Leptospira seroprevalence in dogs, cats, and horses in Tennessee, USA

Abstract

We estimated the Leptospira seroprevalence in dogs, cats, and horses from Tennessee, USA, using the microscopic agglutination test (MAT) against 12 Leptospira serovars. We observed Leptospira seropositivity in 110 of 374 (29.4%) dogs, 21 of 170 (12.4%) cats, and 42 of 88 (47.7%) horses. The highest seroprevalence was observed for serovars Autumnalis (74.6%) in dogs, and Bratislava in cats (42.9%) and horses (95.2%). We found a significant level of potential cross-reactivity between multiple Leptospira serovars tested, with highest cross-reactivity to serovar Autumnalis in dogs. Leptospira seroprevalence was significantly higher in vaccinated dogs (45 of 98 [46%]) compared to unvaccinated dogs (14 of 86 [16%]; p < 0.001). A significant difference in seroprevalence was observed in vaccinated and unvaccinated dogs to all 4 serovars included in canine leptospiral vaccines (p < 0.001). We also evaluated the Leptospira testing results from our diagnostic laboratory submissions from 2021–2023; 103 of 252 (40%) canine serum samples were positive, with the highest positivity rate for serovar Autumnalis. On Leptospira real-time PCR, 35 of 325 (10.7%) urine samples and 15 of 257 (5.8%) blood samples were positive. The cross-reactivity between the Leptospira serovars used in the MAT and vaccination status should be considered when estimating seroprevalence.

Leptospira infection can result in life-threatening disease in humans and animals.^6,12 Leptospirosis is a re-emerging zoonosis that results in ~1 million new human cases and >50,000 deaths annually worldwide. ³ Leptospira infection in domestic animals, such as dogs, cats, horses, and cattle, may lead to fatal disease or reservoir status; however, accurate estimates of animal infections are not available. ⁶ Many mammalian hosts, including rodents, wild animals, cattle, pigs, dogs, and horses, may act as reservoir hosts by harboring Leptospira in their kidneys and reproductive tracts and can continuously excrete the bacteria through urine, thereby contaminating the environment (e.g., raccoons are considered a major reservoir for the serovar Grippotyphosa).^6,24 New infection occurs from direct contact with urine from infected animals and from contact with the contaminated environment.

Animal leptospirosis is well-recognized, and clinical signs may range from mild febrile illness to multi-organ failure and death, or reproductive diseases in livestock species. However, the actual disease incidence is not known given the lack of awareness and the challenges involved in determining an accurate diagnosis. A definitive disease diagnosis can only be achieved by the detection of the infecting Leptospira strain by PCR or culture in a patient with compatible clinical signs. The microscopic agglutination test (MAT) is a widely used method for detecting serovar-specific Leptospira antibodies, and to identify exposure to circulating serovars in a particular geographic region. In humans, a single MAT titer of 800, or a 4-fold difference in acute versus convalescent titer, is recommended for a definitive diagnosis of leptospirosis. ¹² However, interpretation of MAT titer can be confusing in animals due to potential widespread exposure, vaccination, and subclinical infection. The antibody titer depends on the time of exposure, duration of illness, and antimicrobial intervention. In addition, antibody levels may be very low in the early stages of acute infection and may not be detectable at the serum dilution level used in the MAT. The infection prevalence and the infecting or circulating serovar versus species may vary geographically and seasonally; hence, frequent updating of prevalence data is highly desirable.

In a 2020 regional study in the Cumberland Gap region (CGR) close to the intersection of Kentucky, Tennessee, and Virginia, ²¹ 13.1% of the dogs tested were positive for Leptospira DNA in urine by quantitative real-time PCR (qPCR) and 18% were seropositive. Nineteen of these dogs were from shelters in Tennessee, of which only one dog was positive for MAT and none was positive by qPCR. All 50 cats in a study of the seroprevalence of Leptospira in cats in Tennessee tested negative by both MAT and qPCR. ²¹ A literature search using PubMed and Google Scholar did not find any published equine leptospiral studies from Tennessee.

Emerging zoonotic diseases such as leptospirosis are a public health concern, given the potential of animals to become reservoirs and carriers and contaminate their environment. In a 2022 study of soil and water samples from Knoxville, Tennessee, using metagenomic sequencing and analysis, we identified a diverse group of pathogenic Leptospira. ⁹ To improve our understanding of the extent of Leptospira exposure in Tennessee, we investigated the Leptospira seroprevalence in dogs, cats, and horses.

Materials and methods

Samples

We used randomly collected clinical samples submitted for general testing from dogs, cats, and horses between January and August 2022 from the University of Tennessee, College of Veterinary Medicine (UTCVM), Clinical Pathology Diagnostic Laboratory (Knoxville, TN, USA). These samples are considered as biowaste and are discarded after required testing. Sample sizes were calculated as 196, 126, and 381 based on prevalence rates of 15%, 9%, and 45% in dogs, cats, and horses, respectively. ¹ We obtained 374 canine samples, 169 feline samples, and 88 equine samples and stored these samples at −20°C until use. In addition, we evaluated the results from canine Leptospira diagnostic submissions to our laboratory for the years 2021–2023.

Microscopic agglutination test

All serum samples were tested for antibodies against 12 Leptospira serovars, using the standard operating procedure from our Leptospira diagnostic and research laboratory. Briefly, 4–7-d-old cultures of Leptospira serovars, maintained in the laboratory through continuous passage in Ellinghausen–McCullough–Johnson–Harris medium at 28°C, were used for the MAT. The bacterial concentration was adjusted to a spectrophotometer transmittance value of 75–80%. Serum samples (50 µL) at 1:25 dilution in PBS were treated with 50 µL of each serovar in a flat-bottom 96-well microtiter plate and then incubated for 1.5–2 h at 28°C. Homologous antisera prepared in rabbits for each serovar used in the assay served as the positive control and PBS as the negative control. The plates were read using a dark-field microscope, using a 5× long-working-distance objective. Samples with 50% agglutination were considered positive at 1:50 final dilution. Antibody titer was determined for all samples with >50% agglutination.

Data analysis

SPSS Statistics (v.27, v.28; IBM) was used to estimate prevalence. Given that we found MAT positivity to multiple serovars for the same sample, we performed a 2-tailed Spearman rank-order correlation test for the 12 serovars used in our study to determine the potential cross-reactivity between serovars. In dogs, a comparison of seroprevalence between the vaccinated positive and unvaccinated positive groups, as well as the association between serovars included in the leptospiral vaccine, were done using the Fisher exact test. A p ≤ 0.05 was considered statistically significant.

Results

Leptospira seroprevalence in dogs, cats, and horses

We observed Leptospira seropositivity in 110 of 374 (29.4%) dogs, 21 of 170 (12.4%) cats, and 42 of 88 (47.7%) horses (Fig. 1). We observed detectable agglutinating antibody response to 10 of the 12 serovars tested in dogs, but no antibody response to serovars Bataviae and Tarassovi. Among the serovars tested in dogs, the highest seroprevalence was observed for the serovar Autumnalis (82 of 110; 74.6%) followed by Grippotyphosa (44 of 110; 40.0%); the lowest seroprevalence was observed for serovar Hardjo (3 of 110; 2.7%). The MAT titers were 1:50–1:1,600 in canine samples. The highest titers were observed to serovar Bratislava (1:1,600), followed by Autumnalis (1:800), Grippotyphosa (1:800), and Pomona (1:800; Fig. 2).

Figure 1.

Overall Leptospira seroprevalence in dogs, cats, and horses detected by screening serum using microscopic agglutination test.

Figure 2.

Microscopic agglutination test titers to Leptospira serovars tested in dogs. The numbers on the x-axis are the reciprocal titers.

The highest seroprevalence in cats was observed to serovar Bratislava (9 of 21; 43%) followed by Hardjo (8 of 21; 38%); the lowest seroprevalence was observed to serovars Bataviae (1 of 21; 5%) and Mankarso (1 of 21; 5%). Seroprevalence to serovars Ballum and Tarassovi were not observed. The antibody titers were 1:50–1:3,200. The highest titer was observed for serovar Hardjo (1:3,200); however, most of the feline samples only had a 1:50 titer (Fig. 3).

Figure 3.

Microscopic agglutination test titers to Leptospira serovars tested in cats. The numbers on the x-axis are the reciprocal titers.

In horses, all of the serovars tested had a detectable agglutinating antibody response. The highest seroprevalence was to serovar Bratislava (40 of 42; 95%), followed by Copenhageni (24 of 42; 57%); the lowest seroprevalence was to Tarassovi (1 of 42; 2%). The highest titer in horses was observed for both serovars Bratislava and Pomona (1:400; Fig. 4).

Figure 4.

Microscopic agglutination test titers to Leptospira serovars tested in horses. The numbers on the x-axis are the reciprocal titers.

Age, sex, breed, and county distribution

Most of the samples came from 6–10-y-old dogs and cats. For horses, most samples were from 0–5-y-old animals, with the least number of samples in the >30-y-old range (Suppl. Table 1). Of the 374 canine samples, 191 (51.1%) were from female dogs and 183 (48.9%) were from male dogs. Of the 170 feline samples, 78 (45.9%) were from females, 91 (53.5%) were from males, and 1 of was from a cat of unreported sex. Of the 88 equine samples, 26 (30%) were from females, 61 (69%) were from males, and 1 was from a horse of unreported sex.

The sampled canine population was overrepresented by mixed-breed dogs (97 of 374 [25.9%]; Suppl. Table 2). The cat breed distribution was overrepresented by domestic shorthair cats (126 of 170; 74.2%). The horse breed distribution was overrepresented by American Quarter Horses (20 of 88; 23%).

We obtained at least 1 sample from 25 of 95 (26%) counties in Tennessee (Suppl. Fig. 1). Of the 4 regions in Tennessee, most of the samples were submitted from region 1, which includes Knoxville. For the canine samples, 304 (81.3%) were from region 1, 59 (15.8%) from region 2, 11 (2.9%) from region 3, and no samples from region 4. For cats, 145 samples (85.3%) were from region 1, 14 (8.2%) from region 2, 11 (6.5%) from region 3, and no samples from region 4. For horses, we had 65 samples (73.9%) from region 1, 18 (20.4%) from region 2, 4 (4.6%) from region 3, and no samples from region 4. One equine sample (1.1%) did not have a county listed.

Occurrence of potential cross-reactivity between serovars in MAT

Because many of the canine serum samples had positive reactivity to multiple serovars tested, we examined whether this could be the result of cross-reactivity to multiple Leptospira serovars. In addition, the highest seroprevalence was observed for serovar Autumnalis, and isolation of this serovar has not been reported from dogs in our region. A Spearman rank-order correlation assessment suggested potential cross-reactivity between samples (Table 1). The samples positive to serovar Autumnalis also reacted with 7 other serovars (Suppl. Fig. 2). A similar observation was made for horses (data not shown). There were not enough positive feline samples for a Spearman rank-order correlation to be performed.

Table 1.

Spearman rank correlations between microscopic agglutination test positivity to various Leptospira serovars in dogs.

Serovar	Autumnalis	Bratislava	Canicola	Copenhageni	Grippotyphosa	Icterohaemorrhagiae	Mankarso	Pomona
Autumnalis	1.00
Bratislava	0.265	1.00
Canicola†	0.505	0.158	1.00
Copenhageni*	0.625	0.359	0.591	1.00
Grippotyphosa†	0.584	0.190	0.416	0.552	1.00
Icterohaemorrhagiae* †	0.397	0.238	0.398	0.645	0.414	1.00
Mankarso*	0.540	0.323	0.520	0.818	0.529	0.642	1.00
Pomona†	0.731	0.203	0.447	0.608	0.539	0.452	0.596	1.00

All correlations were statistically significant (p < 0.001). Serovars with moderate-to-strong correlation coefficients (>0.4) are in boldface.

^*Serovars in the same serogroup.

^†Serovars present in commonly used vaccines.

Comparison of seroprevalence between vaccinated and unvaccinated dogs

Given that Leptospira vaccination is practiced in dogs, we evaluated the difference in seroprevalence in vaccinated versus unvaccinated dogs. We separated the MAT results from samples for vaccinated and unvaccinated patients using the clinical history obtained from hospital records. Vaccination data were available for 184 of the 374 canine patients. Of the 184 serum samples collected, 98 (53.3%) were from vaccinated dogs and 86 (46.7%) from unvaccinated dogs. Of the samples from the 98 vaccinated dogs, 45 (46%) were MAT positive to one or more Leptospira serovars tested, which was a significantly higher leptospiral seroprevalence than in the unvaccinated dogs (14 of 86 [16%]; p < 0.001). For the vaccinated and MAT-positive dogs, the most common serovars with MAT reactivity were Autumnalis (36 of 45; 80%), followed by Grippotyphosa (22 of 45; 49%), Icterohaemorrhagiae (20 of 45; 44%), and Canicola (18 of 45; 40%).

Leptospira testing results from routine diagnostic submissions to our laboratory

Our Leptospira Diagnostic and Research Laboratory offers real-time PCR (rtPCR) testing of blood and/or urine and MAT, or a panel of both to our veterinary clients. From January 2021–December 2023, we tested 833 samples from canine patients submitted for Leptospira testing (MAT and/or rtPCR); 103 of 252 (40%) of the serum samples were positive by MAT with the highest positivity rate for serovar Autumnalis (Fig. 5). For Leptospira rtPCR, 35 of 325 (10.7%) urine samples and 15 of 257 (5.8%) blood samples were positive.

Figure 5.

Microscopic agglutination test titers to Leptospira serovars in routine samples from dogs. The numbers on the x-axis are the reciprocal titers.

Discussion

The MAT, a test used commonly in Leptospira seroprevalence studies, detects agglutinating antibodies to the lipopolysaccharide components of Leptospira serovars using a limited array of serovars potentially circulating in a geographic area of interest. The test uses live Leptospira cultures maintained continuously in the laboratory and has acceptable specificity because antibodies to Leptospira do not cross-react with other bacterial species. However, sensitivity of this test can be low compared to other tests, especially in the initial stages of infection. In our laboratory, we conduct MAT against 12 Leptospira serovars compared to 6 serovars routinely included in the United States, mainly to track the exposure to new serovars in our region.

We found the highest MAT reactivity in dogs against serovar Autumnalis, followed by Bratislava, Grippotyphosa, and Pomona; our findings were consistent with previous findings.^8,14 MAT is laborious, difficult to standardize, and has levels of subjectivity in reading and interpretation of the results. MAT-positive results may indicate the presence of antibodies from current infection, previous exposure, or recent vaccination against Leptospira. In some cases, infected animals may have a MAT titer below the routinely used titer of 100. A cutoff as low as 1:10 can be used to increase test sensitivity. ⁶ We typically use a cutoff of 1:50 to improve the sensitivity of the assay. It should be emphasized that MAT results should not be used alone because interpretation can be difficult due to low titer levels observed in early infections.

In addition to establishing Leptospira seroprevalence data for dogs, cats, and horses in Tennessee using MAT, we identified multiple issues that may affect the outcome of MAT in surveillance and diagnostic testing. The utility of MAT is limited based on the number of serovars that can be included in the panel. MAT is often described as the reference standard test for facilitating diagnosis of leptospirosis. ²² In our experience, in canine patients in the initial stages of infection, there is little antibody response detected by the MAT, which can confuse the clinician. As peak titers are observed only after 2 wk, a paired titer strategy is recommended for confirmation. Diagnosing clinical leptospirosis in the initial stages of infection in patients with clinical suspicion is critical for implementing intervention strategies to avoid further complications from disease, so pairing the serology tests with agent detection tests such as PCR is highly recommended. Based on the range of titers observed in our study, a single titer at any level should be interpreted with caution. Although this information can help practitioners arrive at a conclusion when combined with other factors, they should be encouraged to conduct combination testing that includes PCR from blood and urine samples.

We identified several factors that could influence the interpretation of serologic data for Leptospira testing and surveillance. One of the significant observations from our study is the shared seroreactivity observed to multiple serovars and even to unrelated serovars. Active infections by multiple serovars may be seen in some rodent reservoirs ¹⁸ ; however, mixed infections are uncommon in clinical cases. In an acute clinical case of leptospirosis, cross-reactivity is common, and results are termed paradoxical titers. Still, in surveillance studies, this type of cross-reactivity is often misinterpreted as true seroprevalence to multiple serovars. Therefore, we pursued the occurrence of potential cross-reactivity of serum to multiple serovars. We found a significant correlation of seroreactivity among multiple serovars, specifically Autumnalis, Canicola, Copenhageni, Grippotyphosa, Mankarso, and Pomona. Autumnalis is not a common serovar isolated from dogs in the United States or elsewhere, and we propose that seropositivity to this serovar should be considered as cross-reactivity. It is noteworthy that some of the serovars that cross-reacted belong to the same serogroup, such as serovars Icterohaemorrhagiae, Copenhageni, and Mankarso, which are in the serogroup “Icterohaemorrhagiae”, and which may explain cross-reactivity among these serovars. However, this is not always true, and data from our previous studies indicate that MAT results may not be serogroup specific. Higher MAT positivity was observed in rats infected with serovar Copenhageni, and no serologic response was detected in rats infected with L. borgpetersenii serogroup Ballum (serovar Castellonis). ¹⁸ Such results should be interpreted with caution to avoid overestimation or underestimation of seroprevalence. Most likely, the serovar Autumnalis MAT reactivity shared with other serovars may result from cross-reactivity to surface antigens common to these serovars.

Vaccination is commonly practiced in dogs using quadrivalent bacterin vaccines that contain serovars Canicola, Grippotyphosa, Icterohaemorrhagiae, and Pomona. Leptospira vaccination status in dogs was only available for 49% of the samples submitted. Although a commercial licensed bacterin vaccine containing L. interrogans serovar Pomona is available for horses, the vaccination status for this species was not available to us, and therefore we could not assess potential effect of vaccination in horses. The higher seroprevalence observed in vaccinated dogs to serovars Canicola, Grippotyphosa, Icterohaemorrhagiae, and Pomona are thought to be related to the vaccine response. It is noteworthy that serovar Autumnalis is not included in the canine vaccine, but the positivity and cross-reactivity observed to this serovar was significantly higher in vaccinated dogs. A pattern consistent with the seroprevalence data from the random canine samples in our study was also noted in routine clinical samples, with highest seroprevalence to serovar Autumnalis. The highest titers consistent with clinical leptospirosis in our routine submissions were observed against serovars Autumnalis, Grippotyphosa, and Pomona. Many of the routine samples positive by PCR, when sequenced, were related to L. kirschneri serovar Grippotyphosa. Leptospira infection is not directly linked to any specific breed of animal, but the literature suggests that for dogs, small breeds and hunting breeds could become infected more often than others.^2,10,19,23

Studies of feline leptospiral infection range from literature reviews aiding diagnoses, ¹⁵ to shedding, seropositivity, and seroprevalence.^16,21 The prevalence in cats in the United States ranges is reported as 4–33%. ¹⁵ All of the cats in a previous study in Tennessee tested negative for Leptospira. ²¹ In our study, the serovars that had the highest number of positive samples were serovars Bratislava and Hardjo, and the highest titer seen in cats was against serovar Hardjo (1:3,200). Positivity to serovar Bratislava agrees with other feline studies both inside and outside the United States^5,11,13,16; a limited number of studies reported samples positive against serovar Hardjo.^11,15 It would have been valuable to assess the difference in seroprevalence between outdoor and indoor cats, but this information was not available to us.

Among the 3 animal species tested, we observed the highest leptospiral antibody prevalence in horses. Studies of equine leptospiral infection range from serologic evidence or diagnosis of leptospirosis in the United States,^17,20 to challenges of establishing experimental Leptospira infection. ²⁶ The reported prevalence in horses in the United States of 2.5–16.4%^1,4 is lower than the 47.7% in our study. In horses, the highest seroprevalence in our study was with serovars Bratislava, Canicola, and Copenhageni, which is in agreement with seropositivity reported to Bratislava and Canicola in previous studies in the United States.^7,25 No studies in the United States have reported the equine breed risk of Leptospira infection. In horses, a bacterin vaccine containing L. interrogans serovar Pomona is available, but we could not retrieve reliable vaccination history for horses to include in our analysis.

One of the limitations of our study was that most samples were from eastern Tennessee, making it difficult to assess the prevalence in other regions accurately. In addition, we used a set of convenience serum samples submitted for other evaluations, not specifically for Leptospira testing; random sampling might have decreased the possibility of selection bias and potentially provided an equal chance of selection. Our results are likely not reflective of the entire Tennessee population. However, the prevalence data from our routine clinical sample submissions from animals suspected to have leptospirosis was similar to the results from the random samples, emphasizing the consistency of our seroprevalence data.