Epidemiology of brucellosis in humans and domestic ruminants in Bangladesh | Epidémiologie de la brucellose bovine chez les ruminants domestiques et chez l’homme au Bangladesh

Background Brucellosis is an ancient and one of the world’s most widespread zoonotic diseases affecting both, public health and animal production. It is endemic in many developing countries of Asia, Africa and Latin America including Bangladesh. Since the first report in 1970, a lot of brucellosis seroprevalence reports are available in cattle, goats, sheep and humans in Bangladesh. Most of the previously reported prevalence studies were based on non-random samples, which may not give a true representation of the status of the disease in respective populations. Some authors also investigated the risk factors in cattle. The tests used for the diagnosis of brucellosis in domestic ruminants and humans are imperfect and their performance was not evaluated in Bangladesh. The true prevalence of brucellosis in domestic ruminants is not known and is essential for analyzing the impact of this disease in domestic ruminants in Bangladesh. Indeed, when diagnostic tests are used without evaluating their performance in a context usually generate unreliable results, which in turn may lead to wrong epidemiological inferences. In addition, information on risk factors of brucellosis in humans and animals is also scarce. Moreover, the different species of Brucella prevalent in animals is scarce and not known in humans in Bangladesh. The overall objective of this thesis was to investigate the epidemiology of brucellosis in humans and domestic ruminants in Bangladesh in terms of the evaluation of commonly used diagnostic tests, estimation of true prevalence, identification of risk factors and detection of Brucella species in order to provide information that will guide the selection of appropriate control strategies. Study design and data analysis Sampling To collect random samples of animals a system of map digitization and selection of one geographical point from selected unions (Sub-Upazilla) using a hand held GPS machine was used. Blood (milk also where applicable) samples were then collected from livestock farmers and their animals within 0.5 km of the selected points. A convenient blood sample of butchers, dairy hands and veterinary practitioners were collected from Dhaka and Mymensingh districts. The sera of pyretic humans were collected from Mymensingh Medical College hospital randomly once in a week. Random milk samples were collected from Sirajgonj and Chittagong districts. Systematic random milk and blood samples of cattle including breeding bulls (semen also) of central cattle breeding and dairy farm (CCBDF) were also collected. Milk and blood samples of gayals of a herd in regional Bangladesh Livestock Research Institute at Naikhonchari, Bandarban were also collected. Convenient samples of placenta and vaginal swabs were also collected from Mymensingh district. Data collection and Analysis Data on serology was generated by using Rose Bengal test (RBT), Slow Agglutination test (SAT) /Standard tube agglutination test (STAT) (animals/humans) and indirect enzyme linked immunosorbent assay (iELISA). Animal, their herd level data and human data on potential risk factors were collected using a pretested questionnaire. The data was stored in Microsoft Excels worksheets and transferred to respective software for analysis. To estimate true prevalence and evaluate three conditionally dependent serological tests, Bayesian latent class models were used. Random effect and Firth’s logistic regression analyses were used to determine the risk factors of human brucellosis. The STATA, R and OpenBUGS softwares were used for data analyses. Staining, culture, genus and species-specific real time PCR assays were applied to isolate and to detect Brucella Spp./DNA in seropositive human sera and animal samples. Main results: Only 0.29% (95% CI: 0.06-0.86) cattle were acutely infected whereas 0.49% (95% CI: 0.16- 8 1.1) were chronically infected with brucellosis in Mymensingh. On the other hand, in CCBDF 15.58% (95% CI:11.89-19.89) cattle were acutely infected with brucellosis and only 3.2%(95%CI: 1.63-5.72) were chronically infected. The true prevalence of brucellosis among cattle in Mymensingh and CCBDF were 0.3% (95%CI: 0.03-0.7) and 20.5% (95% CI: 16.4-26.3) respectively. The performance of iELISA was best in both Mymensingh and CCBDF with the sensitivity of 90.5% and 91.3% and specificity of 99.3% and 99.2% respectively. The performance of RBT was better in Mymensingh than CCBDF with 81.0% and 76.1% sensitivity and 99.0% and 95.6% specificity respectively. Similar to RBT, the performance of SAT was also better in Mymensingh than CCBDF with 63.5% and 79.7% sensitivity and 98.6% and 95.3% specificity respectively. Through this test validation study, a new cut-off of 5 IU/ml for iELISA was recommended both in low (as at Mymensingh) and high prevalence scenarios in cattle populations (as at CCBDF) for routine screening. It was recommended to do nothing for the control of bovine brucellosis under small-scale dairy and subsistence management systems in Bangladesh. However, vaccination should be applied in herds where the prevalence is very high as like CCBDF. The true prevalence of brucellosis in goats and sheep were estimated as 1% (95% CI): 0.7–1.8) and 1.2% (95% CI: 0.6–2.2) respectively. The sensitivity of iELISA was 92.9% in goats and 92.0% in sheep with corresponding specificities of 96.5% and 99.5% respectively. The sensitivity and specificity estimates of RBT were 80.2% and 99.6% in goats and 82.8% and 98.3% in sheep. The sensitivity and specificity of SAT were 57.1% and 99.3% in goats and 72.0% and 98.6% in sheep. The prevalence of brucellosis in occupationally exposed people (HROG) using three tests was observed to be 4.4% based on a parallel interpretation. The results of the multiple random effects logistic regression analysis with random intercept for district revealed that the odds of brucellosis seropositivity among individuals who had been in contact with livestock for more than 26 years was about 14 times higher as compared to those who had less than 5 years of contact with livestock. In addition, when the contact was with goats, the odds of brucellosis seropositivity were about 60 times higher as compared to when contact was with cattle only. The seroprevalence of brucellosis among patients with pyrexia of unknown origin (PUO) was estimated to be 2.7% (95% CI: 1.2-5.2). The age, residence, type of patient, contact with animals, type of animal handled, arthralgia and backache were found to be significantly associated with a positive serological result in bivariable Firth’s logistic regression. Brucella abortus was detected from seropositive pyretic patients. Conclusion: The true exposure prevalence of brucellosis in cattle under small-scale dairy and subsistence/backyard management systems is very low (0.3%; 95% CI: 0.03-0.7). The active/acute infection is also very low (0.29%: 95% CI: 0.06-0.86) and similar to true exposure prevalence. The brucellosis in cattle under such management system is naturally controlled and further control program is not recommended considering the poor socioeconomic conditions. The true exposure prevalence of brucellosis in CCBDF is very high (20.5%; 95% CI: 16.4-26.3). The acute infection in this farm is also very high (15.58%; 95% CI: 11.89-19.89). Immediate control measures by initiating calf hood (female calf) vaccination are recommended to protect a valuable herd which also provides frozen semen for artificial insemination all over the country. The SAT and iELISA may simultaneously be applied to know the stage of brucellosis infection in domestic ruminants in high prevalence scenarios. The true exposure prevalence of brucellosis in goats and sheep are also low and around 1%. Due to lower positive predictive value, these test results should be interpreted with caution to avoid misleading information. Breeding bulls used for artificial insemination all over the country were found to be infected with brucellosis. Brucellosis is not a serious problem for the general population in Bangladesh as drinking raw milk and milk products is unusual and not a risk factor. The apparent prevalence of brucellosis in high risk occupationally exposed people (4.4%; 95% CI: 2.8-6.6) and in pyretic patients (2.7%; 95% CI: 1.2-5.2) are also low. The RBT may be applied as a screening test in humans having signs and symptoms of brucellosis along with the history of animal contact. In case of suspicion, genus or species specific rt PCR may be applied for confirmation. Only B. abortus DNA was amplified from 19 seropositive human samples (both HROG and PUO) and six animal samples (3 cows milk, one goat milk, one gayal milk and one bull semen). No Brucella like organism was observed under microscope in stained smears. Similarly, no Brucella organism was isolated from any of the clinical samples.