Seroprevalence survey of brucellosis among rural people in Mongolia

WPSAR Vol 5, No 3, 2014 | doi: 10.5365/wpsar.2014.5.1.002 www.wpro.who.int/wpsar 1 a National Centre for Communicable Diseases of Mongolia. b Animal Health Project of Swiss Development Agency in Mongolia. c National University of Medicine, Mongolia. d Swiss Tropical and Public Health Institute, Basel, Switzerland. Submitted: 2 January 2014; Published: 11 November 2014 doi: 10.5365/wpsar.2014.5.1.002 Background: After the transition from socialism to a market economy in 1990, human brucellosis re-emerged in Mongolia. The aim of our study was to estimate a representative seroprevalence of Brucella spp. and to determine risk factors for brucellosis seropositivity among rural people.

Mongolia has the second highest incidence of human brucellosis worldwide; another seven republics of the former Soviet Union are included in the 25 countries with the highest incidence.According to data from the National Statistical Office of Mongolia, a rapid increase in notified cases of brucellosis was observed between 1990 and 2000.The increase may have been the result of the evolution from a socialist state to a free market economy which led to the loss of rigorous livestock control. 4During this period, changes to the health system precluded early recognition of the disease or interventions that considered the emerging trends in humans and animals. 5In Mongolia, factors contributing to the incidence of brucellosis include traditional eating habits, standard hygiene measures, methods for processing milk and its products and rapid movement of animals. 3 2011, a national brucellosis serosurvey was conducted that sampled 168 027 head of livestock from 11 528 nomadic camps (two to more than four herder families that share the same pasture and water source) of 337 districts of 21 provinces. 6Twenty-one provinces, 57.3% of all districts and 8.0% of all nomadic camps cluster was optimized according to the feasibility and the available budget.
The study was approved by the Ethics Committee of the Health Sciences University of Mongolia and the Ethics Committee of the Canton of Basel of Switzerland.All participants were informed about the study and what they could expect regarding diagnosis, reporting and treatment; all signed a consent form.A child younger than 16 years of age was included in the study with signed consent from of his/her parents.

Study questionnaire
All study participants were interviewed using a questionnaire which included demographics, risk factors and clinical symptoms for brucellosis.The questionnaire was pre-tested during the 2010 study in Sukhbaatar and Zavkhan 7 and revised for the extended study to improve understanding of questions and to eliminate overlysensitive questions.

Blood sample collection and handling
Venous blood was taken with 5 ml Vacutainer® tubes.The blood samples were centrifuged in 3000 rounds per minute for five minutes.Separated 1.5 ml tubes of serum were kept in a cool box and transported to the provincial laboratories for storage and cooling before shipment had seropositive livestock including camels, cattle, sheep and goats.Livestock seroprevalence was found in 0.7% of camels, 1.8% of cattle, 0.7% of sheep and 0.5% of goats using parallel interpretations of Rose Bengal Tests (RBT), complement fixation tests and competitiveenzyme-linked immunoabsorbent assay (ELISA). 6e aim of our study was to estimate the seroprevalence of Brucella spp.and to determine risk factors for brucellosis seropositivity among rural people.

Study design and population
Eight provinces were selected for the cross-sectional surveys.Between June and September 2010, surveys were conducted in Sukhbaatar and Zavkhan provinces, selected for convenience. 7Between November 2011 and January 2012, the same surveys were conducted in a further six provinces: Arkhangai, Khuvsgul, Selenge, Uvs, Umnugovi and Govi-Altai (Figure 1).In each province, four districts were selected using simple randomization in Excel (the rand () command).Twenty nomadic camps and four to five individual participants were randomly selected based on the required sample size.
The cluster sample size calculation as described elsewhere 7 assumed a human brucellosis seroprevalence among Mongolian rural people of 20%. 8In addition, the number of clusters and number of individuals per

Selected provinces Provinces not selected
Disclaimer: The boundaries shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.

Figure 1. Map of Mongolia by province highlighting provinces where the study was conducted
Brucellosis among rural people in Mongolia Tsend et al ratio test was then constructed.Variables with p values less than 0.05 in the univariate analysis were included in the multivariate model.
To determine the proportion of the general population seroconverting each year due to brucellosis exposure, the seroprevalence data were divided by the duration of seropositivity, assumed to be 10.9 years. 10Using a conservative estimate of 20% of seroconversions representing true clinical cases (note that among all seropositives detected, 58.5% had at least two symptoms and 31.5% had at least three symptoms at time of interview), these proportions were multiplied by 0.3 and converted to rates per 100 000 for the general population.

Seroprevalence
The seroprevalence of Brucella spp.among participants was 11.1% (95% CI: 10.0-12.1)ranging from 2.3% to 22.6% in the eight provinces (Table 1) and 4.1% to 43.8% in the 28 districts.Within nomadic camps, 39.2% (95% CI: 38.2-41.0)had at least one to the serological laboratory of the National Center for Communicable Diseases in Ulaanbaatar where they were tested for brucellosis.

Serological test
Sera were tested with the RBT for detection of antibodies to Brucella abortus/melitensis from Tulip Diagnostic Ltd (Bambolim, India).Positive sera were re-tested with the RBT using ½ to 1/32 serum dilutions, 9 and with enzyme immunoassay for the qualitative determination of IgG class antibodies against Brucella from the NovaTec Immundiagnostica GmbH (Dietzenbach -63128 Germany).The ELISA test was performed according to manufacturer's instruction.

Data entry and statistical analysis
All data were double-entered in Access 2007, compared in Epi Info™ 3.5 to correct entry errors and analysed using STATA 10.1.Study participants who tested positive by either ELISA or RBT were considered seropositive for the statistical analysis.
To assess the association between risk factors and human brucellosis seropositivity we used Pearson  2 or Fisher's exact tests for explanatory variables such as demographics, behaviour-related risk factors and reported clinical symptoms.We also conducted univariate logistic regression using the binary serological outcome with the xtgee command and random effect on the nomadic camp level.A multivariate logistic regression model (with random effect at the nomadic camp) using backward stepwise selection and a removal level for covariates at P = 0.10 based on the likelihood-

History of human brucellosis and clinical symptoms
Of the study participants, 2.7% (n = 76) reported receiving treatment for human brucellosis in the past; the median time since past brucellosis treatment was 14 years (Q1 = 3.3 and Q3 = 20 years).With the exception of testicular pain, there were significant differences between age groups in reporting clinical symptoms; the age groups of 20 to 44 years and 45 years and above reported more clinical symptoms for human brucellosis.Females also reported more headaches; joint, back and muscle pain; weakness and sleeping disturbances than males (Table 4).
Reported clinical symptoms at the time of the study were compared to the sero-status of participants.Overall, 165 of the 316 (52.2%) brucellosis seropositive participants and 1186 of the 2540 (46.7%) seronegative participants reported symptoms.Among all seropositives, 36.7% reported more than three symptoms; among the seronegatives, 23.1% reported more than three symptoms (P < 0.001).Headache; joint, back and muscle pain; night sweats and sleeping disturbances were significantly associated with brucellosis seropositivity (Table 5).

DISCUSSION
We report a seroprevalence of Brucella spp.among rural people of 11.1% (with a range between provinces from 2.3% to 22.6%) and an annual incidence of 229 per 100 000.The high incidence in the study likely reflects an increase in human brucellosis after to four seropositive members (Table 2).This equated to an annual incidence of seroconversion of 1145 per 100 000 and an overall annual incidence of 229 clinical cases per 100 000.
In the multivariate analysis, only two variables remained associated with being seropositive: being 45 years old and above (adjusted odds ratio [AOR] = 6.9, 95% CI: 5.1-8.7) and being a veterinarian (AOR = 2.8, 95% CI: 1.5-5.0).Among veterinarians who participated in the study, 72.7% assisted in livestock obstetric work, and 50% had direct contact with aborted animal fetuses and placentas.The risk factors for veterinarians was also much higher compared with other occupations (P < 0.001).Brucellosis among rural people in Mongolia Tsend et al the transition in Mongolia from socialism to a market economy leading to livestock privatization and collapse of the veterinary sector. 4 Although several earlier studies also estimated the seroprevalences of Brucella spp. in Mongolia among high-risk people including herders, veterinarians and raw animal processing technicians, 11-14 these differed from our study in time, study design and methodology and should not be compared.The result from our study was higher than the 0.1% to 10.1% reported among high-risk people in other countries, 10,15-21 which is not surprising as Mongolia is ranked second in the world for brucellosis incidence. 5We also estimated a much higher incidence compared with that reported from notification data, 22 despite the fact that we have taken a conservative assumption that 20% of seropositive cases are clinical cases.
According to the multivariate analysis, adults aged 45 years and above and veterinarians had a higher risk for brucellosis.This age group plays an important role in livestock herding and birthing, and veterinarians have direct contact with animals and aborted materials when doing veterinary examinations.We also found seropositives in all age groups, including in young children (four to nine years), which may indicate ongoing exposure and transmission of brucellosis in rural Mongolia.These groups should be targeted with material about protection against brucellosis infection.This study will serve as a baseline of the seroprevalence of Brucella spp. in rural people in Mongolia before the implementation of a nationwide livestock vaccination campaign; it also will be used for ongoing brucellosis surveillance.A decrease of human incidence and repeated sero-surveillance surveys

Funding
The study was carried out in Sukhbaatar and Zavkhan provinces in 2010 with funding from the Swiss Agency for Development and Cooperation in Mongolia.We thank the Mongolian Ministry of Health, the Health Promotion Foundation of Mongolia and the Research Institute of Veterinary Medicine for funding the study in 2011-2012.We also wish to thank the staff of these agencies for their assistance on the study.in humans will indirectly assess the efficacy of the vaccination campaign in livestock. 23ere were several limitations to the study.First, association between human and livestock seropositivity was not assessed in provinces (with the exception of Zavkhan and Sukhbaatar 7 ).There also may have been temporal variations in risk factors for childhood brucellosis, interpretation of reported clinical symptoms for brucellosis based on seropositivity and pathogen exposure that were not captured by the cross-sectional study design.

CONCLUSION
Our study confirms that human brucellosis seroprevalence among rural people in Mongolia is high and that the incidence is much higher than the notification data suggests.As recommended by the Food and Agriculture Organization of the United Nations, the World Organization for Animal Health and the World Health Organization, mass livestock vaccination is required in Mongolia in the mobile livestock production system.Safety measures to avoid brucellosis include wearing protective clothes such as gloves, using metal hooks to collect aborted fetuses and placentas for burial or burning, washing hands after handling livestock and completely cooking liver from small ruminants.This information should be included in educational materials to prevent as many as possible new cases, especially at the beginning of the mass vaccination campaign while strains still circulate.We have developed written and pictorial educational materials mainly for children.The literacy rate in Mongolia is extremely high and thus printed media are appropriate.In parallel, the surveillance, treatment and diagnostic capacities for human brucellosis must be increased in provinces and districts.Education and awareness programmes should be implemented particularly before the livestock birthing season.

Confl icts of interests
None declared.

Table 1 . Number of participiants seropositive for Brucella spp.* by province, Mongolia, 2010 to 2012
* Based on parallel interpretation of the RBT and ELISA test.

Table 2 . Number of nomadic camps with members seropositive for Brucella spp., Mongolia, 2010 to 2012
* Based on parallel interpretation of the RBT and ELISA test

Table 3 . Univariate analysis of risk factors of brucellosis seropositivity* in Mongolia, 2010 to 2012
* Based on parallel interpretation of RBT and ELISA

Table 4 . Reported clinical symptoms among study participants by age group and sex, Mongolia, 2010 to 2012 (N = 2856)
* Either derived from the  2 test or Fisher's exact test.