Introduction Equine viral arteritis (EVA) is one of the most important equine viral diseases both because of the clinical presentation of the disease and as one of the single most important causes of abortion in mares. Despite the global widespread distribution of the disease, the first evidence of the equine viral arteritis virus (EAV) circulation in the Republic of Croatia was confirmed in 2005. The confirmation of positive animals in different parts of Croatia in 2008 led to the introduction of mandatory disease control measures in 2009, which have been in place ever since. For disease control purposes, the frequent asymptomatic infections that are dominant among the carrier stallions who can remain lifelong virus carriers, have been the biggest challenge. Therefore, the control measures are based on the serological testing of animals and the exclusion of carrier stallions from the breeding programme. In line with the recommendation of the World Organisation for Animal Health (OIE), serological diagnostics is based on the virus neutralisation test (VN-test). As this method is very demanding in practice, sometimes the enzyme-linked immunoassay (ELISA) is used as an alternative. However, the reliability of the use of the ELISA method has to be assessed against the area-specific individual epizootiological situation. The epizootiological characteristics of the EVA in an area depend on the breed structure, the methods of keeping and use of horses, as well as the circulating virus strains, which will ultimately determine the clinical presentation of the disease. In Croatia, horsebreeding is characterised by the indigenous horse breeding programmes, with the most dominant being the coldblooded breeds such as the Croatian Posavina horse and the Croatian Coldblood horse, which are mostly bred in Central Croatia and often by using extensive farming methods; and the lipizzaner horse, which is the most widely bred horse breed in Eastern Croatia. An analysis of these two biggest groups of horses already clearly shows that their respective epizootiological characteristics are mutually very different. Additionally, we also have to take into account the growing population of warmblood horses, which are kept for sports and leisure, and have a different set of epizootiological characteristics. The epizootiological characteristics of the EVA have not been examined in detail in Croatia so far, and neither has been evaluated the effect of the mandatory measures despite the fact that they have been implemented for more than a decade. Furthermore, the diagnostic potential of the commercial ELISA assays in our epizootiological situation has not been sufficiently examined and no molecular typing of the circulating EAV strains has been carried out to the present day. These existing gaps, in addition to the fact that the disease is present in Croatia, were the main reasons that lead to this survey with the aim of investigating of the epizootiological situation, determining the risk factors, evaluating the effectiveness of the mandatory control measures, examining the serological diagnostic alternatives and expanding the knowledge with respect to the EVA molecular epizootiology; all against the backdrop of the specific epizootiological features of the horse breeding in Croatia. Methods In order to achieve the objectives of this study, the test results for a total number of 4445 horses from across the Croatian territory that had been serologically tested by VN-test in the period 2010-2015, were analysed. For each individual sample, the analysis included the year of sampling, the area where the tested animal was kept, the breed, age, sex and category of the animal, as well as VN-test result. Based on the above, the seroprevalence was established for the reference period, as well as for each individual year that fell within the temporal scope of the survey at the level of the entire country and then at the level of the three individual regions (Eastern, Central and Coastal Croatia). These regions were selected for the survey based on the grouping of the counties with similar epizootiological situations, so as to obtain a satisfactory sample for determining the risk factors. Due to an imbalanced sample, such an analysis could not be conducted at the county level. The data that were thus grouped were then analysed to determine the differences and risk factors in relation to the region, age, sex, gender, category and the year covered by the survey. All data were statistically processed to determine the significance of the obtained result and the risk factors and their significance were determined by establishing the odds ratio with a confidence interval of 95% and p<0.05. Additionally, the correlation was determined by using the Pearson correlation coefficient (r) and the results were shown as r and a confidence interval of 95%. A logistic regression analysis was also carried out, as well as an analysis of the relative significance of individual factors in the outcome of the VN-test. With the aim of analysing the possibilities of using commercial ELISA test kits for the serological diagnostic purposes, 94 horses were tested using VN test and two commercial immunoenzyme tests: ID Screen® Equine Viral Arteritis Indirect (ID.Vet, Grabels, France) and INgezim Arteritis 2.0 (Eurofins Technologies Ingenasa, Madrid, Spain). The sensitivity and specificity, as well as the reliability of the commercial ELISA tests, were determined by comparing the test results. A total of 16 archived RNAs isolated from the viruses confirmed in the Osijek-Baranja County and the Brod-Posavina County were used for the molecular epizootiology examination. For molecular typing, the nucleotide sequences of the EAV ORF5 gene were analysed. After the successful multiplication of the ORF 5 gene confirmed by gel electrophoresis, the nucleotide sequences were determined in Macrogen Europe (Amsterdam, the Netherlands) and obtained sequences were phylogenetically compared with the 138 nucleotide sequences of the same EAV gene isolated in the world with the known year and country of isolation, and which were freely available in the gene bank (GenBank). Results The seroprevalence in the territory of Croatia decreased from 21.3% to 9.7% in the 2010- 2015 period. It was demonstrated that with every successive year, the odds of infection for horses was 0.77 lower (OR=0,77, 95% CI 0,7-0,84, p<0,001). Eastern Croatia is the area with the statistically significantly higher seroprevalence of EAV infections and there is a high correlation between the seroprevalence in this area and the seroprevalence at the level of Croatia (r=0.86; 95%CI=0.16-0.98; p=0.03). The same was demonstrated for the seroprevalence of stallions in the same region (r=0.88; CI 95% 0.22-0.99; p=0.02). Among the risk factors, the most impactful on the seropositivity throughout the period is the rising age of the animal so that, when compared to the animals below five years of age, the probability of the positive result increases by factor 3.99 in animals aged 5-10 years (OR=3,99, 95% CI 2,55-6,43, p<0,001), and by as much as factor 16.5 in animals older than 10 years (OR=16,53, 95% CI 10,6-26,65, p<0,001). Female animals are 1.96 more likely to have a positive test result (OR=1,96, 95% CI 1,44-2,68, p<0,001), as well as the horses of lipizzaner breed, who are 6.46 times more often positive than other warmblood horses (OR=6,46, 95% CI 4,17-10,31, p<0,001). The scope of application of the mandatory control measures were from 60.3% to 91.5% at the level of the breeding animals, and from 18.6% to 23.6% at the level of the general animal population. In addition, incomplete epizootiological data was present in 23.9- 45.9% of the cover letters received with the laboratory sample. Serological diagnosis of EVA monitoring in the equine population in the Republic of Croatia can be carried out with specific ELISA tests but only after validation in local epizootic conditions. The phylogenetic analysis confirmed a close link between the EAV strains isolated in the territory of the Osijek-Baranja County and they belong to the EU-1 virus group. The strains confirmed in the territory of the Slavonski Brod-Posavina County were mutually matching and phylogenetically distant from the strains found in the Osijek-Baranja County. These viruses, together with the strain isolated in the Republic of Serbia and the strain isolated in the United States in 1963, form a separate branch on the phylogenetic tree. Conclusions Equine viral arteritis is widespread across the Republic of Croatia with the highest seroprevalence in Eastern Croatia. After the introduction of the mandatory control measures, the seroprevalence has been continuously decreasing, and an even better result could be achieved by the full implementation of the mandatory measures. The most significant risk factor is the increasing age of the animal, and female animals are more often positive than male animals. In addition, the infections are significantly more frequent in horses of lipizzaner breed. Seroprevalence results analysis indicate two previously unreported epizootics of the EVA in Central and Eastern Croatia, which additionally emphasises the deficiency in mandatory measures conducting. In the serological diagnostics of EVA, the method of choice can be ELISA tests but only after validation in local epizootic conditions. Although the ID Screen® Equine Viral Arteritis Indirect test has a higher matching rate of results in terms of routine laboratory diagnostics, it is not acceptable due to the potential false-negative results. In view of all of the above, and the proven co-circulation of at least two different virus strains of EAV in Croatia, it is necessary to continue the implementation of the control measures to the fullest extent, placing an additional emphasis on the lippizaner horse populations in Eastern Croatia in order to successfully control and eradicate this disease that has a significant impact, both economically and in terms of the breeding programmes. This is the only way to ensure the future development of the horsebreeding in Croatia and protect the national heritage that is the Croatian indigenous breeds.