• 1.

    Shapiro BI , Gebru G , Desta S , Nigussie K , 2017. Rwanda Livestock Master Plan. Available at: http://extwprlegs1.fao.org/docs/pdf/rwa172923.pdf. Accessed June 30, 2020.

    • Search Google Scholar
    • Export Citation
  • 2.

    Dean AS , Crump L , Greter H , Hattendorf J , Schelling E , Zinsstag J , 2013. Clinical manifestations of human brucellosis: a systematic review and meta-analysis. PLoS Negl Trop Dis 612: e1929.

    • Search Google Scholar
    • Export Citation
  • 3.

    McDermott JJ , Arimi SM , 2002. Brucellosis in sub-Saharan Africa: epidemiology, control and impact. Vet Microbiol 90: 111134.

  • 4.

    McDermott JJ , Deng KA , Jayatileka TN , EL Jack MA , 1987. Cross-sectional cattle disease study in Kongor rural council, Southern Sudan. II. Brucellosis in cows: associated factors, impact on production and disease control considerations. Prev Vet Med 5: 125132.

    • Search Google Scholar
    • Export Citation
  • 5.

    Megersa B , Biffa D , Abunna F , Regassa A , Godfroid J , Skjerve E , 2011. Seroprevalence of brucellosis and its contribution to abortion in cattle, camel, and goat kept under pastoral management in Borana, Ethiopia. Trop Anim Health Prod 43: 651656.

    • Search Google Scholar
    • Export Citation
  • 6.

    Rujeni N , Mbanzamihigo L , 2014. Prevalence of brucellosis among women presenting with abortion/stillbirth in Huye, Rwanda. J Trop Med 2014: 13.

    • Search Google Scholar
    • Export Citation
  • 7.

    Gafirita J et al. 2017. Seroprevalence of brucellosis among patients attending a district hospital in Rwanda. Am J Trop Med Hyg 97: 831835.

    • Search Google Scholar
    • Export Citation
  • 8.

    Chatikobo P , Manzi M , Kagarama J , Rwemarika JD , Umunezero O , 2008. The prevalence of bovine brucellosis in milking dairy herds in Nyagatare and its implications on dairy productivity and public health. The 3rd International Conference on Appropriate Technology, Ministry of Agriculture and Animal Resources, Kigali, Republic of Rwanda Kigali, Republic of Rwanda: The Ministry in the Office of the President in Charge of Science and Technology, 247255. Available at: http://www.dcmetrosftp.org/appropriate%20tech/AT%20articles/Proceedings3 rdInternationalConference OnAppropriateTechnologyRwanda2008.pdf.

    • Search Google Scholar
    • Export Citation
  • 9.

    Manishimwe R , Ntaganda J , Habimana R , Nishimwe K , Byukusenge M , Dutuze F , de Dieu Ayabagabo J , Umurerwa L , Rukundo JC , 2015. Comparison between rose bengal plate test and competitive enzyme linked immunosorbent assay to detect bovine brucellosis in Kigali city. Rwanda J Vet Sci Technol 6: 1.

    • Search Google Scholar
    • Export Citation
  • 10.

    Ndazigaruye G , Mushonga B , Kandiwa E , Samkange A , Segwagwe BE , 2018. Prevalence and risk factors for brucellosis seropositivity in cattle in Nyagatare district, Eastern province, Rwanda. J South Afr Vet Assoc 89: a1625.

    • Search Google Scholar
    • Export Citation
  • 11.

    Stärk KDS , Regula G , Hernandez JA , Knopf L , Fuchs L , Morris RS , 2006. Concepts for risk-based surveillance in the field of veterinary medicine and veterinary public health. BMC Health Serv Res 6: 20.

    • Search Google Scholar
    • Export Citation
  • 12.

    Gall D , Nielsen K , 2004. Serological diagnosis of bovine brucellosis: a review of test performance and cost comparison. Revue Scientifique Tech 23: 9891002.

    • Search Google Scholar
    • Export Citation
  • 13.

    AusVet Animal Health Services , 2015. Calculate Confidence Limits for a Population Proportion. Available at: http://epitools.ausvet.com.au/content.php?page=CIProportion.

    • Search Google Scholar
    • Export Citation
  • 14.

    Hosmer DW , Lemeshow S , 2000. Applied Logistic Regression, 2nd edition. New York, NY: Wiley Series in Probability and Statistics, John Wiley and Sons.

    • Search Google Scholar
    • Export Citation
  • 15.

    Makita K , Fèvre EM , Waiswa C , Eisler MC , Thrusfield M , Welburn SC , 2011. Herd prevalence of bovine brucellosis and analysis of risk factors in cattle in urban and peri-urban areas of the Kampala economic zone, Uganda. BMC Public Health 7: 18.

    • Search Google Scholar
    • Export Citation
  • 16.

    Ebrahim WOMK , Elfadil AAM , Elgadal AA , Shuaib YA , 2016. Seroprevalence and risk factors of anti-brucella antibodies in cattle in Khartoum State, the Sudan. J Adv Vet Anim Res 3: 134144.

    • Search Google Scholar
    • Export Citation
  • 17.

    Rikin EU , 1988. Brucellosis of cattle in Nigeria. Proposals for a control under intensive and extensive husbandry systems. Acta Vet Scand 84: 9497.

    • Search Google Scholar
    • Export Citation
  • 18.

    Mai HM , Irons PC , Kabair J , Thompson PN , 2012. A large seroprevalence of brucellosis in cattle herds under diverse production systems in northern Nigeria. BMC Vet Res 8: 144.

    • Search Google Scholar
    • Export Citation
  • 19.

    Muma JB , Munyeme M , Samui KL , Skejerve E , Oloya BC , 2007. Risk factors for brucellosis in indigenous cattle reared in livestock–wildlife interface areas of Zambia. Prev Vet Med 80: 306317.

    • Search Google Scholar
    • Export Citation
  • 20.

    Gomo C , de Garine-Wichatitsky M , Caron A , Mubika-Pfukenyi D , 2012. Survey of brucellosis at the wildlife–livestock interface on the Zimbabwean side of the great limpopo transfrontier Conservation area. Trop Anim Health Prod 44: 7785.

    • Search Google Scholar
    • Export Citation
  • 21.

    Magona JW , Walubengo J , Galiwango T , Etoori A , 2009. Seroprevalence and potential risk of bovine brucellosis in zero-grazing and pastoral dairy systems in Uganda. Trop Anim Hlth Prod 41: 17651771.

    • Search Google Scholar
    • Export Citation
  • 22.

    Dohoo I , Martin W , Stryhn H , 2009. Veterinary Epidemiologic Research, 2nd edition. Prince Edward Island, Canada: VER, Inc.

Past two years Past Year Past 30 Days
Abstract Views 642 275 0
Full Text Views 263 127 31
PDF Downloads 150 57 8
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

Seroprevalence of and Risk Factors Associated with Exposure to Brucella Spp. in Dairy Cattle in Three Different Agroecological Zones in Rwanda

David Kiiza Heifer Project International, Kigali, Rwanda;
Rwanda Agriculture and Animal Resources Development Board, Kigali, Rwanda;

Search for other papers by David Kiiza in
Current site
Google Scholar
PubMed
Close
,
Savino Biryomumaisho Department of Veterinary Pharmacy, Clinical and Comparative Medicine, Makerere University, Kampala, Uganda;

Search for other papers by Savino Biryomumaisho in
Current site
Google Scholar
PubMed
Close
,
Ian D. Robertson School of Veterinary Medicine, Murdoch University, Perth, Australia;

Search for other papers by Ian D. Robertson in
Current site
Google Scholar
PubMed
Close
, and
Jorge A. Hernandez College of Veterinary Medicine, University of Florida, Gainesville, Florida;
University of Florida’s Institute of Food and Agricultural Sciences, Feed the Future Innovation Lab for Livestock Systems, Gainesville, Florida

Search for other papers by Jorge A. Hernandez in
Current site
Google Scholar
PubMed
Close

ABSTRACT

Livestock production is a key element for poverty alleviation, food security, and economic growth in Rwanda. In 2017, the national average milk production per cow was about 2.5 L per day; in 2020–2021, it is projected to increase to 3.5 L per day if improvement interventions including those designed to reduce the burden of brucellosis in cattle are implemented. The objective of the study reported here was to estimate the seroprevalence of and identify risk factors associated with dairy farms and cattle classified as seropositive to Brucella spp. in three different agroecological zones in Rwanda. Most study farms (40/85 or 47%) had one head of cattle only. Using the Rose Bengal test, the seroprevalence of brucellosis was 28/85 or 33% (95% CI = 24%, 43%) at the farm level and 63/465 or 14% (95% CI = 11%, 17%) at the animal level. Using logistic regression, at the farm level, the presence of seropositive cattle was associated with herd size (2–45 cattle, odds ratio = 21.2; 95% CI = 2.4, 184.5) (46–220 cattle, OR = 288.5; 95% CI = 24.3, 3,423.1) compared to farms with one animal, after controlling for main breed (local breeds, crossbreeds) on the farm. In addition, the odds of testing seropositive were 10.7 (95% CI = 2.3, 49.1) and 149.5 (95% CI = 19.3, 1,158.7) times higher in farms in Nyabihu district and Nyagatare district, respectively, than in farms in Muhanga district, after controlling for main breed on the farm. The odds of seropositivity to Brucella spp. were 2.8 times higher in farms with mostly local breeds, than in those with mostly crossbreeds; but the association was confounded by herd size and geographic location. At the animal level, the odds of seropositivity to Brucella spp. were 2.6 times higher in adult cattle than in young cattle (95% CI = 1.1, 6.3). Finally, we observed a high frequency of adult cattle (86%) and a high seroprevalence of brucellosis in adult cattle (25%) in Nyagatare; an indication that, in the absence of culling and other control measures, Brucella spp. infection pressure can be relatively constant and a steady source of disease transmission in pastoral systems in that district.

INTRODUCTION

Livestock production is a key element for poverty alleviation, food security, and economic growth in Rwanda. 1 The national cattle population is 1.39 million, and although local cattle breeds represent 43% of this population, they only contribute 9% of the total milk production. 1 In 2017, the national annualized average milk production per cow was about 909 L (2.5 L per day), and it is projected to increase by 41% to 1,281 litters (3.5 L per day) in 2021–2022 if improvement interventions in key areas, such as feed, genetics, animal health services, marketing and processing, and policy are implemented. 1 Brucellosis is a priority disease because of its public health and economic consequences. Brucellosis can be transmitted to humans by consumption of unpasteurized milk from or direct contact with infected animals, particularly aborted fetuses 2 ; infected people often suffer from a chronic, debilitating illness. In cattle, the disease can cause one or multiple abortions, reduced milk yield, and delayed calving. 35 Information on the burden of and risk factors associated with Brucella spp. infection in human and livestock populations in Rwanda is very limited. In one study conducted in women presenting with abortion/stillbirth at a hospital in Huye district, Southern Province in 2006, 15 of 60 (25%) patients tested positive to Brucella spp. antibodies by using the Rose Bengal test (RBT) 6 ; however, the temporal relationship between exposure to Brucella spp. and subsequent abortion in study subjects was not established. In another study, 12 of 198 (6%) patients with clinical signs of brucellosis (onset of fever and sweating, chills, arthralgia, weight loss, or joint pain) admitted to a hospital in Nyagatare, Eastern Province, in 2014 tested positive to the RBT 7 ; a positive result was associated with consumption of unboiled milk. Four studies have reported different prevalence estimates of exposure to Brucella spp. in selected cattle herds or populations in Rwanda. The seroprevalence of brucellosis in cattle, using the RBT, varied from 125/1,297 (10%) in 2002 to 108/2,340 (5%) in 2006 at one university farm in Huye district, Southern Province. 6 In addition, in that study, 50/603 (8%) cattle on a government research farm in the Southern Province in 2004 tested positive to the RBT. 6 In another study, 99/998 (10%) cattle in Nyagatare district, Eastern Province, in 2008 tested positive to the RBT. 8 In that study, 62/205 (30%) farms had one or more cattle that tested positive. 8 In a third study, 41 (2.0%) and 35 (1.7%) of 2017 cattle on 217 farms in suburban areas of Kigali city in 2015 tested positive to the RBT and C-ELISA, respectively 9 ; the seroprevalence of brucellosis at the farm level was not reported. Finally, in a recent study, 114/604 (19%) cattle on 120 farms in Nyagatare district tested positive to the RBT 10 ; the seroprevalence of brucellosis at the farm level was not reported.

In Rwanda’s Livestock Master Plan, 1 a key strategy to reduce the burden of brucellosis in the national cattle herd is the implementation of enhanced surveillance systems for early detection and risk management of infected cattle. Enhanced disease surveillance systems can include risk-based approaches for the efficient management of limited resources allocated to national veterinary services for disease control and prevention in cattle populations. 11 The objective of the study reported here was to estimate the seroprevalence of and identify risk factors associated with dairy farms and cattle classified as seropositive to Brucella spp. in three different agroecological zones in Rwanda.

MATERIALS AND METHODS

Study sites.

The study was conducted in three provinces of Rwanda (representing three agroecological zones and livestock systems) during September–December 2013: 1) Southern Province (plateau land, zero-grazing, and cut/carry system); 2) Western Province (highland and extensive system); and 3) Eastern Province (plateau land and pastoral system). Within each province, one district was randomly selected: Muhanga district in Southern Province (i.e., 1/8 districts was selected), Nyabihu district in Western Province (1/7 districts), and Nyagatare district in Eastern Province (1/7 districts). Finally, a total 920 households or farms with dairy cattle in Muhanga, 420 farms in Nyabihu, and 360 farms in Nyagatare were initially targeted for inclusion. A convenience sample of 85 farms (∼5%) from Muhanga (n = 46), Nyabihu (n = 21), and Nyagatare districts (n = 18). The 85 study farms were selected in coordination with district animal resources officers in the respective districts, and based on farmers’ interest and willingness to participate.

Study animals.

In Muhanga (where farmers keep less than 10 cattle in shelters and feed them in-doors), all dairy cattle present on selected farms were sampled and tested for the detection of antibodies to Brucella spp. In Nyabihu and Nyagatare, 20% of cattle present on the selected farms were sampled and tested for the detection of antibodies to Brucella spp. In Nyabihu and Nyagatare, sampled animals were those presented by the farmer on the day of sampling (i.e., selection of study animals was not conducted by using probabilistic random sampling procedures).

Collection of blood samples.

Approximately 10 mL of blood was collected from either the jugular or coccygeal vein of each selected animal into sterile plain vacutainer tubes without anticoagulant (Becton Dickson, United Kingdom). Each sample was labeled using codes to identify the animal, study household or farm, district, and sampling date. Blood samples were allowed to clot overnight at room temperature and transported in iceboxes to a designated laboratory at the National Veterinary Laboratory of Rwanda in Kigali city for processing. Serum was separated by centrifugation at 3,000 rpm for 15 minutes (503 × g) and stored in 2-mL cryovials at −20°C until tested.

Detection of Brucella spp. antibodies.

Serum samples were tested for the detection of antibodies to Brucella spp. by using the RBT and C-ELISA. Laboratory tests were conducted at the National Veterinary Laboratory of Rwanda in Kigali city. First, all serum samples were tested by using the RBT following procedures recommended by the World Organization for Animal Health. 9 Next, samples that tested positive by using the RBT were tested again by using the C-ELISA following procedures recommended by the manufacturer (SVANOVIR® Brucella-Ab C-ELISA, Svanova, Sweden). Samples that tested positive to both the RBT and C-ELISA were classified as seropositive (i.e., testing in series). The reported sensitivity and specificity of the RBT = 81.2% and 86.3% 12 and for the C-ELISA = 100.0% and 100.0% (SVANOVIR Brucella-Ab C-ELISA), respectively.

Data collection.

For each study household or farm, the following data were collected: household or farm identification, province, district, herd size, and main breed. The variable for main breed included three categories: 1) mainly local breeds (≥ 51% of herd size), 2) mainly crossbreed (≥ 51% of the herd), or 3) local breed (50%) and crossbreed (50%). In addition, for each study animal, the following data were collected: animal identification, household or farm identification, sex (male and female), breed (local and cross), and age (young = has not reached sexual maturity; adult = has reached sexual maturity). The local breed was Ankole long horn, and crossbreeds were Ankole long horn crossed with Brown Swiss, Sahiwal, Boran, Jersey, Friesian, or Guernsey.

Data analysis.

Descriptive statistics (median; minimum, and maximum) were calculated for the continuous variable of herd size. A household or farm was classified as positive if one or more cattle tested positive to antibodies to Brucella spp. The farm-level prevalence was calculated by dividing the number of seropositive households or farms by the total number of tested households or farms. The same approach was used to estimate the prevalence of individual animals classified as positive to antibody titers to Brucella spp. Ninety-five percent CIs (95% CI) were calculated for each prevalence estimate by using the software EpiTools. 13 At the household or farm level, the distributions for herd size were compared between districts and between the main breeds by using the nonparametric Kruskal–Wallis test. In addition, the frequency (%) of households or farms with mostly local breeds was compared between districts by using a chi-square test. At the animal level, the frequency (%) of adult cattle was compared between districts by using a chi-square test. In addition, among adult cattle, the frequency (%) of seropositive cattle was compared between districts by using a chi-square test.

Logistic regression was used to identify exposure factors associated with a positive diagnosis (yes, no) of antibodies to Brucella spp. at the household or farm level and at the animal level. 14 At the farm level, investigated exposure factors were main breed (cross, local, and cross and local), herd size (1, 2–45, and 46–220), and district (Muhanga, Nyabihu, and Nyagatare). At the animal level, exposure factors were age (young and adult), sex (male and female), and breed (cross and local). Initial screening of investigated exposure factors associated with a positive diagnosis of antibodies to Brucella spp. was performed by use of univariable logistic regression. Variables with a P-value ≤ 0.20 were considered for multivariable analyses. Associations between independent variables (exposure factors) were examined, and when a pair of variables was significantly associated (P < 0.05) by use of a chi-square test, the exposure variable judged as most biologically plausible was further examined in the analysis. In the multivariable analyses, a manual forward selection technique was used by adding one variable at the time and assessing the model goodness of fit using the likelihood ratio test statistic. Exposure factors retained in final models were examined for confounding by adding each of the variables to the model, and assessing the changes in the odds ratios (i.e., ≥ 20%) of the remaining variables in the model. Interaction effects at the farm level (e.g., herd size × district) were not examined because the number of seropositive farms was small in several categories. Logistic regression models were assessed for goodness of fit using the Hosmer–Lemeshow test. Odds ratios and 95% CI were calculated to measure the magnitude of association between investigated exposure factors and seropositivity to Brucella spp. In all analyses, P-values < 0.05 were considered significant.

RESULTS

Herd size was smaller in Muhanga district (mean = 1, median = 1, minimum = 1, and maximum = 7) than that in Nyabihu (44; 35; 15, 100) or Nyagatare (60; 55; 15, 220) (P < 0.05). Herd size was smaller in households or farms with mostly (≥ 51%) crossbreeds (mean = 22; median = 1; minimum = 1, and maximum = 220) than in those with local breeds (30; 28; 1, 75) (P < 0.05). The frequency of households or farms with mostly local breeds was different between districts: Muhanga (5/46 or 11%), Nyabihu (6/21 or 29%), and Nyagatare (9/18 or 50%) (P < 0.05).

Seroprevalence of brucellosis at the household or farm level.

Overall, the farm-level seroprevalence of brucellosis was 28/85 or 33% (95% CI = 23.8, 43.4). The farm-level seroprevalence was different (P < 0.01) between districts: Muhanga (3/46 or 7%), Nyabihu (9/21 or 43%), and Nyagatare (16/18 or 89%).

Using univariable logistic regression analysis, the variables for main breed, herd size, and district had P-values ≤ 0.20. The variables for herd size and district were associated (P < 0.05). The odds of testing positive to Brucella spp. antibodies were 2.87 times higher in households or farms that mostly had local breeds than in those that mostly had crossbreed cattle (crude OR = 2.87; 95% CI = 1.01, 8.17) (Table 1). Using multivariable logistic regression analysis, the odds of testing positive to Brucella spp. antibodies were associated with herd size or district (Nyabihu and Nyagatare) (Table 2). In one model, adding the variable for herd size changed the OR for the observed association between main breed (local) and a positive diagnosis of Brucella spp. antibodies from 2.87 to 2.18 (32% change): an indication that the effect of local breed on a positive diagnosis of Brucella spp. antibodies was confounded by the variable for herd size (i.e., the observed association between local breed and a positive diagnosis of Brucella spp. antibodies was overestimated by 32% before herd size was taken into account). The Hosmer–Lemeshow goodness of fit tests (2.6; degrees of freedome [df] = 3; P = 0.53) indicated that the model provided a good fit of the data. In another model, adding the variable for district changed the OR for the observed association between main breed (local) and a positive diagnosis of Brucella spp. antibodies from 2.87 to 0.68 (422% change): an indication that the effect of local breed on a positive diagnosis of Brucella spp. antibodies was confounded by the variable for district. The Hosmer–Lemeshow goodness of fit tests (3.6; df = 2; P = 0.16) indicated that the model provided a good fit of the data.

Table 1

Univariable analysis for identification of risk factors associated with a positive diagnosis of Brucella spp. antibodies in 85 households or farms with dairy cattle in three districts of Rwanda

Variable Category Negative Positive OR* 95% CI P-value
Main breed Cross 46 16 1.00 Reference NA
Local 10 10 2.87 1.01, 8.17 0.04
Cross and local 1 2 5.75 0.49, 67.70 0.16
Herd size 1 39 1 1.00 Reference NA
2–45 16 11 26.81 3.21, 223.88 < 0.01
46–220 2 16 311.99 26.53, 3,669.59 < 0.01
District Muhanga 43 3 1.00 Reference NA
Nyabihu 12 9 10.75 2.51, 45.97 < 0.01
Nyagatare 2 16 114.67 17.54, 749.60 < 0.01

Odds ratio.

95% CI.

In farms with two or more dairy cattle, the main breed (≥ 51%) was cross or local breeds.

Table 2

Multivariable analysis for identification of risk factors associated with a positive diagnosis of Brucella spp. antibodies in 85 households or farms with dairy cattle in three districts of Rwanda

Variable Category Adjusted odds ratio 95% CI P
Model 1*
Main breed Cross
Local 2.18 0.52, 9.23 0.28
Cross and local 2.37 0.14, 38.83 0.54
Herd size 1
2–45 21.27 2.45, 184.50 < 0.01
46–220 288.51 24.32, 3,423.11 < 0.01
Model 2
Main breed Cross
Local 0.68 0.14, 3.29 0.63
Cross and local 7.33 0.40, 134.68 0.17
District Muhanga
Nyabihu 10.74 2.35, 49.10 < 0.01
Nyagatare 149.55 19.30, 1,158.71 < 0.01

Hosmer–Lemeshow statistic = 2.6; df = 3; P = 0.53.

Hosmer–Lemeshow statistic = 3.6; df = 2; P = 0.16.

Seroprevalence of brucellosis at the animal level.

The animal-level seroprevalence of brucellosis was 63/465 or 14% (95% CI = 11%, 17%). The animal-level seroprevalence was higher (P < 0.01) in Nyagatare (48/217 or 22%) than in Muhanga (3/61 or 5%) or Nyabihu (12/187 or 6%). The frequency of adult cattle was different between districts: Muhanga (38/61 or 62%) and Nyabihu (147/187 or 79%) and Nyagatare (186/217 or 86%) (P < 0.05). In addition, among adult cattle, the seroprevalence of brucellosis was different between districts: Muhanga (0/38 or 0%; 95% CI = 0%, 9%) or Nyabihu (11/147 or 7%; 95% CI = 4%, 13%), and Nyagatare (46/186 or 25%; 95% CI = 19%, 31%). Using univariable logistic regression analysis, the odds of testing positive to Brucella spp. antibodies were 2.66 times higher in adult cattle than in young cattle (OR = 2.66; 95% CI = 1.11, 6.37; P = 0.02) (Table 3). A multivariable logistic regression analysis was not conducted because the variables for age and breed were correlated (P < 0.05).

Table 3

Univariable analysis for identification of risk factors associated with a positive diagnosis of Brucella spp. antibodies in 465 dairy cattle in three districts of Rwanda

Variable Category Negative Positive OR 95% CI P-value
Age Young 88 6 1.00 Reference NA
Adult 314 57 2.66 1.11, 6.37 0.02
Sex Male 43 4 1.00 Reference NA
Female 359 59 1.77 0.63, 5.01 0.28
Breed Cross 269 18 1.00 Reference NA
Local 133 45 5.06 2.82, 9.07 < 0.01

DISCUSSION

This study provides new information on the seroprevalence of brucellosis in dairy cattle in three different ecological zones and livestock systems in Rwanda. The seroprevalence of brucellosis in dairy cattle was higher in Nyagatare district (associated with pastoral systems) and lowest in Muhanga district (associated with zero-grazing, and cut and carry systems). Cattle in households or farms with mostly local breeds tested positively more often to Brucella spp. antibodies, but this association was confounded by herd size and geographic location. In 2013, the cattle population in Rwanda was not vaccinated against Brucella spp. The detection of Brucella spp. antibodies using the RBT represents natural exposure.

The seroprevalence of brucellosis at the household or farm level and at the animal level was higher in Nyagatare district than that in Nyabihu and Muhanga. Three broad factors that can explain this finding are herd size, geographic location, and livestock management practices. First, herd size was higher in study households or farms in Nyagatare (mean = 60; median = 55) than in those in Nyabihu (44; 35) and Muhanga (1; 1). Large pastoral herds that can be at high risk are of Brucella spp. infection through direct and/or indirect contact with other pastoral herds infected with Brucella spp. 15 Other pathways of exposure to Brucella spp. in cattle in Nyagatare could include bull exchange for mating 10,16 and introduction of Brucella spp.–infected cattle onto the premises. Second, Nyagatare is located bordering Uganda and Tanzania, where there is a constant movement of livestock across the border with those two countries. Pastoral and semi-pastoral livestock systems in Nyagatare allow different herds to mix at watering points, a known risk factor for Brucella spp. infection in cattle herds. 10,17,18 In addition, Nyagatare contains Akagera National Park, located along the border with Tanzania. Although the presence or absence of Brucella spp. infection has not been determined, the national park has wild animals that can act as reservoirs for Brucella spp. if they are infected, and cattle often graze alongside wild animals. Close contact between cattle and susceptible wild animals (e.g., wildebeest, eland, and African buffalo—Syncerus caffer) is a known risk factor for exposure to Brucella spp. in cattle. 3,19,20 Finally, the observed lowest prevalence of exposure to Brucella spp. in Muhanga can be attributed to herd size and livestock management. In Muhanga, most study households had one cow or one bull and practiced zero-grazing, where farmers cut and carry forage or crop residues to feed livestock kept below a shed. A low frequency of exposure to Brucella spp. in zero-grazing systems with very low level of cattle-to-cattle contacts, even in the absence of specific control measures, has been reported in East African highlands. 3,21 In Nigeria, a high prevalence of exposure to Brucella spp. was observed in cattle in zero-grazing systems (24/101 or 24%) and pastoral systems (561/1,244 or 45%). 18 In Nigeria, however, cattle in zero-grazing systems are generally purchased in livestock markets for a fattening program. By contrast, pastoral systems in Nigeria are characterized by long distance movement of cattle in search of pasture and water, comingling in communal grazing areas and at watering points, particularly during the dry season, 18 leading to direct and/or indirect contact and potential disease transmission.

At the animal level, the odds of testing seropositive to Brucella spp. were 2.6 times higher in adult cattle than in young cattle. The higher seroprevalence observed in adult cattle can be attributed to a longer duration of potential exposure to Brucella spp. 18 In addition, the high frequency of adult cattle and the high seroprevalence of brucellosis in adult cattle in Nyagatare indicate that Brucella spp. infection pressure is relatively constant. 3 Because culling and other methods of brucellosis control are not practiced in most pastoral systems, long-term chronic Brucella spp. infections can cause multiple abortions and a steady source of Brucella spp. exposure and disease transmission. 3

This study had several imitations. The study sample of 85 households or farms with dairy cattle in Muhanga, Nyabihu, and Nyagatare districts was not justified by using sample size calculations for disease prevalence or selected by using a probabilistic random sampling approach. Thus, it is possible that prevalence estimates of Brucella spp. exposure at the household or farm level reported here were under- or overestimated. In addition, although the sample (20%) of cattle tested on each farm in Nyabihu and Nyagatare did not target sick or clinically healthy cattle, it is possible one or more farms were misclassified as negative if the farm had Brucella spp.–infected cattle that were not sampled. Study results apply to the sample of households or farms with dairy cattle in Muhanga, Nyabihu, and Nyagatare. In future epidemiologic studies to estimate disease prevalence, probabilistic random sampling approaches should be adopted to mitigate against selection bias. 22 Finally, the RBT used for classification of cattle as positive or negative to Brucella spp. antibodies is not 100% accurate. In this study, testing in series (C-ELISA) was used to reduce false-positive results. Limited funding and local diagnostic capacity prevented our ability to collect and test milk samples for diagnosis and identification of Brucella spp. by using bacteriologic culture, PCR, or nucleotide sequencing techniques.

CONCLUSION

Overall, the seroprevalence of brucellosis was 28/85 or 33% (95% CI = 24%, 43%) at the household or farm level and 63/465 or 14% (95% CI = 11%, 17%) at the animal level. At the household or farm level, the seroprevalence of brucellosis was higher in Nyagatare district than that in Nyabihu and Muhanga. Herd size and geographic location were associated with a positive diagnosis of Brucella spp. antibodies. Breed (local) was associated with seropositivity to Brucella spp., but the association was confounded by herd size and geographic location. The high frequency of adult cattle and the high seroprevalence of brucellosis in adult cattle in Nyagatare district indicate that Brucella spp. infection pressure can be relatively constant. Because culling and other methods of brucellosis control are not practiced in most pastoral systems, long-term chronic Brucella spp. infections can cause multiple abortions and a steady source of Brucella spp. exposure and disease transmission. Future studies can confirm that the risk of Brucella spp. exposure is higher in pastoral systems than in zero-grazing systems in Rwanda, and justify the use of risk-based surveillance systems for early detection and risk management of brucellosis in livestock populations in Rwanda.

Figure 1.
Figure 1.

Map of Rwanda showing three selected districts (black).

Citation: The American Journal of Tropical Medicine and Hygiene 104, 4; 10.4269/ajtmh.20-1426

REFERENCES

  • 1.

    Shapiro BI , Gebru G , Desta S , Nigussie K , 2017. Rwanda Livestock Master Plan. Available at: http://extwprlegs1.fao.org/docs/pdf/rwa172923.pdf. Accessed June 30, 2020.

    • Search Google Scholar
    • Export Citation
  • 2.

    Dean AS , Crump L , Greter H , Hattendorf J , Schelling E , Zinsstag J , 2013. Clinical manifestations of human brucellosis: a systematic review and meta-analysis. PLoS Negl Trop Dis 612: e1929.

    • Search Google Scholar
    • Export Citation
  • 3.

    McDermott JJ , Arimi SM , 2002. Brucellosis in sub-Saharan Africa: epidemiology, control and impact. Vet Microbiol 90: 111134.

  • 4.

    McDermott JJ , Deng KA , Jayatileka TN , EL Jack MA , 1987. Cross-sectional cattle disease study in Kongor rural council, Southern Sudan. II. Brucellosis in cows: associated factors, impact on production and disease control considerations. Prev Vet Med 5: 125132.

    • Search Google Scholar
    • Export Citation
  • 5.

    Megersa B , Biffa D , Abunna F , Regassa A , Godfroid J , Skjerve E , 2011. Seroprevalence of brucellosis and its contribution to abortion in cattle, camel, and goat kept under pastoral management in Borana, Ethiopia. Trop Anim Health Prod 43: 651656.

    • Search Google Scholar
    • Export Citation
  • 6.

    Rujeni N , Mbanzamihigo L , 2014. Prevalence of brucellosis among women presenting with abortion/stillbirth in Huye, Rwanda. J Trop Med 2014: 13.

    • Search Google Scholar
    • Export Citation
  • 7.

    Gafirita J et al. 2017. Seroprevalence of brucellosis among patients attending a district hospital in Rwanda. Am J Trop Med Hyg 97: 831835.

    • Search Google Scholar
    • Export Citation
  • 8.

    Chatikobo P , Manzi M , Kagarama J , Rwemarika JD , Umunezero O , 2008. The prevalence of bovine brucellosis in milking dairy herds in Nyagatare and its implications on dairy productivity and public health. The 3rd International Conference on Appropriate Technology, Ministry of Agriculture and Animal Resources, Kigali, Republic of Rwanda Kigali, Republic of Rwanda: The Ministry in the Office of the President in Charge of Science and Technology, 247255. Available at: http://www.dcmetrosftp.org/appropriate%20tech/AT%20articles/Proceedings3 rdInternationalConference OnAppropriateTechnologyRwanda2008.pdf.

    • Search Google Scholar
    • Export Citation
  • 9.

    Manishimwe R , Ntaganda J , Habimana R , Nishimwe K , Byukusenge M , Dutuze F , de Dieu Ayabagabo J , Umurerwa L , Rukundo JC , 2015. Comparison between rose bengal plate test and competitive enzyme linked immunosorbent assay to detect bovine brucellosis in Kigali city. Rwanda J Vet Sci Technol 6: 1.

    • Search Google Scholar
    • Export Citation
  • 10.

    Ndazigaruye G , Mushonga B , Kandiwa E , Samkange A , Segwagwe BE , 2018. Prevalence and risk factors for brucellosis seropositivity in cattle in Nyagatare district, Eastern province, Rwanda. J South Afr Vet Assoc 89: a1625.

    • Search Google Scholar
    • Export Citation
  • 11.

    Stärk KDS , Regula G , Hernandez JA , Knopf L , Fuchs L , Morris RS , 2006. Concepts for risk-based surveillance in the field of veterinary medicine and veterinary public health. BMC Health Serv Res 6: 20.

    • Search Google Scholar
    • Export Citation
  • 12.

    Gall D , Nielsen K , 2004. Serological diagnosis of bovine brucellosis: a review of test performance and cost comparison. Revue Scientifique Tech 23: 9891002.

    • Search Google Scholar
    • Export Citation
  • 13.

    AusVet Animal Health Services , 2015. Calculate Confidence Limits for a Population Proportion. Available at: http://epitools.ausvet.com.au/content.php?page=CIProportion.

    • Search Google Scholar
    • Export Citation
  • 14.

    Hosmer DW , Lemeshow S , 2000. Applied Logistic Regression, 2nd edition. New York, NY: Wiley Series in Probability and Statistics, John Wiley and Sons.

    • Search Google Scholar
    • Export Citation
  • 15.

    Makita K , Fèvre EM , Waiswa C , Eisler MC , Thrusfield M , Welburn SC , 2011. Herd prevalence of bovine brucellosis and analysis of risk factors in cattle in urban and peri-urban areas of the Kampala economic zone, Uganda. BMC Public Health 7: 18.

    • Search Google Scholar
    • Export Citation
  • 16.

    Ebrahim WOMK , Elfadil AAM , Elgadal AA , Shuaib YA , 2016. Seroprevalence and risk factors of anti-brucella antibodies in cattle in Khartoum State, the Sudan. J Adv Vet Anim Res 3: 134144.

    • Search Google Scholar
    • Export Citation
  • 17.

    Rikin EU , 1988. Brucellosis of cattle in Nigeria. Proposals for a control under intensive and extensive husbandry systems. Acta Vet Scand 84: 9497.

    • Search Google Scholar
    • Export Citation
  • 18.

    Mai HM , Irons PC , Kabair J , Thompson PN , 2012. A large seroprevalence of brucellosis in cattle herds under diverse production systems in northern Nigeria. BMC Vet Res 8: 144.

    • Search Google Scholar
    • Export Citation
  • 19.

    Muma JB , Munyeme M , Samui KL , Skejerve E , Oloya BC , 2007. Risk factors for brucellosis in indigenous cattle reared in livestock–wildlife interface areas of Zambia. Prev Vet Med 80: 306317.

    • Search Google Scholar
    • Export Citation
  • 20.

    Gomo C , de Garine-Wichatitsky M , Caron A , Mubika-Pfukenyi D , 2012. Survey of brucellosis at the wildlife–livestock interface on the Zimbabwean side of the great limpopo transfrontier Conservation area. Trop Anim Health Prod 44: 7785.

    • Search Google Scholar
    • Export Citation
  • 21.

    Magona JW , Walubengo J , Galiwango T , Etoori A , 2009. Seroprevalence and potential risk of bovine brucellosis in zero-grazing and pastoral dairy systems in Uganda. Trop Anim Hlth Prod 41: 17651771.

    • Search Google Scholar
    • Export Citation
  • 22.

    Dohoo I , Martin W , Stryhn H , 2009. Veterinary Epidemiologic Research, 2nd edition. Prince Edward Island, Canada: VER, Inc.

Author Notes

Address correspondence to David Kiiza, Capacity Development Consultants (CDC) LTD., P.O. Box 7005, Kigali, Rwanda. E-mail: dkiiza07@yahoo.com

Disclosure: Supporting information Excel file with data is available on request.

Financial support: This study was funded in part by the Rwanda Agriculture Board through the Directorate of Animal Resources and AusAID.

Authors’ addresses: David Kiiza, Rwanda Agriculture and Animal Resource Development Board (RAB), Animal Resources, Kigali, Rwanda, and Capacity Development Consultants Limited, Livestock Value Chain, Kigali, Rwanda, E-mail: dkiiza07@yahoo.com. Savino Biryomumaisho, Veterinary Pharmacy, Makerere University, Kampala, Uganda, E-mail: sbiryomumaish15@gmail.com. Ian D. Robertson, School of Veterinary Medicine, Murdoch University, Perth, Australia, E-mail: i.robertson@murdoch.edu. Jorge A. Hernandez, College of Veterinary Medicine, University of Florida, Gainesville, FL, E-mail: hernandezja@ufl.edu.

Save