Material and Methods

Study design—A cross-sectional study design was used to determine the prevalence of antibodies against various Leptospira serovars and risk factors associated with positive Leptospira titers in healthy client-owned dogs in Michigan. A structured questionnaire was mailed to 767 private small animal veterinary practices in the lower peninsula of Michigan to solicit participation in the study. The state was divided into 4 regions (northern, central, southeastern, and south central-western; Figure 1), and a proportional random sample of 150 practices was selected from the practices interested in participation, to ensure that the number of practices in the study would be proportional to the actual number of practices found in each geographic region. Participating practices were asked to submit serum samples from up to 20 healthy client-owned dogs between August and December 1999, when the highest antibody concentrations in dogs exposed to Leptospira species would be expected to occur, on the basis of the seasonal nature of canine leptospirosis in the United States.^1,11,13

Map of the lower peninsula of Michigan illustrating the geographic regions in a study of the prevalence of antibodies against 6 Leptospira serovars in 1,241 healthy dogs in 1999. Values indicate overall percentages and numbers of dogs with positive titers against any of the 6 serovars.

Citation: Journal of the American Veterinary Medical Association 230, 11; 10.2460/javma.230.11.1657

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Map of the lower peninsula of Michigan illustrating the geographic regions in a study of the prevalence of antibodies against 6 Leptospira serovars in 1,241 healthy dogs in 1999. Values indicate overall percentages and numbers of dogs with positive titers against any of the 6 serovars.

Citation: Journal of the American Veterinary Medical Association 230, 11; 10.2460/javma.230.11.1657

• Download Figure
• Download figure as PowerPoint slide

Map of the lower peninsula of Michigan illustrating the geographic regions in a study of the prevalence of antibodies against 6 Leptospira serovars in 1,241 healthy dogs in 1999. Values indicate overall percentages and numbers of dogs with positive titers against any of the 6 serovars.

Citation: Journal of the American Veterinary Medical Association 230, 11; 10.2460/javma.230.11.1657

• Download Figure
• Download figure as PowerPoint slide

Selection of participants—Dogs judged as healthy on the basis of results of a physical examination and history were candidates for the study. Samples were acquired while veterinarians were collecting blood for clinical diagnostic tests (eg, routine preanesthetic testing). Dogs with a history of leptospirosis or current liver, renal, or other systemic disease were excluded from the study. Because maternal antibodies may interfere with interpretation of results, all participants were at least 4 months of age.

Sample collection and testing—Veterinarians collected 3 mL of blood in a serum separator tube. Serum was harvested, transferred to a 3-mL clot tube, and stored at −20°C. Samples were stored at Michigan State University until Leptospira antibody titers were measured by use of an MAT. Serum samples were analyzed for agglutinating antibodies against 6 Leptospira serovars: Bratislava, Canicola, Grippotyphosa, Hardjo, Icterohaemorrhagiae, and Pomona. Two-fold dilutions of serum from 1:50 to 1:6,400 were tested with the MAT,³ and the titer was recorded as the reciprocal of the highest dilution of serum that agglutinated ≥ 50% of the leptospires.

Definition of a positive titer—A positive titer in vaccinated dogs was defined as a positive reaction at a dilution of ≥ 200 for serovars Bratislava, Grippotyphosa, Hardjo, and Pomona; ≥ 1,600 for Icterohaemorrhagiae; and ≥ 3,200 for Canicola. A positive titer in unvaccinated dogs was defined as a positive reaction at a dilution of ≥ 200 for serovars Bratislava, Canicola, Grippotyphosa, Hardjo, Icterohaemorrhagiae, and Pomona. A dog could be classified as having a positive titer to > 1 serovar. Among seropositive dogs, the serovar with the highest titer was considered to be the serovar to which the dog was likely exposed. Dogs in which the highest titer was to more than 1 serovar were classified as likely having been exposed to each of those serovars.

Risk factors for positive Leptospira titers—To identify risk factors associated with a positive antibody titer to 1 or more Leptospira serovars, participating owners completed a demographic questionnaire about their dog. Information collected included sex (sexually intact female, spayed female, sexually intact male, or castrated male), age, breed, location of residence (northern, central, southeastern, or south central-western), housing (indoors, outdoors, or both), degree of restraint when outdoors (leashed, fenced, or free roaming), travel outside of Michigan, exposure to other animals (dogs, cats, wildlife, or domestic livestock), exposure to natural water sources (creeks, ponds, or lakes), swimming, and vaccination status (date and type of vaccines and whether Leptospira bacterin was in the vaccine).

Statistical analysis—All statistical analyses were performed on data from the 309 seropositive dogs and the serovars to which they were likely exposed. Test prevalence of seropositive dogs and the serovar or serovars to which they were likely exposed was computed for each serovar of interest for different subject groups (all dogs combined, dogs vaccinated with Leptospira bacterin, and unvaccinated dogs) by geographic region by use of the following equation:

View Expanded

Univariable and multivariable analyses were conducted to assess the association between risk factors of interest and the serovars of likely exposure. The Mantel-Haenszel χ² test was used to assess the association of categoric risk factors (eg, geographic region and sex) with positive titers, and ANOVA was used to test associations between serovar of likely exposure and the continuous risk factor of age. Multivariable logistic regression models were used to simultaneously assess the associations of various risk factors on the probability of a dog being exposed to a specific leptospiral serovar. The multivariable models were generated by use of a hierarchical backwards model building procedure. Risk factors with P b 0.25 in univariable analysis were considered for inclusion in the full multivariable model. Interaction terms were developed on the basis of biological plausibility and the results of descriptive statistics. Risk factors were selected for elimination by removal of the risk factor with the highest P value. The reduced model was executed, and confounding was assessed by computing odds ratios for risk factors that remained in the model. If odds ratios of any of the remaining risk factors changed > 10%, the removed variable was considered to be a confounder for the remaining risk factors and was retained in the model. The model with the best combination of lowest model deviance (as assessed by comparing model Akaike information criterion before and after removing a variable) and covariate P values (P < 0.05) was considered to be the final reduced form of the multivariable model.

Results

A total of 736 small animal practices responded to enrollment letters, and 150 practices were randomly selected for sampling. Serum samples from 1,379 dogs were collected, for which 1,241 had complete demographic data and were included in the study. Serum samples from dogs with complete risk factor data were tested. The distribution of dogs with various titers against the 6 serovars was determined (Table 1). Three hundred nine dogs had 463 positive titers (certain dogs had 2 or more positive titers).

Results of univariable analyses—The serovars that were most frequently associated with likely exposure of seropositive dogs, Bratislava and Grippotyphosa, were serovars that were not present in canine vaccines during the study period (Table 2). Vaccination was associated with greater prevalence of likely exposure to serovars Bratislava and Grippotyphosa (P ≤ 0.05). Of positive dogs, only 9.1% of unvaccinated dogs and 9.5% of vaccinated dogs had likely exposure to > 1 serovar. Older age was associated with likely exposure to serovars Grippotyphosa and Icterohaemorrhagiae, although the number of dogs with likely exposure to serovar Icterohaemorrhagiae was small (Table 3). Significantly fewer dogs in the northern region had likely exposure to serovar Bratislava, compared with other regions (Table 4). No other geographic associations were significant. No significant associations were found between likely exposure to Leptospira spp and breed. Although associations were not significant, the breed groups with the highest likelihood of exposure were terriers for all serovars and for serovars Bratislava and Icterohaemorrhagiae, toy breeds for serovar Canicola, and working breeds for Grippotyphosa and Pomona (Table 5).

Results of multivariable analyses—Because of the low numbers of positive results for serovar Hardjo (n = 2), multivariable logistic regression models were only attempted and successfully developed for the remaining 5 serovars (Table 6). Dogs (all and vaccinated) living in the northern region of the state were at decreased risk for likely exposure to serovar Bratislava, and unvaccinated dogs with high livestock exposure were at higher risk, as were all dogs with wildlife exposure. The risk factor for likely exposure to serovar Grippotyphosa was age (all dogs and unvaccinated). Unvaccinated dogs living in the south central-western region were at decreased risk for likely exposure to serovar Grippotyphosa. Exposure to livestock and travel outside Michigan were associated with increased risk for likely exposure to serovar Pomona for all dogs and vaccinated dogs.

Discussion

The study population was chosen to minimize bias towards clinically affected dogs that were more likely to be seropositive for leptospiral antibodies. The fact that the participating dogs were chosen randomly and that all participating dogs were healthy minimized sampling bias. Because leptospiral vaccination status was known and all samples were collected prior to use of vaccines containing serovars Grippotyphosa and Pomona, confounding associated with vaccination was also minimized. Because dogs were assessed for positive titers independent of disease, positive titers likely represented exposure to Leptospira spp.

The MAT is the most commonly used diagnostic test for detecting leptospiral antibodies and is relatively serovar specific, although dogs with positive titers generally have sera that cross-react to some extent with other serovars.^1,5,6 The serovar with the highest titer is interpreted as the infecting serovar, and lower titers are interpreted as antibody cross-reactivity between serovars. The MAT does not differentiate between antibodies induced by vaccination and those produced because of natural exposure. Antibody titers associated with vaccination are typically < 400 and measurable for < 4 months after vaccination, although there is considerable variation in the titers detected in individual vaccinated dogs.^5,8

The definition of positive titers in leptospiral serosurveys varies widely and ranges from 10 to 800, depending on the nature of the population surveyed and the purposes of the study.^{18,19,24,26–28} Many surveys have been done in dogs with unknown vaccination status, which further complicates interpretation of titers. In this study, we had definitive information regarding the vaccination history of the dogs, which afforded the opportunity to use different cutoff titers for dogs that had been vaccinated and those that had not. A cutoff titer of ≥ 100 is commonly used for nonvaccinal serovars. In this study, a more conservative cutoff of ≥ 200 was used to minimize the effect of cross-reacting antibodies. However, in seroprevalence surveys among vaccinated dogs, a higher cutoff for serovars contained in the vaccines is warranted to minimize the overestimation of seroprevalence in the vaccinated population. Therefore, in this study, the cutoff titers for vaccinal serovars (Canicola and Ictero-haemorrhagiae) were different for vaccinated and nonvaccinated dogs and were chosen in consideration of the ranges seen in healthy vaccinated dogs at the Diagnostic Center for Population and Animal Health, where > 15,000 dog sera are tested for leptospiral antibody annually. The distribution of positive titers among vaccinated and nonvaccinated dogs in the present study supports the contention that these cutoff titers were sufficiently high to minimize the effect of vaccination on the seroprevalence of Canicola and Icterohaemorrhagiae reported here. A cutoff titer of ≥ 200 was used for serovars Canicola and Icterohaemorrhagiae in nonvaccinated dogs. Because the titer cutoffs chosen for this study were more conservative than are typically used when assessing for Leptospira exposure, interpreting vaccine response as actual exposure was minimized. The overall prevalence of 24.9% for seropositive dogs determined in this study was more likely an underestimate than an overestimate of actual exposure.

There is some controversy regarding the role of serovar Bratislava as an important pathogen for dogs in the United States. An experimental study²² involving exposure of young Beagles to an isolate of serovar Bratislava did not result in infection of the dogs. However, the authors of that study concluded that the particular isolate they used may have lost virulence during in vitro culture and that the results of the study with a single isolate should not be interpreted to indicate that serovar Bratislava is not a canine pathogen. Evidence to support the potential of serovar Bratislava as a pathogen of dogs includes evidence that serovar Bratislava is present in animal populations in the United States²⁹ and clinical reports from the United States^16,30 and United Kingdom^19,31 that indicate that serovar Bratislava is infectious for dogs and can be associated with clinical disease.

Serovar Autumnalis was not included in this study, although positive canine titers have been reported.²³ Interpretation of serovar Autumnalis titers is not fully understood at this time because of erroneous or paradoxical high reactivity.^6,15,23 In 1 study,⁶ dogs vaccinated with only serovars Grippotyphosa and Pomona developed titers to serovar Autumnalis that were significantly higher than the titers to either vaccinal serovar. The sera collected for the present study, which were collected prior to the availability of vaccines containing serovar Grippotyphosa and Pomona, will be useful in determining the prevalence of titers to serovar Autumnalis in a nonvaccinated population.

Canine Leptospira serologic surveys from areas corresponding to the southeastern and south central-western regions in Michigan have been reported. In association with 75 cases of human leptospirosis in Detroit reported in 1939, 16% of examined rats were seropositive for leptospiral infections and 10 of 13 dogs had positive results of MAT against serovar Icterohaemorrhagiae but lacked titers against serovar Canicola.³² Results of a substantial leptospiral serologic survey³³ of Detroit-area dogs were also reported in 1980. Dog sera were tested by use of the MAT for antibodies against serovars Canicola, Grippotyphosa, Icterohaemorrhagiae, and Pomona. Of 433 stray dogs, 23.8% were seropositive against serovar Icterohaemorrhagiae, followed by serovars Canicola (8.3%), Grippotyphosa (6.9%), and Pomona (6.7%), whereas seropositvity among suburban dogs was 16.3% for Icterohaemorrhagiae, 4.0% for Canicola, and 1.6% for Pomona. No dogs were seropositive against serovar Grippotyphosa. In contrast, in the present study, 15.9% of dogs in the southeastern region had likely exposure to serovar Grippotyphosa, with no dogs likely exposed to serovar Icterohaemorrhagiae, but overall likely exposure in the southeastern region (25.7%) was similar to that reported in 1980. The marked decrease in likely exposure to serovar Icterohaemorrhagiae may reflect use of Icterohaemorrhagiae bacterins and, probably more importantly, better Norway rat control as a result of curbside garbage pickup.^a The increased likely exposure to serovar Grippotyphosa may reflect increased exposure to maintenance hosts for Grippotyphosa (raccoons, opossums, and skunks) as suburban expansion occurs.

Of 500 healthy dogs from Lansing, Mich, tested for antibodies against 2 serovars in 1950, 26.8% were seropositive against serovar Canicola and 2.6% against serovar Icterohaemorrhagiae.³⁴ The author noted that dogs in the study were unlikely to have had exposure to rats. Lansing is located in the south central-western region of the present study where likely exposure to serovar Canicola was 2.8% and likely exposure to serovar Icterohaemorrhagiae was 1.8%, but 16.7% of dogs were seropositive against Grippotyphosa. The decrease in prevalence of likely exposure to serovar Canicola may reflect use of Canicola bacterins, whereas the nearly equivalent overall exposure rate was a reflection of increased exposure to serovars Grippotyphosa and Bratislava, perhaps as a result of suburban encroachment on wildlife.

Serovar prevalence results for all dogs from the 4 Michigan study regions can be compared with published serovar reports for dogs with leptospirosis and with reports of dog sera submitted to diagnostic laboratories presumably because clinical leptospirosis was a diagnostic consideration. In most leptospirosis case report series published in the preceding 15 years, Grippotyphosa was the predominant serovar.^{12–15,17,18} An exception is the only report¹⁶ originating from western North America. In that report of California dogs with clinical leptospirosis, serovars Bratislava and Pomona predominated and only 2 dogs had any titer to serovar Grippotyphosa. In contrast, published reports from midwestern and eastern North America all found grippotyphosa to be the predominant or nearly predominant infecting serovar. It is possible that this difference reflects geographic maintenance-host differences as supported by a report³⁵ of low serovar Grippotyphosa seropositivity among California raccoons, in contrast to high Grippotyphosa seropositivity among raccoons east of the Mississippi River.^36–38 Similarly, recently published seropositivity reports^24,25 based on submissions to diagnostic laboratories indicate that serovar Grippotyphosa titers predominate.

Risk factors identified in the present study of healthy dogs were both similar and dissimilar to some of the risk factors identified in dog populations not known to be healthy. Age was significantly associated with increased likely exposure to serovars Grippotyphosa and Icterohaemorrhagiae regardless of vaccination status, and this was previously reported in a retrospective study²⁵ of dogs with leptospirosis caused by unspecified serovars. Age may have allowed greater cumulative exposure to maintenance hosts for serovars Grippotyphosa and Icterohaemorrhagiae. The finding that unvaccinated healthy dogs living in the south central-western region of Michigan were at reduced risk for likely exposure to serovar Grippotyphosa cannot be easily explained but may be related to complex factors of exposure to maintenance hosts and unidentified environmental influences. For example, serovar Grippotyphosa–infected raccoons may have higher population densities in urban than in suburban areas.³⁹

Risk factors were also identified for serovar Bratislava exposure. Dogs living in the northern and central regions were at reduced risk, possibly because reservoir hosts such as rats, pigs, and horses have lower population densities in those regions. Unvaccinated dogs with high livestock exposure were at highest risk for serovar Bratislava exposure. The overall high risk for serovar Bratislava exposure associated with wildlife exposure remains unexplained.

Travel outside Michigan and exposure to livestock were risk factors for likely exposure to serovar Pomona for all dogs and for vaccinated dogs. It is possible that this finding reflected increased exposure to cattle, although other reservoir hosts such as pigs, skunks, and opossums cannot be ruled out.

Leptospiral vaccination reduced the risk of dogs being seropositive to serovars Canicola and Icterohaemorrhagiae, the 2 serovars included in commercially available bacterins at the time of the study. Because these 2 serovars were in the vaccine that was in use during the period the study was conducted, findings support the notion that the vaccine was effective in protecting dogs from developing positive titers against these serovars. Additionally, findings support the cutoffs used in this study as not classifying titers in a false-positive manner. Exercise in a fenced yard was associated with reduced risk of likely exposure to both serovars, whereas exercising leashed or free was associated with increased risk of likely exposure to serovar Icterohemorrhagiae in unvaccinated dogs. Vaccinated dogs located in the south central-western region were at increased risk for likely exposure to serovar Canicola, a finding not easily explained in light of the lack of similar findings for other serovars. Exposure to surface water (eg, lakes, ponds, rivers, and streams) was not associated with likely exposure to serovar Canicola; exposure to water has been identified both as a risk factor and as not associated with leptospirosis.^40,41

In addition to geographic location, degree of restraint when exercised, travel outside Michigan, and vaccination status, other potential risk factors including sex, breed, and housing were evaluated. None of these factors were significantly associated with risk in this population of healthy dogs. In previous reports²⁵ of dogs with leptospirosis or dogs tested because leptospirosis was a diagnostic consideration, hounds, herding dogs, working dogs, and gun dogs were at increased risk.

Results of the present study indicated that > 20% of healthy, client-owned dogs from the lower peninsula of Michigan likely had been exposed to Leptospira serovars that were previously uncommon. In addition, results indicated that exposure to Leptospira was widespread and was not associated with breed as has been suggested. Because serovar Grippotyphosa, a serovar maintained by wildlife, was the predominant serovar identified, certain strategies can be used to either mitigate or minimize risk. Administration of vaccine containing leptospiral serovar Grippotyphosa and limiting exposure to wildlife reservoirs may be appropriate considerations.