BRUCELLOSIS OF RUMINANTS IN BOSNIA AND HERZEGOVINA : DISEASE STATUS , PAST EXPERIENCES AND INITIATION OF A NEW SURVEILLANCE STRATEGY

Th e current animal health situation in Bosnia and Herzegovina requires the prioritization of diseases for the application of control measures. One of the diseases requiring high priority is brucellosis of ruminants. Brucellosis is a zoonotic infectious disease and one of the most important zoonoses in the world. Brucellosis has been recognized during the past fi ve decades as an important infectious disease in ruminants in Bosnia and Herzegovina. Control and eradication of brucellosis in animals is based on test and slaughter control policy. When the existing brucellosis control program was instituted, the veterinary and animal production sector was almost exclusively owned by the government, an arrangement that promoted compliance with the program and resulted in the successful control of the disease. Th is paper provides an overview of the current institutional and legislative framework for brucellosis control including the laboratory detection system and the epidemiological status of brucellosis in ruminants in Bosnia and Herzegovina. Relevant data were collected during the period spanning from the beginning of  until the middle of . Data we collected reveal an increase in the number of reported outbreaks in ruminants as well as a related increase in the number of human cases. Th is has brought serious consequences to public health, animal health and production and international


Introduction
Brucellosis is an nfectious disease with a long history in the Mediterranean region ().Brucella agents are highly pathogenic for humans, causing one of the most serious zoonoses known ().Since disease is generally not transmissible from human to human (,), prevention of human cases generally depends on the control of the disease in animals ().For example, in Spain, from  to ,  fold decrease of brucellosis prevalence was recorded along with concurrent sharp decline in the number of reported human cases, what was direct result of the national ruminant brucellosis eradication program initiated in early -ties ().Brucellosis poses many challenges in designing effective surveillance system; the infection is chronic in both humans and animals, symptomatology and incubation periods are variable and microbiological confi rmation is the only fully credible mean of fi nal diagnosis.Brucellosis has been recognized during the past fi ve decades as an important infectious disease in ruminants in Bosnia and Herzegovina (B&H).A government program designed in  is currently in eff ect to control the disease.Th e aims of this paper are: to provide an overview of the currently applied brucellosis detection and control measures, to demonstrate the disease patterns based on available data collected for the last seven years, and to propose a surveillance strategy for disease control.

Materials and Methods
Th is paper provides an overview of the current institutional and legislative framework for animal disease control including the laboratory detection system and the epidemiological status of brucellosis in ruminants in B&H.Relevant data were collected during the period spanning from the beginning of  until the end of June, .A presentation and descriptive analysis of current measures used for detection, control and prevention of brucellosis in ruminants is provided.Th is includes detailing protocols existing in the animal health laboratory network and providing the current scheme of disease information distribution among involved government veterinary agencies and animal health laboratories.Th e analysis is accomplished primarily through a retrospective assessment of relevant legal acts and their consecutive changes over time.Data on methods, frequency and extent of serum sampling were collected through interviews with representatives of animal health laboratories.Disease levels and trends were obtained using data from the national infectious diseases data base, pub-lished scientific papers and an investigation into the numbers and origins of samples submitted for routine laboratory testing.Th e data on disease occurrence in animals and humans were presented in respect to their temporal and spatial distribution in order to establish epidemiological trends.Established trends for brucellosis in animals were presented using the frequency of reported outbreaks, while a trend of brucellosis in humans was presented using frequency of reported cases.Definition of an outbreak of brucellosis in animals is adopted from the offi cial Regulations on measures for control and eradication of brucellosis in cattle, sheep, goats and swine, were an outbreak represents occurrence of one or more brucellosis cases among animals of same species owned by the same farmer ().Univariable linear regression models were constructed separately for cattle and sheep in order to quantitatively describe observed temporal trends in brucellosis occurrence in animals ().Same regression analysis was used to describe temporal trend of brucellosis occurrence in humans.Statistical signifi cance of the regression models was assessed using Student t test for significance of regression coeffi cients and F test for signifi cance of regression model, at  level of statistical signifi cance.In order to establish relationship between brucellosis occurrence in ruminants and humans we used correlation coefficients ().Correlation coefficients were calculated for every pair of variables; number of outbreaks in cattle vs. number of outbreaks in small ruminants, number of outbreaks in cattle vs. number of cases in humans and number of outbreaks in small ruminants vs. number of cases in humans.Calculated correlation coefficients were statistically assessed using Student t test at  level of statistical signifi cance for two sided test ().Regression analysis, calculation of correlation coefficients and all statistical testing were accomplished using MINITAB  (Statistical software for Windows, © Th omson learning).Lack of reliable estimates on ruminant population in the country and inconsistent and incomplete data on the number of serum samples collected for brucellosis testing among animals had excluded possibility of reliable estimates of disease rates.We also identifi ed constraints and limitations for the interpretation of the results provided.

Detection and control measures for brucellosis in animals in Bosnia and Herzegovina
Control measures for brucellosis have undergone several revisions since they were first instituted in  (-).Almost two decades have passed since the last revision of brucellosis detection and control measures.When the existing brucellosis control program was instituted, the veterinary and animal production sector was exclusively owned by the government, an arrangement that promoted compliance with the program and resulted in the successful control of the disease ().Sporadic local outbreaks were reported in small ruminants (mainly sheep) after the introduction of new animals originating from within the former Yugoslavia into unexposed herds (, ).Human cases of brucellosis were rare and principally occurred in veterinarians and farmers as a consequence of occupational exposure.
Currently, brucellosis detection is provided through serological screening of: all imported ruminants during quarantine, all dairy cattle (once a year) and all clinically suspect ruminant animals ().Th e primary cause for clinical suspicion of brucellosis is the occurrence of abortion in late pregnancy.After laboratory confi rmation, positive reactors are slaughtered and their remains buried.Quarantine measures are imposed on the herd/fl ock of origin.During quarantine, all animals in the herd/fl ock are sampled and tested using reference laboratory procedures at least twice within a  day period after identifi cation of the fi rst case.Th e absence of cases in some parts of the country cannot be interpreted as absence of disease, because of the small numbers or lack of samples submitted for testing from those areas during the past seven years (absence of evidence is not evidence of absence).Th e north-south pathway of disease spread coincides with the seasonal movement of sheep towards mountain pastures in the central south region in spring and towards valley pastures in the northern region in autumn.

Occurrence of brucellosis in ruminants in
The seasonal trend of brucellosis occurrence over the last seven years also demonstrates an increase in the numbers of identified outbreaks in ruminants.Within a year the largest number of brucellosis cases in ruminants is observed during the second and third quarters of the year (Figure ).At fi rst glance, it appears that the identifi ed pattern of brucellosis occurrence does not coincide with the increased disease occurrence usually seen during the lambing/calving season (fi rst quarter of the year).Th e lambing/calving season is quite extended in B&H, due to relatively significant climate variations in some regions.Also, the time period from sampling to laboratory confirmation and identification of an outbreak can be, in some instances, over one month.Therefore, at least some of the infections that occurred during the fi rst quarter of the year were accounted for in the second quarter.This assumption is supported by data from years -, when a more extensive and rigorous sampling campaign were undertaken.Th e spatial trend of spreading ruminant brucellosis in the country correlates with the observed pattern of increase in the number of reported brucellosis outbreaks in animals over the last seven years.Thus, it can be claimed that the current brucellosis detection system in B&H can detect trends, but it fails to provide reliable estimates of brucellosis levels in ruminant population.
The reported outbreaks and cases shown here provide limited input for precise quantification of disease levels.Due to the limitations of our data sources (national disease data base contains only confirmed positive cases/outbreaks), disease proportions and rates could not been calculated.

Evaluation of the current brucellosis control measures
Compliance with animal brucellosis control requirements is currently lower than when these control measures where fi rst instituted, due to the economic transition that has occurred.The primary components of the disease reporting system (farms, dairies, veterinary practices and laboratories) are no longer exclusively under government control.The effectiveness of the implementation of regulations is also compromised by the small ruminant management system (free range pasture, and transhumance), and the lack of coordination between veterinary service agencies and veterinary laboratories.Additionally, established high correlation between increase of the number of reported brucellosis cases humans and outbreaks in small ruminants, indicates that small ruminants, rather than cattle are the primarily source of infection in humans.Screening of imported animals is mandatory and is applied to all legally imported ruminants.Animals whose milk is bought by milk processing plants are tested once a year, and their milk is pasteurized.According to reported experiences of other countries, widespread application of pasteurization during milk processing significantly reduces the probability of indirect transmission of brucellosis from animals to humans ().Bosnian Dairies generally process milk from cattle only.Th erefore, pasteurization is almost exclusively applied to cow's milk.Large quantities of dairy products, especially those of small ruminants (sheep), are frequently home-produced by farmers and sold on local markets.Consumption of these products is widespread, due to the public perception that they are organically produced.Inconsistent, and in some instances inadequate, processing of these products, traditional dietary habits and lack of education on methods of disease transmission to humans represent signifi cant public health risk factors () ().Th e lack of motivation by farmers to report suspect cases to the local veterinarian, aside from fi nancial and practical reasons, can also be explained by community judgment that occurs towards owners of fl ocks in which infected animals are detected.Th is arises primarily from poor education of the public about this disease.Th e application of Rose Bengal and complement fi xation tests in series is the prescribed testing protocol in current brucellosis detection requirements and it is also recommended by relevant international authorities such as OIE and EU.Nevertheless, many studies have demonstrated substantial inadequacies in this testing regime when applied in the detection of brucellosis in small ruminants ().As a result of the above mentioned limitations, the offi cial data on disease occurrence are often unreliable, incomplete, and unrepresentative.

Conclusion
Th e approach to control, prevention or eradication of brucellosis in a country or region will depend on many factors, such as the level of infection in the herds or fl ocks, type of husbandry, economic resources, public health impacts, and potential international trade implications.Decision-making by those charged with policy making is likely to be intuitive unless accurate and current epidemiological information is available.Even though the brucellosis detection system in Bosnia and Herzegovina indicates an increasing trend for disease in ruminants, the numbers of cases/outbreaks that have been reported are still low relatively to the estimates for the overall ruminant population.However, the evaluation of the brucellosis detection system provided in this paper indicates available data on brucellosis occurrence are not reliable.
An assumption that large numbers of brucellosis cases in animals are not reported is likely valid.Th is is supported by reports of increased numbers of brucellosis cases in humans.Th is fi nding would not be possible if disease levels among animals were low and sporadic as offi cial data indicate.Also, the spatial distribution of brucellosis outbreaks in animals tracked over time presents an indirect verifi cation of this underreporting.Th ese data indicate ruminant brucellosis spreads to wider areas from year to year.Brucellosis persistence in some areas, despite the implementation of control measures following initial outbreaks, raises questions about the effi ciency of measures used.Based on this, our conclusions are: y Th e implemented brucellosis control measures have not halted its spread, y Human health in this country is seriously threatened by brucellosis and y Th e current brucellosis detection system cannot be used to provide reliable estimates of disease prevalence and incidence and cannot be used to establish a scientifi cally based surveillance program.A well-functioning surveillance system, fed by valid data collected from the fi eld, must be established before a brucellosis control program can be designed and implemented.
SABINA ŠERIĆHARAČIĆ ET AL.: BRUCELLOSIS OF RUMINANTS IN BOSNIA AND HERZEGOVINA: DISEASE STATUS, PAST EXPERIENCES AND INITIATION OF A NEW SURVEILLANCE STRATEGY Bosnia and HerzegovinaData on brucellosis occurrence we collected indicate an increase in the number of established outbreaks in ruminants, especially sheep and goats, as well as a disturbing increase in the number of human cases (Figure).SABINA ŠERIĆHARAČIĆ ET AL.: BRUCELLOSIS OF RUMINANTS IN BOSNIA AND HERZEGOVINA: DISEASE STATUS, PAST EXPERIENCES AND INITIATION OF A NEW SURVEILLANCE STRATEGYThe summary statistics for the regression models that describe observed increase is provided in Table .According to results of the regression analysis, each year the number of recorded cases in humans is increased for about  cases more than previous year, while the number of outbreaks in sheep is for  outbreaks larger than year before.A signifi cant increase in the numbers of human cases, followed by the resultant public outcry, triggered a more rigorous testing policy, and increased government funds were allocated for brucellosis control during the past three years.An expanded testing program had presumably contributed to an increased number of detected brucellosis cases in animals during this period.Th erefore, regression models established using data predominantly from offi cial sources as it is case in this paper, can not reliably quantify increase of ruminant brucellosis occurrence since it is not clear what portion of this increase is contributed by the expanded testing program and what is the direct consequence of disease spread.Th e spatial pattern of disease indicates a spreading trend (Map ).Th e most prominent clusters of outbreaks of animal brucellosis over the last seven years period are seen in areas of Banjaluka, Bugojno, Konjic, Nevesinje, Novi Travnik, Sarajevo, Tomislavgrad, Travnik and Zenica.

TABLE 1 .
Th e summary statistics for the temporal trend models for the occurrence of brucellosis over last seven years (*p value is based on Student t test, ** p value is based on F test) TABLE 2. Th e correlation coeffi cients with corresponding p values for every pair of assessed variables (the contents of cells represents Pearson' s correlation coeffi cient and corresponding p value)