Risk Factors Associated with Leptospirosis in Dogs from Northern California: 2001–2010

Introduction

Leptospirosis is an infectious disease of humans and other animals resulting in potentially life-threatening illness caused by the zoonotic spirochetes Leptospira interrogans and L. kirschneri (Farr 1995, Levett 2001, Galloway et al. 2009, Levett and Haake 2010). Pathogenic leptospires have a worldwide distribution, persist in soil and wet climates for months, and have been described as the most common zoonotic agents in the world (Palaniappan et al. 2007, Levett and Haake 2010). A large number of strains and serovars can infect many different hosts, typically in host-affiliated but interconnected epidemiological cycles affecting humans, dogs, livestock, wildlife, and even marine mammals (Colagross-Schouten et al. 2002).

In dogs, clinical manifestations of leptospirosis range from subclinical infection to multiple-organ failure and death, most commonly from acute renal failure (Greene et al. 2006, Sykes et al. 2011). Serovars Canicola and Icterohemorrhagiae were historically the most common serovars thought to infect dogs; however, since introduction of bivalent vaccines containing these serovars, most dogs with leptospirosis at least in North America, seroconvert predominantly to serovars Pomona, Grippotyphosa, and Bratislava (Goldstein et al. 2006). Previous assessments of risk factors for canine leptospirosis in northern California have documented associations with exposure to water and periods of heavy rainfall (Ward 2002, Ward et al. 2004, Ghneim et al. 2007). Increases in the incidence of canine leptospirosis at one hospital in northern California followed above-average annual rainfall, particularly after an approximately 3-month lag (Ward 2002, Ward et al. 2004). In the present study, we evaluate risk factors for canine leptospirosis in northern California over a modern timeframe.

Materials and Methods

Data were extracted electronically from medical records at the University of California, Davis William R. Pritchard Veterinary Medical Teaching Hospital (VMTH) of dogs diagnosed with leptospirosis between March, 2001, and November, 2010. An electronic search of the VMTH database was performed using the terms “leptospir*”, “acute renal failure”, “acute hepatic failure”, “kidney failure”, “liver failure”, and “pulmonary hemorrhage”. A case was defined as a dog with a home address in northern California that had clinical and clinicopathologic abnormalities consistent with acute renal failure and/or hepatic failure with at least one positive microscopic agglutination test (MAT) serology result with a reciprocal titer ≥800 or positive Leptospira immunohistochemistry (IHC) in kidney tissue. Four controls from the same time period were randomly selected from the VMTH database for every leptospirosis case using a random number generator (www.random.org). Inclusion criteria for the control group were dogs with no clinicopathologic data consistent with leptospirosis, such as azotemia or increases in liver enzyme activities. Matching of cases and controls was not performed.

Data that were extracted from case and control records included signalment, date of initial examination, home address with five-digit zip code, history of vaccination for leptospirosis, potential exposure history to Leptospira (including access to wildlife and rodents, a yard, various water sources and domestic farm animals, and a history of travel or hunting), clinical signs at the time of initial examination, and the results of hematological and serum biochemistry testing. Where available, acute and convalescent Leptospira serology results and outcome (defined as survival or fatality) were recorded as well. Leptospira serology was performed using a MAT at the California Animal Health and Food Safety Laboratory (Davis, CA) using serovars Pomona, Grippotyphosa, Bratislava, Canicola, Icterohemorrhagiae, and Hardjo as test antigens. This laboratory voluntarily participates in the National Veterinary Service Laboratory diagnostic proficiency scheme.

The temporal and spatial distributions of leptospirosis among cases were reported previously (Hennebelle et al. 2013), comparing annual prevalence with rainfall. For all dogs, demographic factors, including age, breed, and gender, geographical region of home zip code, and vaccination history for leptospirosis, were evaluated to identify differences between the case and control groups. Age was determined by year of birth and divided into the categories less than 1 year old, 1–5 years old, 5–10 years old, and older than 10 years. Breeds were grouped based on the American Kennel Club classification as herding dogs, hound dogs, mixed-breed dogs, nonsporting dogs, sporting dogs, terrier dogs, toy dogs, and working dogs. Regions were divided into Modoc and the Cascade Mountains, south and central coasts, the Sierra Nevada Foothills, Mojave and Tehachapi, north coast, Sierra Nevada Mountains, San Francisco Bay Area, and Central Valley. Vaccine status was defined as leptospirosis vaccine administered within 12 months prior to evaluation, vaccine administered >12 months prior, leptospirosis vaccine not administered, or vaccination history unknown.

Data were maintained in Excel (Microsoft, Redmond, WA) and analyzed with the R statistical package (R Development Core Team, www.r-project.org) including calculating odds ratios (ORs) with 95% confidence intervals (CIs). For all tests, statistical significance was inferred if p≤0.05. Hematological and serum biochemical test results as well as cost incurred on the first visit to the VMTH were compared among cases and controls using two-tailed Student t-tests. Differences in the number of dogs that survived for cases and controls were assessed with a chi-squared contingency test. Age, breed, region, gender, and vaccination history among cases and controls were evaluated using logistic regression.

Results

There were 338 dogs enrolled in the study, comprising 67 dogs with leptospirosis (cases) and 271 controls. Clinical characteristics of leptospirosis in the case dogs were vomiting in 53 (79%) dogs, inappetence or anorexia in 51 (76%) dogs, and lethargy in 47 (70%) dogs (Table 1). Lameness, stiffness, or locomotory abnormalities were described in 20 (30%) dogs, and 14 (21%) dogs had at least one episode of fever. Less common signs were decreased water intake, polyuria, polydipsia, decreased urination, and apparent pain on abdominal palpation. All of these clinical signs were significantly more common in dogs with leptospirosis than in the controls with the exception of polyuria/polydipsia (Table 1). A synopsis of clinical chemistry values and white blood cell counts is given in Table 2. Cases had significantly higher mean white blood cell counts compared with controls as well as higher neutrophil and monocyte counts, and blood urea nitrogen, creatinine, and bilirubin concentrations. Platelet numbers were significantly lower in cases of leptospirosis, whereas liver enzymes showed no significant difference. Death occurred before animals could be released from the hospital in 13% (9/67) of the dogs with leptospirosis, compared to 8% (22/271) among controls; this difference was not statistically significant (p=0.26). The average amount charged for hospitalization among cases was $5459 and among controls was $2802; this difference was not statistically significant (p=0.19). However, the duration of hospitalization among cases was statistically different from controls (11 days vs. 3 days, p<0.001).

All 67 cases had positive Leptospira serology. Leptospira serology was generally not performed on control dog sera because testing had not been clinically indicated in these dogs. Forty-five (67.2%) of the case dogs had the highest titer to serovar Pomona, five (7.5%) had the highest titer to Bratislava, and three (4.5%) had the highest titer to Icterohemorrhagiae. The remaining cases had equally high titers to multiple serovars (Table 3). Two dogs were assessed for Leptospira antigen using IHC of renal tissue and both were positive. Among cases, 44 dogs had paired MAT titer results reported, and all 44 dogs had a four-fold change in titer to at least one Leptospira serovar.

Many of the case dogs had a history that suggested environmental exposure to Leptospira (Table 4). Access to a yard was reported for 21 (31.3%) dogs; exposure to wildlife (including rodents) for 17 (25.4%); travel, hiking, or hunting for 16 (23.9%); and exposure to water sources (e.g., pond, creek, puddles) reported in 14 (20.9%). Thirteen dogs lived on a ranch or farm. The risk factors with the greatest and significant odds ratios were exposure to wildlife (OR=23), living on a ranch or farm (OR=9), and exposure to water (OR=6). Vaccination for leptospirosis within a 12-month period prior to being seen at the VMTH was reported in eight (12%) of 67 cases (Table 5). Vaccination for leptospirosis more than 12 months prior to evaluation was reported in seven (10%) of cases. The vaccine type administered was known for five case dogs, all of which had received a bivalent Icterohaemorrhagiae and Canicola vaccine within the last year. Vaccination was not reported for the other case dogs.

ORs for the signalment and spatial variables of age, breed, gender, and region are shown in Table 5. Dogs with leptospirosis tended to be older dogs, with ORs of 3.2 for dogs in the 5- to 10-year age group and 2.8 in the >10-year age group. Hound breed (OR=4.6) and herding breeds (OR=3.1) were at greater risk for leptospirosis, but there was no effect due to gender. Cases were 7.3 times more likely to live in the central or south coast region, 4.5 times more likely to live in the foothills region, 4.2 times more likely to live in the San Francisco Bay region, and 2.8 times more likely to live in the north coast region of California than controls.

Discussion

Leptospirosis has been described as the most frequent zoonosis in the world, but most case descriptions in people originate from tropical areas. Yet the abundance of animal, particularly canine, cases in many nontropical areas suggests possible underrecognition of the true prevalence in many hosts, including humans. The epidemiology of leptospirosis may have changed over the last few decades due to changing climate, land use, activities of dogs such as hiking and hunting, and vaccination. Given such changes, regular re-examination of animal cases becomes an important tool for understanding the full spectrum of leptospirosis ecology. Dogs are an important part of this strategy because their risk factors are often shared with people, they appear to be highly susceptible to the disease, and their disease may reflect the diversity of serovars that could impact human and canine health in an ecosystem.

Previously, risk factors were reported for leptospirosis in a smaller number of dogs from northern California, reflecting primarily exposure to water and rain (Adin and Cowgill 2000, Ghneim et al. 2007). In this larger, recent study, we found 67 cases of leptospirosis that were diagnosed at the VMTH from 2001 through 2010, which is approximately seven cases/year, compared with approximately 11 dogs/year between 1998 and 2000 (Ghneim et al. 2007). The clinical signs we observed in dogs diagnosed with leptospirosis largely reflected renal failure and azotemia, including vomiting and lethargy (Sykes 2014). We did not report fever, which occurs primarily in early-stage disease, likely before dogs would have been seen at the VMTH. On hematology, dogs had leukocytosis and thrombocytopenia, thought possibly to be due to disseminated intravascular coagulation or immune-mediated mechanisms (Sykes 2014). Clinical chemistry in this study emphasized elevations in urea nitrogen, creatinine, and bilirubin. Such changes in clinical presentation, hematology, and clinical chemistry have been reported before for severely affected dogs, reflecting more kidney and less liver disease as has been reported (Sykes 2014).

Leptospirosis is an expensive disease, both in the incalculable toll it exacts on affected dogs and families and in actual costs of treatment. In the present case series, dogs with leptospirosis presented with vomiting, lethargy, and inappetence with leukocytosis and thrombocytopenia as well as evidence of renal, but not hepatic, failure. At the VMTH, 45% of dogs received hemodialysis treatment (Adin and Cowgill 2000), although this option is not often available outside of referral, tertiary care hospitals. Costs for management of leptospirosis at this hospital were on average more than $5000, as described previously (Adin and Cowgill 2000). Case dogs spent significantly more days in the hospital than controls. The case fatality rate at the VMTH (where hemodialysis is available) was 13%, compared with rates as high as 20% in some series of dogs and people (Birnbaum et al. 1998, Baumann and Fluckiger 2001, Levett 2001, World Health Organization 2001, Townsend et al. 2006, Ko et al. 2009). Previous work in people has shown that prophylaxis and prompt therapy for cases improve outcome and considerably reduce costs (Galloway et al. 2009).

In the present study, an antibody response to serovar Pomona predominated, as was reported beginning in 1990 (Adin and Cowgill 2000, Ghneim et al. 2007). We do not know the background rate of serovar Pomona seroreactivity in dogs without leptospirosis because serology was typically not performed for control dogs. Such a lack of disease-specific diagnostic test results is a common problem in retrospective studies. In part to address this problem and because serology for Leptospira is known to be positive in some vaccinated or healthy dogs, we required the case definition to include clinical disease as well as seroreactivity. While it would be helpful to know the serological status of control dogs, we nevertheless defend the use of this control group because it was required that controls be drawn from the same at-risk population. However, because our case definition required clinical disease, it is possible that some milder cases of leptospirosis were overlooked. Another option for case dogs would have been culture or molecular confirmation of Leptospira organisms or DNA, but this was not practical because of the retrospective nature of the study, with some cases dating to early availability of molecular diagnostic testing. Additionally, the majority of dogs, even with acute leptospirosis, are PCR and culture-negative in blood and urine, limiting the usefulness of these tests for a study such as performed here (Sykes 2014).

Sources of exposure to Leptospira are often water, wildlife, or livestock. In the present study, the greatest risks were associated with wildlife exposure and living on a farm or ranch. Residence on a farm or ranch may result in exposure to either wildlife or domestic farm animal reservoir hosts. However, suburban and even urban dogs may be exposed to periurban mesocarnivores such as raccoons that commonly host serovar Pomona but can be reservoirs for serovar Icterohemorrhagiae as well (Meyer 1939, Heath et al. 1965, Richardson and Gauthier 2003, Meites et al. 2004, Brod et al. 2005, Gouveia et al. 2008, Maciel et al. 2008, Ko et al. 2009, Jimenez-Coello et al. 2010). Most dogs with leptospirosis for which vaccine histories were available had inadequate vaccination, either because it was not recent, only included two serovars, or had not occurred.

Ward et al. reported that male dogs are at higher risk of leptospirosis than females, whereas other studies indicated that females are at greater risk or that there is no identifiable gender predilection (Rentko et al. 1992, Harkin and Gartrell 2002), as was seen in the present study. Certain breeds, such as herding and hound breeds, are at greater risk (Ward 2002). Although these breed groups include popular dogs such as beagles and German shepherd dogs, these breeds may also have disproportionately increased risk by virtue of their uses on farms or in other outdoor activities. Although only large breeds had significantly elevated risk, as reviewed previously (Greene et al. 2012), small breed dogs may still have some risk as our data show. In contrast to a previous study (Geisen et al. 2007) that showed a bimodal distribution of age susceptibility (dogs <1 year of age and dogs ≥8 years of age), only older dogs were at increased risk for infection in our data. A possible explanation for this may be increased vaccination of young dogs with more effective quadrivalent vaccines, which is reinforced by the fact that dogs for which vaccination status data were available that had leptospirosis had not recently received the quadrivalent vaccine.

The present study identified an increased risk of leptospirosis among dogs living in the central and south coast, foothills, and San Francisco Bay regions of California when compared with control dogs. The risk for cases in the San Francisco |Bay region was also corroborated by a recent spatial analysis in our laboratory (Hennebelle et al. 2013). It is not known what features of these areas confer this risk, although we speculate that coastal areas are cooler and moister, allowing for longer persistence of the spirochete, whereas inland high-risk areas may still have patchy moist areas or almost certainly some wild or domestic animal reservoirs. These data clearly suggest that additional work on periurban wildlife is warranted.

Assessing wildlife-human–domestic animal cycles is difficult, however, because of cross-reactivity in serological assays among serovars. Nevertheless, it was striking that the strongly predominant serovar in our data, both in terms of numbers of dogs seroreacting and intensity of the serological titer (data not shown), Serovar Pomona, predominates in animal species in the western United States ranging from horses to dogs, raccoons, skunks, and even marine mammals, although genetic analysis suggests that sea lion strains may form a clade separate from terrestrial strains (Colagross-Schouten et al. 2002, Zuerner and Alt 2009). This contrasts with other parts of the United States; e.g., antibodies to serovar Grippotyphosa are more common in the upper Midwest and on the east coast (Sykes et al. 2011) where voles and some other rodents have been implicated as reservoirs (Woods 1974, Hathaway et al. 1983, Kuiken et al. 1991). In the past, canine and rat-origin serovars Icterohemorrhagiae and Canicola were most commonly identified in dogs in North America, but the prevalence of seroreactivity to these serovars in dogs appears to have declined relative to serovars Grippotyphosa, Pomona, Bratislava and Autumnalis following increased use of the bivalent canine vaccine (Goldstein et al. 2006, Campbell 2007, Geisen et al. 2007, Miller et al. 2007, Ortega-Pacheco et al. 2008, Picardeau et al. 2008). Thus, changing vaccine regimes and changing epidemiological patterns from rodent-acquired leptospirosis to more mesocarnivore sources may account for some of the data we have.

A complex, intersecting ecology unites risk of potentially fatal leptospirosis among multiple host species. Given the close relationship between dogs and humans, high prevalence and particular risk factors for this disease, even in an area where human cases are rarely diagnosed, suggests possible underreporting in people. Data derived from studying leptospirosis in Californian dogs may provide management targets that could allow us to improve canine health and better understand human risk. We show an ongoing presence of canine leptospirosis, primarily associated with seroreactivity to serovar Pomona, in northern California. At least in this region, older sporting or herding breed dogs are at increased risk and the clinical abnormalities primarily reflect acute renal failure. Although overall survival was high, 40% of dogs required hemodialysis, the cost of treatment was high, and length of hospital stay was significantly longer than that for other dogs seen at a tertiary care referral veterinary hospital. Improved understanding of the epidemiology of leptospirosis and updated vaccination strategies may help to reduce the impact of this disease on the canine population and provide insights for optimal protection of other species, including humans.