Most horses infected with Lyme disease are asymptomatic [7], but some show encephalitis [8], meningitis, cranial neuritis, radiculoneuritis [9], lameness [10], arthritis, panuveitis [11], and pseudolymphoma [12]. In infected humans, erythematous rash, neurological signs, and arthritis have been reported, but as with horses, some patients do not experience any specific symptoms [13]. According to the Centers for Disease Control and Prevention, 300,000 people are infected with Lyme disease every year in the United States alone, and this number is increasing.

Lyme disease has been reported in other countries of Asia, including China [14], Malaysia [15], and Japan [16]. In Korea, Park et al. [17] first reported B. burgdorferi infection from Ixodes ticks and mice, and until now, serological and molecular investigations have been performed in humans, dogs, and ticks [2,18,19]. According to Kim et al. [20], Haemaphysalis longicornis is the most prevalent tick species in Korea, but the existence of other species of Haemaphysalis and Ixodes has also been documented.

Recently, the equine industry in Korea has expanded, and the number of horses is increasing. However, only a few studies exist about disease distribution in horses, and none is concerned with tick-borne diseases [21,22]. Because of global warming and climatic changes, it is expected that tick-borne disease will become more prevalent in Korea [23]. Thus, this study was initiated to assess the infection status of B. burgdorferi, using ELISA and PCR, in horses reared in Korea.

Blood samples were collected from the jugular vein of 727 horses, from 2009 to 2013 throughout different localities of Korea. The overall number of horses in Korea is approximately 30,000. The sample size in this study was determined using the following formula, with an expected prevalence of 5%, and a desired absolute precision 5% with a simple random sampling design [24]:

n=1.962pexp (1-pexp)d2, where n=required sample size, p_exp =expected prevalence, and d=desired absolute precision. According to the formula, a minimum of 292 samples were needed, and the samples were collected from various regions of Korea. Data for each animal including age, gender, breed, and region of sampling were collected for epidemiological analysis. Cases with insufficient data were grouped as ‘unknown’. The studied areas were divided into 4 regions: northern, central, southern, and Jeju-do (Fig. 1).

To detect anti-B. burgdorferi antibodies in sera, the SNAP^® 4Dx test (IDEXX Laboratories, Westbrook, Maine, USA) was used according to the manufacturer’s instructions. Although the SNAP^® 4Dx test was originally manufactured to detect antibodies in dogs or cats, it also showed useful results to detect antibodies in horses [25,26]. In horses, the sensitivity and specificity of the SNAP^® 4Dx test are 95% and 100% when compared with immunofluorescence assay [25].

For PCR, DNA was extracted from horse blood using the DNeasy^® Blood & Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. To detect a specific gene of B. burgdorferi and increase the possibility of detection, nested PCR (nPCR) was carried out using 2 sets of primers; Bb23S3/Bb23Sa and Bb23S3nF/Bb23SanR which amplify a 226-266 bp fragment of B. burgdorferi 5S-23S rRNA intergenic spacer [27], and BbospF/BbospR and BbospnF/BbospR which amplify a 314 bp fragment of the B. burgdorferi ospC gene [4]. For each nPCR analyses, B. garinii DNA isolated from a dog-attached tick was included as a positive control. The amplicons were determined on 1.0% agarose gel stained with ethidium bromide.

The observed prevalence (OP) was calculated as the number of positive samples/total number of samples. The true prevalence (TP) was estimated using the following formula: TP=(OP+Sp-1)/(Se+Sp-1), where Se and Sp are sensitivity and specificity, respectively [28]. The chi-square or Fisher’s exact tests were used for statistical analysis among variables. The ‘unknown’ group was excluded from the statistical analysis. Statistical analyses were carried out using SPSS 21.0 (IBM Corporation, New York, USA), and P values <0.05 were regarded as significant. The 95% confidence interval for the adjusted prevalence of each estimate was calculated using Blaker’s method [28].

Out of 727 horse blood samples, 38 (5.2%; TP: 5.5%) tested positive for B. burgdorferi using ELISA (Table 1). None of the tested samples showed positivity for B. burgdorferi by 2 different nPCR experiments. The seroprevalence was 5.6% (11/197; TP: 5.9%), 1.7% (4/236; TP: 1.8%), 3.0% (6/200; TP: 3.2%), and 18.1% (17/94; TP: 19.0%) for horses aged <5 years, 5-10 years, >10 years, and of unknown age, respectively. With respect to gender, 3.2% (4/124; TP: 3.4%), 4.0% (10/250; TP: 4.2%), 2.7% (7/259; TP: 2.8%), and 18.1% (17/94; TP: 19.0%) of male, female, castrated, and unknown gender horses were seropositive, respectively. In terms of breed, 3.6% (17/477; TP: 3.8%), 15.6% (17/109; TP: 16.4%), 6.6% (4/61; TP: 6.9%), and 0% (0/80; TP: 0%) of serology results were positive for thoroughbreds, Korean native ponies, warm-bloods, and mixed breeds, respectively. Regarding regions, 2.4% (5/211; TP: 2.5%), 2.2% (4/179; TP: 2.4%), 4.9% (12/243; TP: 5.2%), and 18.1% (17/94; TP: 19.0%) were seropositive in the northern, central, southern, and Jeju-do regions, respectively. There were statistically significant differences according to breed and region (P <0.001).

In this study, statistically significant differences were observed between our ELISA results in terms of region and breed. Climate conditions influence the distribution and abundance of ticks that transmit B. burgdorferi [29]. In Korea, lower latitude regions (the southern and Jeju-do regions) are warmer and receive more precipitation than the higher latitude regions (the northern and central regions); these factors are important in tick ecology. Therefore, the more favorable climate conditions in the southern and Jeju-do regions may increase tick abundance and result in greater exposure of horses to B. burgdorferi.

Although a statistically significant difference according to breed was observed in the ELISA results, the authors do not think that the seroprevalence was affected by breed. Korean native ponies and Jeju-do were the breed and region with the highest seroprevalence, respectively. Among the 109 Korean native pony samples, 94 (86.2%) were collected in Jeju-do, and all the seropositive samples were collected from Jeju-do (data not shown). Therefore, the higher seroprevalence in Korean native ponies is likely due to the higher seroprevalence in the Jeju-do region.

There were no statistically significant differences in the seroprevalence according to age and gender. Older horses might be expected to show higher seroprevalence than younger horses, as older horses are more likely to have been exposed to vector ticks with a reservoir of B. burgdorferi, and antibodies against B. burgdorferi can exist for long periods of time [30]. From our ELISA data and the fact that Lyme borreliosis is a tick-borne disease, the authors cautiously suggest that regional factors related to climate are more important than age and gender in rates of equine B. burgdorferi infection.

In contrast to the ELISA results, when tested using 2 different nPCR experiments, none of the samples tested positive. The authors attribute this to the properties of the PCR and ELISA used, which detect antigens and antibodies, respectively. Borchers et al. [13] reported that several weeks are needed for humans to develop antibodies against B. burgdorferi. In the same study, IgM was not detected until 1-2 weeks after infection; IgG required 4-6 weeks to develop, but then persisted for several years. Although there are no known data on B. burgdorferi antibody duration in horses, antibodies could persist for long periods of time, influenced by the timing and frequency of exposure to B. burgdorferi [30]. Therefore, a positive ELISA result can indicate only that the tested horse was previously exposed to B. burgdorferi, with no certainty over the time of exposure. Comparing the PCR and ELISA results from this study, the authors can conclude that most of the horses tested were exposed to B. burgdorferi at a previous time, but not at the current time.

To our knowledge, this study describes the nationwide prevalence of B. burgdorferi in horses reared in Korea for the first time. Conclusively, this study indicates that B. burgdorferi exists in Korean horses, and that the prevalence is greatest in Jeju-do region. Since global warming is likely to increase the abundance of ticks in Korea, continuous monitoring of tick-borne diseases in Korean horses is needed.