Materials and Methods

Study sample—Results of all Leptospira MATs of canine serum samples from January 1, 2000, through December 31, 2007, were obtained electronically from a commercial veterinary diagnostic laboratory^a providing animal disease diagnostic services in the United States. All MATs were conducted at 1 facility, regardless of sample origin. The MAT was performed by use of 2-fold serial dilutions of serum samples, starting with a dilution of 1:100, for detection of each of 7 serovars (Canicola, Grippotyphosa, Pomona, Bratislava, Icterohemorrhagiae, Hardjo, and Autumnalis). The MAT result for each serovar was reported as the highest dilution of serum at which at least 50% of the antigens were agglutinated, compared with the control suspension.

Data extraction—A subset of the original data set from the laboratory was prepared to extract information for dogs only. Data were extracted for dog age, sex, and breed as well as postal code of the hospital of origin, date of test request, and MAT titer results for all assayed serovars. Sex was entered by the laboratory as male or female; neuter status could not be determined. Breed information was entered as free text by veterinary clinics, resulting in several different entry formats for each breed. All entries corresponding to a particular breed were grouped together and identified with a standard breed name. Entries for breed with a frequency of < 10 were not grouped into a specific breed but, rather, were grouped together under the category of not classified.

A serum sample with an MAT titer ≥ 1:1,600 for at least 1 of the 7 serovars was considered positive. This cutoff was selected to reduce the likelihood of identifying vaccine-induced antibodies as those resulting from natural infection.^14,15

Statistical analysis—Seroprevalence in this study was defined as the proportion of submitted samples with positive MAT results as identified with the defined cutoff titer. Overall annual and serovar-specific seroprevalences with 95% CIs were calculated. Computation of serovar-specific seroprevalence was performed by dividing the number of tests with positive results for any given serovar by the total number of tests performed for that serovar. A linear change in annual seroprevalence was assessed by use of a simple weighted regression model. In the model, annual seroprevalence was the dependent variable, year (2000 through 2007) was the independent variable, and total number of serum samples submitted in a year was the weighting variable.

The Spearman correlation coefficient (R) for all 2-way comparisons among serovar results was calculated by rank ordering the MAT dilutions for each serovar from 1 to 9, starting with 1:100 as 1 and ending with 1:12,800 as 9. When 2 or more serovars had the same positive MAT titer in a serum sample from 1 dog, all serovars with equal titers were assigned the same rank score appropriate for the titer. Those serum samples that had negative results for all serovars at a titer of 1:100 were excluded from correlation calculations, because in that situation, all serovars would have been scored as 0. All statistical computations were performed by use of a commercial statistical software program.^b Values of P < 0.05 were considered significant.

Results

Serum samples—A total of 33,119 serum samples were submitted for Leptospira MAT testing during the 8-year study period. The number of canine serum samples received for testing increased steadily each year from 1,528 in the first year (2000) to 8,189 in the final year (2007). Samples were submitted from 3,224 geographic areas (postal codes) representing all continental states. Age was available for 32,856 (99.2%) of the dogs for which serum samples were submitted. The distribution of samples submitted for testing did not differ significantly across age strata. Dog breed data were available for 24,444 (73.8%) submissions.

More than 31,000 tests were performed for each of 5 Leptospira serovars (Canicola, Icterohemorrhagiae, Pomona, Bratislava, and Grippotyphosa). Testing for serovar Autumnalis was initiated at the beginning of 2006, and 13,814 tests were performed for this serovar. The total number of MATs performed for serovar Hardjo was 19,247; testing was discontinued early in 2006.

Seroprevalence of leptospirosis—Of the 33,199 submitted serum samples, 2,680 (8.1%; 95% CI, 7.8% to 8.4%) had positive MAT results (Table 1). The number of submissions and number of seropositive dogs varied geographically within the continental United States (Figure 1). Among the serum samples with positive MAT results, 2,110 (78.7%) had the highest titer against 1 serovar and 570 (21.3%) had the highest titer against > 1 serovar. Among the 2,110 serum samples with the highest titer against a single serovar, Grippotyphosa was the most common, followed by Pomona and Autumnalis (Table 2).

Geographic distribution by postal code of positive results (titer ≥ 1:1,600) of MAT for antibodies against Leptospira serovars in canine serum samples (n = 33,119) submitted from various regions of the United States to a commercial laboratory from 2000 through 2007. Samples were not received from unshaded regions.

Citation: Journal of the American Veterinary Medical Association 237, 3; 10.2460/javma.237.3.293

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Geographic distribution by postal code of positive results (titer ≥ 1:1,600) of MAT for antibodies against Leptospira serovars in canine serum samples (n = 33,119) submitted from various regions of the United States to a commercial laboratory from 2000 through 2007. Samples were not received from unshaded regions.

Citation: Journal of the American Veterinary Medical Association 237, 3; 10.2460/javma.237.3.293

• Download Figure
• Download figure as PowerPoint slide

Geographic distribution by postal code of positive results (titer ≥ 1:1,600) of MAT for antibodies against Leptospira serovars in canine serum samples (n = 33,119) submitted from various regions of the United States to a commercial laboratory from 2000 through 2007. Samples were not received from unshaded regions.

Citation: Journal of the American Veterinary Medical Association 237, 3; 10.2460/javma.237.3.293

• Download Figure
• Download figure as PowerPoint slide

The highest percentage of positive MAT results (4.9%; 95% CI, 4.6% to 5.3%) was for serum antibodies against serovar Autumnalis, followed by serovar Grippotyphosa (4.3%; 95% CI, 4.1% to 4.5%; Table 3). Serovars Pomona and Bratislava were the third and fourth most common, with seroprevalences of 3.8% (95% CI, 3.6% to 4.0%) and 2.7% (95% CI, 2.5% to 2.9%), respectively, in submitted serum samples.

The highest correlations between seroprevalences of the various serovars were for Pomona and Grippotyphosa (R = 0.47), Pomona and Bratislava (R = 0.45), and Grippotyphosa and Bratislava (R = 0.31). Correlations of positive MAT results for Autumnalis with those of the aforementioned serovars were generally weaker; all of the aforementioned rank correlations were significant (P < 0.001). Although a weakly positive correlation (R = 0.25) was evident between seroprevalences of Canicola and Icterohemorrhagiae, these serovars had virtually no correlation with any other serovar (all R < 0.10).

The highest annual seroprevalence of 10.2% (95% CI, 9.6% to 10.9%) was observed for the year 2007, followed by the year 2004 with 10.1% positive (95% CI, 9.2% to 11.1%). A linear trend in annual seroprevalence for canine leptospirosis was not evident for the study period. The highest percentage of positive MAT results was for the months of November and December, with occasional highs also evident during early summer months in some years (Figure 2).

Monthly variation in positive results (titer ≥ 1:1,600) of MAT for antibodies against Leptospira serovars in canine serum samples (n = 33,119) submitted from various regions of the United States to a commercial laboratory from 2000 through 2007.

Citation: Journal of the American Veterinary Medical Association 237, 3; 10.2460/javma.237.3.293

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Monthly variation in positive results (titer ≥ 1:1,600) of MAT for antibodies against Leptospira serovars in canine serum samples (n = 33,119) submitted from various regions of the United States to a commercial laboratory from 2000 through 2007.

Citation: Journal of the American Veterinary Medical Association 237, 3; 10.2460/javma.237.3.293

• Download Figure
• Download figure as PowerPoint slide

Monthly variation in positive results (titer ≥ 1:1,600) of MAT for antibodies against Leptospira serovars in canine serum samples (n = 33,119) submitted from various regions of the United States to a commercial laboratory from 2000 through 2007.

Citation: Journal of the American Veterinary Medical Association 237, 3; 10.2460/javma.237.3.293

• Download Figure
• Download figure as PowerPoint slide

The percentage of positive test results was highest for dogs > 2 to b� 6 years of age, whereas the lowest percentage was for dogs > 10 years of age (Table 4). The seroprevalence of leptospirosis in samples from mixed-breed dogs was 9.2% (95% CI, 8.5% to 10.1%). The proportion of positive tests for 19 specific breeds ranged from 2.9% to 15.3%, with both small and large breeds represented in proportions that were greater than and less than that of mixed-breed dogs.

Discussion

In the study reported here, we evaluated MAT results for canine serum samples submitted to a commercial diagnostic laboratory by practicing veterinarians. Submission bias was presumed to exist in that clinicians would be expected to submit samples out of concern of leptospiral infection in a patient. This submission bias may have been further influenced by a clinician's suspicion or lack of suspicion of infection when considering patient factors such as size and breed. Despite this bias, analyses conducted with this type of data may still yield informative results on the basis of the absolute number of submitted samples, the percentage of those samples that had positive results, or the relative distribution of seroprevalence within subgroups of interest.

A positive MAT result for leptospirosis in dogs can be indicative of the presence of circulating antibodies against Leptospira spp from current natural infection, past exposure to leptospires, or recent vaccination against Leptospira spp.⁴ High reciprocal antibody titers are usually associated with acute leptospirosis from recent natural infection,⁴ as postvaccinal antibody titers are often > 1:1,600 and decrease rapidly.^14–16 Therefore, a conservative cutoff titer of ≥ 1:1,600 was used to minimize the possibility of misclassification of vaccine response as natural infection. It was assumed that canine serum samples were predominantly submitted by veterinarians for testing because of clinical suspicion of leptospiral infection. However, the lack of information regarding clinical signs and vaccination status is a limitation of leptospiral studies in which only laboratory-based records are used. The percentage of positive MAT results reported here may have been an underor overestimation of the prevalence of clinical disease, but comparative aspects regarding serovars and dog signalment are informative regarding the changing epidemiology of this reemerging disease.

Among the dogs with a single highest reacting titer, Grippotyphosa, Pomona, and Autumnalis were the most common Leptospira serovars, which is consistent with findings of other researchers.⁹ Serovar Autumnalis was the most commonly reacting serovar when the percentage of positive MAT results for a serovar was computed as the proportion of the total number of tests performed for that serovar. It is difficult to accurately interpret the results for serovar Autumnalis because it may purely indicate cross-reactivity or paradoxical reactions with other serovars.^14,17 In our clinical experience, titers against Autumnalis may be high on an initial test, indicating leptospiral infection, but may decrease and another (presumably the true infective) serovar may increase up to 4-fold on paired titer analysis. To the authors' knowledge, Leptospira Autumnalis has not been isolated from dogs in the United States in the last 50 years, and its detection by the MAT (other than an early antibody response to another serovar) may not mean it was the infective serovar. Similarly, there are limited reports of canine infections with Leptospira Bratislava in the United States,¹⁸ where the reservoir host has not yet been identified; however, horses are probably a maintenance host for Bratislava in other countries.¹⁹

Reported cross-reactions between various Leptospira serovars, coupled with the correlation results in the present study, suggest molecular mimicry between serovars such that a host's IgM or IgG antibodies will agglutinate leptospires of multiple serovars. Leptospira Pomona was associated with the highest correlations between anti-serovar antibody titers in our study, although Autumnalis was the most highly correlated serovar in another study.⁹ One explanation could be a difference in the antigenic variations of laboratory-maintained leptospires used to perform MATs, as these results are based on records from 2 commercial laboratories. The positive correlation detected between selected serovars here suggested the possibility of gaining some cross-protection from vaccination given that some antigenic proteins are highly conserved in pathogenic leptospires.^20,21 Existence of cross-protection between serovars Canicola and Icterohemorrhagiae has been reported, mainly related to nonlipopolyosidic antigens of the leptospires,²² but cross-protection with other serovars has not been reported. Serovars Canicola or Icterohemorrhagiae were seldom considered the infective serovar (< 5%) when the cutoff of ≥ 1:1,600 was used in the present study but often yielded low titers when other serovars were considered infective. These low titers likely contributed to the correlation between the seroprevalences of these 2 serovars. Dogs were rarely seropositive against the Hardjo serovar in both this and a previous study.⁹

Marked increases in percentages of serum samples with positive MAT results were noticeably evident in November, December, or both for several years in the study. A seasonal (fall) predilection for this disease has been reported, with suggested correlation between the number of cases of leptospirosis in dogs and the amount of rainfall.^6,23 Increased seroprevalence in spring (eg, March or May) may be related to increased rainfall, but rainfall data were not assessed in our study. An association of disease with sporting or hunting dogs and season would correspond also with open hunting periods in the fall and potential exposure to wildlife and their environment.⁶ However, evidence that urban or suburban dogs are also at risk,¹³ the seropositive rates in both small-and large-breed dogs in the present study, and recognition that November is not a period of great monthly rainfall in much of the United States suggest other factors may be important in explaining this aspect of the seasonal pattern. Urban or suburban dogs may be exposed to leptospires shed by periurban wildlife such as raccoons, which are known reservoir hosts for leptospirosis,^24,25 and the young of such wildlife may disperse in the fall when seeking new homes for winter.²⁶ Thus, increased seasonal activity of a reservoir host may explain seasonal exposure increases in incidental hosts. The number of submissions and proportions of seropositive dogs varied across the continental United States, and it is expected that the risk of leptospirosis also varies geographically because of varied infection rates and population densities of reservoir hosts.

Results of the present study indicated the annual Leptospira seroprevalence in dogs varied slightly during the study period, suggesting a stable background risk of exposure for leptospirosis in dogs may exist. Annual fluctuations in seroprevalence might have been attributable to differences in other environmental risk factors such as the amount of rain, flooding, outdoor activities, and contact of dogs with reservoir hosts. Trend analysis for seroprevalence did not indicate a significant linear pattern in the percent of positive tests for canine leptospirosis during 2000 through 2007 in the United States, and it is possible that the increasing trend that was detected in a previous study⁹ has stabilized.

Leptospires are unlikely to have particularly high affinity for any particular dog breed, and although dogs of some breeds may have increased outdoor activity, difference in seroprevalence among breeds may also reflect submission biases by veterinarians. Perhaps also indicative of sampling bias was our finding that some sporting breeds such as Labrador and Golden Retrievers had a lower seroprevalence than mixed-breed dogs. Although differing from earlier published work,⁶ this finding is consistent with a recent investigation in Canada.²⁷ Veterinarians might have had a higher index of suspicion of leptospirosis in sporting or large breeds, leading to a greater number of submitted serum samples from these breeds for testing. These breeds also may have been more likely to have been vaccinated and protected against leptospirosis than other breeds. Although fewer serum samples were generally submitted for small-breed dogs, the percentage of positive test results when the study titer of ≥ 1:1,600 was used indicated an exposure risk for these dogs.

The distribution of submitted serum samples between male and female dogs in the present study was approximately equal, as were the percentages of positive test results for each sex (7% to 9%). Age was identified as a risk factor for leptospirosis in dogs in another study,⁶ suggesting that dogs between 4 and 10 years of age are at increased risk of disease, compared with dogs < 1 year of age. Although the seroprevalence of leptospirosis in our study was highest for dogs 2 to 6 years of age, the distribution of submitted serum samples across the 2-year incremental age strata was fairly homogeneous, as was the percentage of positive MAT results (8% to 10%). The lower percentage of positive test results in dogs > 10 years of age (5%) was consistent with other reported findings,⁶ although it is unknown whether this reduction in apparent disease risk is attributable to lower exposure risk, possible previous vaccination, or other factors.

Practitioners should consider leptospirosis in their differential diagnoses based on appropriate clinical signs, rather than ruling out the disease because of a dog's breed, size, or age or the season of year. Identification of the infective Leptospira serovar allows a veterinarian to consider, and inform the owner of, possible reservoir host or hosts and the potential for immunoprophylaxis against that serovar. Given the serovars evaluated in our study and their potential for crossreactions in the MAT, bacterial culture of urine or blood samples is highly recommended for identification of the infective serovar.