with small stock (sheep and goats)?
0. no:______
1. yes:______
13. Animal census
Beef cattle Dairy Goats Sheep Pigs Other cattle animals
14. A breakdown of herd structure
Cattle Number Small ruminants Sheep Goats Cows Does/Ewes Bulls Bucks/rams Heifers>1yr Females>8months Steers/oxen Males>8months Calves<1yr Kids<8months
179
SECTION THREE: FARM MANAGEMENT AND CARE
A: GENERAL MANAGEMENT
15. What type of grazing does the farm use?
0. communal:______
1. own pasture:______
16. What type of feeding management do you use?
0. pasture only:______
1. Supplementary feeding and pasture:______
17. Water source?
0. communal:______
1. own supply:______
18. Who looks after the animals?
0. self:______
1. hired caretaker:______
19. How far away is the nearest farm?
0. less than 200 m:______
1. more than 200m:______
20. Do animals come in contact with wild animals?
0. no:______
1. yes:______
180
B: DISEASE CONTROL
21. Which methods do use to control ticks?
0. communal diptank:______
1. own spray:______
22. Where do you normally receive veterinary support?
0. Veterinary Department ______
1. AREX/DDP Officers______
2. None of the above______
23. Do you keep any records?
0. no:______
1. yes:______
24. Have you heard of brucellosis?.
0. no:______
1. yes:______
25. If yes, do you vaccinate your animals against brucellosis?
0. no:______
1. yes:______
C: BREEDING
26. What kind of breeding methods do you use on this farm?
0. artificial insemination: ______
1. natural methods: ______
27. Where do you get bulls?
0. use hired bulls:______
1. use own bull:______
181
28. Where do cows calve?
0. calving pens:______
1. on pasture:______
29. Have any animals aborted in the last three years?
0. no:______
1. yes:______(specify identity)
D. HOUSING
30. Where are animals kept overnight?
0. in pens:______
1. on pasture:______
31. How is manure removed?
0. taken to the fields:______
1. left in pens:______
D: MARKETING
32. Where do you sell milk?
0. local community:______
1. commercial processor.______
2. both above:______
33. Did you buy animals in the last three years?
0. no:______
1. yes:______
34. What is the source of your stock?
0. animal market:______
1. neighbours:______
182
2. others specify:______
35. Did you sell any animals in locality during the last three years?
0. no:______
1. yes:______
36. Do you hire animals for use?
0. no:______
1. yes:______
37. Do you lease animals to neighbours?
0. no:______
1. yes:______
38. Do you hire animals from neighbours for use?
0. no:______
1. yes:______