Over this five-year investigation, 1168 wild boar were tested. The results show that brucellosis is endemic in wild boar in southern Belgium. By comparison with a previous study carried out in the same region with the same iELISA , an upward tendency was observed in the present work. As expected, the four serological tests showed different results, the highest apparent seroprevalence being observed with the iELISA (54.9%). These differences are due not only to the intrinsic factors of each test but also to the immunoglobulin classes that the tests target . The iELISA detects IgG1 and IgG2 (isotopes detected with protein G). These immunoglobulins are present in the later stage of infection and persist over a long period of time . The CFT and RBT mainly detect IgG1, and the principle of SAT is to detect anti-agglutinin antibodies mainly of the IgM isotype, markers of acute infection. In this study, the large proportion of wild boar showing positive results in the iELISA and a lower seroprevalence in the SAT suggests that brucellosis is chronic in wild boar populations in the investigated area. Godfroid and coworkers report the same trends, with apparent seroprevalences of 39.72% (31.64-47.79) and 0.71% (0.00-2.09) measured by iELISA and SAT respectively . Although the relative sensitivities of the CFT, RBT, and SAT are low compared to that of the iELISA, their relative specificities are above 90%. A few ELISA-negative samples were positive by CFT or RBT, but positive reactions were weak in these tests. On the other hand, both ELISA-negative, SAT-positive samples showed a high titer in the SAT test. These boar might have been sampled at an early stage of the humoral response, before the appearance of IgGs. The lesser sensitivities of the CFT, RBT, and SAT are confirmed by the serological profiles of the PCR/ELISA-positive boar.
The apparent seroprevalences reported in the present study are higher than those mentioned in other European reports. Results obtained with an iELISA were around 35% in the Canton of Jura, Switzerland , 22% in northeastern Germany , and ranged from 25% to 46% in different regions of Spain . In France, the apparent seroprevalence varied from 20% to 35% according to the department , but these results were based on CFT and RBT used in serial or parallel testing. Serological results should be compared with caution because of the different tests used.
The circulation of pathogens within wild populations may be influenced by artificial management, and notably by fencing, translocations, and feeding. The impact of these factors is not fully understood, however. In one study, several pathogens including Brucella showed higher prevalences in fenced estates than in open ones , but other surveys conducted in Spain  and France  showed no relationship between the apparent prevalence of brucellosis and wild boar management or density.
In southern Belgium, it is forbidden to fence hunting estates, but artificial feeding has been practiced in recent years as supplementation during the winter, as a dissuasive measure aiming to reduce crop damage by wild boar or as attractive measure during hunting season. Clearly, artificial feeding causes spatial aggregation of wild boar and thus, presumably, increased contacts among animals. This factor, associated with the steady increase in wild boar populations, could be linked to the rise in seroprevalence detected between 1994 and 2007 in Belgium.
Our study, however, shows no significant differences in prevalence between regions, even though inter-regional disparities in wild boar densities are observed. According to the official 2007 census of wild boar populations , the Famenne region has the highest density: 68 individuals per 1000 ha, versus 39 per 1000 ha in the Condroz. These official data are means for defined regions, and do not take local aggregation into account. The 21 hunting zones investigated in this study were all characterized by high wild boar density, as evidenced by annual hunting bags.
In the Famenne and Ardenne regions, where high wild boar densities overlap with outdoor pig farming, there is an increased risk of disease transmission between wild and domestic suids. A similar situation is reported in Switzerland . Preventive actions should concern both farmers and hunters. Outdoor pig farming requires setting up double fences and maintaining sows indoors during heat. On the other hand, hunting stakeholders must control translocations and the release of farm-bred « wild boar », which is strictly prohibited. Likewise, surveillance strategies for brucellosis must include both pigs (in-depth analysis of reproductive disorders) and wild boar. In the latter, surveillance is all the more important because Brucella infection is silent in most cases. As observed previously [4, 8], the seroprevalence increases with age and does not differ between females and males.
The seroprevalence was significantly higher in 2007 than in previous years of sampling. The reason for this increase is unknown, but as mentioned above, it could be related to the steady rise in wild boar numbers. Further studies will confirm or not an upward trend.
Brucella infections can be diagnosed unequivocally only by isolating and identifying Brucella spp. or detecting the bacterial DNA by PCR . The PCR used in this study is known to be highly specific for Brucella spp. [23, 24]. Isolating the bacteria by culture appears less sensitive than the PCR, since only 1 out of 4 PCR-positive samples yielded a positive culture, whatever the organ. A lower sensitivity of the culture method has been reported in other studies [25, 26] and is partly related to wildlife sampling conditions, which do not always guarantee good quality of the sampling, adequate transport, and proper tissue conservation. This suggests that a PCR approach, considered more robust under field conditions, should be preferred in large-scale and long-term wildlife surveys. Accordingly, a field study performed in Switzerland and using IS711 real-time PCR gave excellent results for detection of Brucella spp. infections in wild boar .
Whatever the method used in our study, Brucella or its DNA was more frequently detected in tonsils than in spleen. The recorded bacteriological prevalences were 8.33% (culture) and 32.20% (PCR) for tonsils versus 4.89% and 5.63% for spleen. This could mean that in splenic tissue the concentration of Brucella DNA is lower, and often below the detection limit of the test. Real-time PCR assays, having higher sensitivity , could increase detection of the bacteria within the selected organs. Another cause of false-negative PCR results might be the presence in the spleen of polymerase inhibitors such as blood constituents [26–28]. Variability in the location and quantity of Brucella bacteria among targeted organs has been documented  and is likely to correlate with the stage of infection in individual animals.
Between the iELISA and PCR results, discrepancies were observed. The high number of iELISA-positive samples that were PCR-negative is worth stressing. This might reflect a lack of sensitivity of the PCR method, a lack of specificity of the iELISA, or both, to some extent. No “gold-standard” test is available for determining the true serological prevalence towards Brucella spp. . The high-sensitive iELISA was used in first step for optimizing the detection of carriers at the expense of specificity. Thus the overall prevalence could be overestimated. Nevertheless the iELISA detected all the PCR-positive boar, unlike the three other tests. Cross-reactions seen with other bacteria are a well-known problem in serological diagnosis of brucellosis [25, 29]. Because of a shared O-chain on the lipopolysaccharides (LPS), humoral responses induced by some bacteria and notably Y. enterocolitica O:9 (Y09) are indistinguishable in iELISAs from those induced by smooth Brucella species. In the present study, although no Yersinia sp. was isolated from tonsils, the real impact of Y09 on our serological results remains unknown. Y09 has been found in 2.6% of examined wild boar in Switzerland . Alternatively, the discrepancy between the number of seropositive animals and the number of PCR-positive animals might be due, in part, to past infections, with clearance of the bacteria from the host and persistence of anti-Brucella antibodies. This “self-limiting” process has been observed in cattle experimentally infected with B. abortus biovar 1 or B. suis biovar 2 .
All wild boar samples (n = 35) isolated in the present study were identified as B. suis biovar 2. These results are in agreement with previous studies carried out on wild boar and hares in Europe [2–9, 31]. Yet the isolation of B. suis biovar 3 from pigs, wild boar, and horses in Croatia shows the emergence of zoonotic biovars in Europe [14, 16]. Translocation of wild boar for breeding or hunting purposes increases the risk of spreading zoonotic brucellosis among neighboring countries. In this context it remains important to biotype Brucella strains isolated from wildlife in surveillance programs throughout Europe.