Introduction

The expansion of tick-borne disease in populations of domestic dogs is a one health issue that poses a significant and increasing health threat to canids and to humans.¹ Dogs can be infected with numerous tick-borne zoonoses and may serve as sentinels of disease risk due to their close interaction with humans.^2,3 Dogs also support tick populations that may bite and infect humans and spread disease between dogs, and are a critical component of the rapidly changing landscape of tick-borne rickettsial pathogens.⁴

The genus Rickettsia, along with Anaplasma, Ehrlichia, and several other genera, falls within the order Rickettsiales, a group of gram negative, obligately intracellular bacteria of varying pathogenicity to humans and animals.^5,6 The genus Rickettsia currently comprises around 30 validly-described species including 20 documented pathogens.⁷ In the western hemisphere, Rocky Mountain spotted fever (RMSF; called Brazilian spotted fever or spotted fever rickettsiosis outside the United States), caused by Rickettsia rickettsii, causes the most severe disease in both humans and dogs of any of the Rickettsia species.^8,9 In the US, until the early 2000s, R rickettsii was thought to be associated exclusively with Dermacentor species ticks, particularly D variabilis, the American dog tick, and occurred primarily in the midwestern and eastern US.^5,10 However, in 2002, a human epidemic of RMSF began in Arizona, associated with Rhipicephalus sanguineus, the brown dog tick.⁵ Simultaneously, cases emerged in northern Mexico, also associated with brown dog ticks, and the outbreaks were accompanied by startlingly high human fatality rates ranging from 12% to over 30%.^10,11 While health impact of RMSF on dogs in these outbreaks has not been quantified, serologic data suggests that dog exposure in epidemic areas is widespread, while exposure is much lower in areas where human cases have not been reported.^12–14

The brown dog tick, unlike other vectors of R rickettsii, is dependent on dogs to complete its life cycle, and the emergence of RMSF is associated with heavy environmental burdens of ticks resulting from large numbers of dogs, particularly free-roaming dogs.^2,15 The brown dog tick thrives in peridomestic environments in and around homes and kennels, and, while it typically feeds on dogs at all life stages, incidental feeding on humans occurs, especially when ticks are at high density and possibly at elevated temperatures.^4,16 In North America, there are 2 species of brown dog tick (called the tropical and temperate lineages, both members of the Rh sanguineus species complex), which cannot be easily distinguished by appearance but have very similar ecology and life history.^4,17 Both are responsible for RMSF outbreaks, though they have not been directly compared for transmission efficiency of R rickettsii.¹⁷

In addition to its role as a vector of R rickettsii, the brown dog tick transmits Ehrlichia canis, the cause of canine monocytic ehrlichiosis, and Anaplasma platys, the cause of infectious cyclic thrombocytopenia in dogs.^5,18,19 Dogs and brown dog ticks support other spotted fever group rickettsial zoonoses in North America as well, including R massiliae, which may represent an emerging threat in the same regions at risk of RMSF outbreaks.²⁰ Ehrlichia canis and A platys were historically regarded as strictly canine diseases, but both have now been proven to infect humans as well.^21,22 In addition, because they are transmitted by the same tick, their presence may be a surrogate indicator of risk for spotted fever group rickettsiosis.²³ There is significant overlap in the clinical signs of RMSF, ehrlichiosis, and anaplasmosis in dogs, with nonspecific signs like fever and thrombocytopenia common to all 3 diseases.²⁴ Infection with R rickettsii may lead to severe RMSF and death in dogs, or infection may be subclinical.⁹ Monocytic ehrlichiosis and anaplasmosis often progress to chronicity, and chronic E canis infection may manifest in multiple-organ disease impacting kidneys, eyes, and joints among other tissues, and result in immune suppression and increased susceptibility to other infections.²⁵ Coinfection with both A platys and E canis results in greater hematologic abnormalities, increased surgical or perioperative bleeding, and longer duration of infection than infection with either alone.^26,27

While RMSF has not been reported in California, E canis and A platys infections have been diagnosed in clinically ill dogs in southern California, suggesting that transmission of RMSF could occur.²⁸ In addition, the presence of an ongoing, severe epidemic of human RMSF in northern Baja California and Arizona, immediately adjacent to Imperial and San Diego counties in southern California, suggests risk of disease emergence in California as well.^29,30

Sheltered and stray dogs have been identified as a critical reservoir of zoonotic pathogens,^15,31 and the movement and re-homing of sheltered dogs is a documented risk factor for vector and pathogen spread.³² Sheltered dogs may have higher parasite exposure than dogs observed in other contexts (eg, those presenting to veterinary clinics) due to social and economic biases.³¹ Surveillance of sheltered dogs, therefore, offers a potential opportunity to identify emerging zoonoses that threaten both human and canine health. Previous studies on RMSF risk in dogs have depended primarily on serologic data on dogs from wider populations, rather than assessing the utility of sheltered dogs in and around outbreak areas.^12–14,33 We hypothesized that combining diagnostic modalities—PCR and serology—for multiple pathogens in the shelter setting would provide a more complete picture of rickettsial disease risk in a region, as well as providing valuable data on the health and detection of disease of sheltered dogs. The purpose of this study, therefore, was to assess exposure to and infection with 3 pathogens (R rickettsii, A platys, and E canis) vectored by brown dog ticks in sheltered dogs in southern California and northern Baja California, with the goal of determining a baseline for current and potential transmission of RMSF in this region.

Materials and Methods

Sample collection

A cross-sectional survey was conducted at 7 animal shelters in southern California and northern Baja California between October 2021 and May 2022. The study area was divided into 4 quadrants: San Diego County and Imperial County in California, and Tijuana and Mexicali in Baja California (Figure 1). Dogs at least 6 weeks of age of any size and breed were enrolled as a convenience sample based on presence in shelter at time of survey. All dogs were part of the general shelter population (ie, not quarantined or hospitalized) and not under treatment for any condition at the time of sample collection. Demographic data collected included breed, sex, estimated weight and age (based on dentition), body condition (scored out of 5),³⁴ and, where possible, location of origin and shelter intake date. Dogs were examined for ticks on pinnae of ears, interdigital areas, and around collar area, and ticks were collected when present and stored in 70% ethanol. Whole blood (1 to 3 mL per dog) was collected from each dog and stored at –20 °C until analysis. This research was approved by the UC Davis IACUC (Protocol #22235).

View Full Size

Map showing locations of shelters in the 4 study quadrants (San Diego County, Imperial County, Tijuana, and Mexicali) along the western US–Mexico border. Map data by OpenStreetMap contributors. Map layer by Esri.

Citation: Journal of the American Veterinary Medical Association 261, 3; 10.2460/javma.22.08.0388

• Download Figure
• Download figure as PowerPoint slide

Results

A total of 239 dogs were sampled across the 4 quadrants at 7 shelters. Of these, 114 (47.7%) were female. Of the Mexicali and San Diego samples, approximately two-thirds were male, while two-thirds were female in Tijuana (Table 2). The majority (197/239 [82.4%]) of dogs across all sites were medium or large (> 10 kg). Most dogs (203/239 [84.9%]) were adults, with a similar age structure between study sites, though no puppies were sampled in San Diego.

A total of 200 ticks, all brown dog ticks, were collected from 36 (57.1%) dogs in Mexicali and 8 (10.3%) dogs in Tijuana; the first 10 ticks from each dog were used for PCR analysis (Table 3). No ticks were found on dogs in Imperial County or San Diego County in California. The majority of ticks were adults (66/200 [33%] female, 75/200 [37.5%] male) and nymphs (57/200 [28.5%]); only 2 larvae (1%) were collected. A single tick—collected from a dog that was co-infected with E canis and A platys—was positive for both E canis and A platys.

Two dogs were positive on pan-Rickettsia PCR, and both were subsequently positive on R rickettsii–specific PCR. The 2 positive dogs were from 2 different shelters in the Tijuana area sampled in December. Both were adult males; one was a mid-sized shepherd mix, and the other a Labrador retriever mix. Neither showed overt signs of illness at time of sampling, though one of the dogs was also seropositive for Ehrlichia and Anaplasma, while the other dog was PCR-positive for A platys as well as R rickettsii. Rickettsia rickettsii was further confirmed via DNA sequencing of the 17kDa, gltA, and OmpA genes for each. Rickettsial seroprevalence differed significantly (P < .001) among the shelters, ranging from 15.2% (7/46) of dogs in Imperial up to 81% (51/63) in Mexicali. Two ticks collected from different dogs (neither of which was PCR-positive for R rickettsii) in 1 shelter in Tijuana were positive for Rickettsia spp.; one was positive for R rickettsii, while the other rickettsial agent could not be identified to species. The R rickettsii positive sample could only be successfully sequenced using the 17kDa gene, but was 100% homologous to the R rickettsii 17kDa sequences obtained from the dogs. The single tick from Mexicali that was positive by pan-Rickettsia screening was determined to be infected with R felis via sequencing of the 17kDa gene.

E canis and A platys seroprevalence and PCR-prevalence were significantly different among locations (Table 4). Both PCR (25/78 [32.1%]) and antibody (35/78 [44.9%]) prevalence for Anaplasma were significantly higher in Tijuana than the other locations (Figure 2). For Ehrlichia, rates were much higher in Mexicali than other locations (17/63 [27.0%] PCR-positive and 31/63 [49.2%] seropositive). No dogs were positive by serology for B burgdorferi. A single dog from San Diego was positive for D immitis infection and was also seropositive for Rickettsia exposure, but otherwise negative on all assays.

View Full Size

Proportion with 95% CI of dogs positive by serology and PCR for tick-borne pathogens at each location.

Citation: Journal of the American Veterinary Medical Association 261, 3; 10.2460/javma.22.08.0388

• Download Figure
• Download figure as PowerPoint slide

Detecting evidence of recent exposure to tick-borne disease among dogs would better inform assessment of and response to tick-borne disease risk. Rickettsemia in RMSF is highly transient and thus the PCR-positive samples we detected confirm high risk at the time of sampling. In contrast, both A platys and E canis can cause chronic infection and therefore a positive PCR result is not necessarily indicative of recent exposure. Because such chronically ill dogs would be expected to be seropositive, we identified PCR-positive cases of E canis and A platys that were negative on serology as a proxy measure for possible acute cases. Out of the 30 total A platys PCR-positive dogs, 25 (83.3%) were from Tijuana, and of those, 12 were seronegative (48.0%, 95% CI [28.3 to 68.2]), suggesting widespread recent infection. This is in contrast to Mexicali, where only 5 individuals were PCR-positive for A platys, but 4 of those were acute (80%, 95% CI [29.9 to 98.9]). Only 3 acute cases of E canis were identified, all from Mexicali.

Results of generalized linear mixed models showed that being severely underweight (BCS of 1.5/5 or below) was associated with a 7.24 (95% CI, 1.77 to 32.63) times higher odds of being positive by PCR for E canis. Puppies had a 4.07 (95% CI, 1.16 to 12.92) times higher odds of being PCR-positive for E canis than adult dogs in a univariable model, though age was no longer significant when included in the model with underweight status. For Anaplasma spp antibodies, small size (< 10 kg) was associated with a 77.6% reduction in odds. Infestation with ticks was associated with a 4.05 times increase in the odds of being PCR-positive for A platys (95% CI, 1.21 to 15.36; P < .001). Being seropositive for Ehrlichia was associated with a 3.42 times increase in the odds of being seropositive for Rickettsia (P < .001; 95% CI, 1.63 to 13.31), while Anaplasma seropositivity was not significantly associated with Rickettsial seropositivity (P = .20). No other demographic characteristics were associated with positivity by serology or PCR for any of the 3 pathogens.

χ² testing suggested that active co-infection with Ehrlichia and Anaplasma spp was not significantly (χ² 0.368; df = 1; P = .554) more likely than infection with either pathogen individually. However, there was significant evidence of co-exposure based on serology (χ² = 51.79; df = 1; P < .001).

Discussion

Shelter-based surveillance is a valuable tool to explore community risk for RMSF or other zoonotic or canine diseases. In our study, we found very high regional variability in rates of infection and exposure to tick-borne pathogens, indicating that risk of infection in dogs is markedly different even in adjacent locations. Even where we did not find active infections, there was considerable exposure to Ehrlichia, Anaplasma, and Rickettsia species, suggesting that pathogens are circulating in those areas. While they may frequently go undetected, there is likely a greater burden of disease in dogs than is currently recognized as well as potential for transmission of tick-borne zoonoses.

Detection of R rickettsii in blood samples from 2 dogs in Tijuana suggests that the infection prevalence at the time sampling occurred was high. Rickettsemia is transient, lasting only 3 to 14 days in infected dogs.⁹ Studies in humans demonstrate that the diagnostic sensitivity of PCR on blood is low,⁴² even though PCR has excellent analytic specificity for Rickettsia species, and can detect a single copy of bacterial DNA in a reaction.³⁶ The 2 PCR-positive dogs were seronegative and, given that seroconversion occurs 10 to 14 days post-inoculation, this suggests they were early in the course of disease.⁹

While experimental and clinical data on R rickettsii infections in dogs have confirmed fever, petechiation, anorexia, and lethargy as primary clinical signs,^9,24 syndromic surveillance of RMSF in shelters would have significant limitations. Neither RMSF PCR-positive dog in Tijuana appeared clinically ill on brief observation, though subtle clinical signs such as changes in appetite or behavior would be difficult to appreciate in these shelter settings. Given that RMSF is likely to occur in the same populations where other tick-borne pathogens are prevalent, routine clinical assessment in the absence of fulminant infection or death may not be adequately sensitive in the shelter setting to detect RMSF. In addition, co-infection and co-exposure to multiple tick-borne pathogens, including 1 R rickettsii PCR-positive, Ehrlichia and Anaplasma seropositive dog and 1 R rickettsii and A platys co-infected dog in our survey, could obscure the etiology of illness in dogs. Finally, because tetracycline antibiotics are used for all 3 pathogen genera, dogs with clinical signs may be diagnosed and treated, with subsequent resolution, without any recognition of R rickettsii infection.

Reliance on PCR can yield an incomplete picture of the community’s risk if acute infection prevalence is low, with a low sample size, or outside of the most active tick-transmission season, although PCR has the benefit of allowing for precise identification of the etiologic agent in most cases. There is also considerable benefit to using serology to detect risk, given that it provides evidence for exposure over multiple months. Serology has the added benefit of detecting population level susceptibility; while high levels of seropositivity suggests a history of high level of exposure, lower levels may indicate that a population has minimal immunity and is therefore at higher risk of disease emergence with rapid spread. In the US–Mexico border region, rickettsial serologic positivity ranged from 15% in Imperial County, 29% in San Diego, 60% in Tijuana, to 80% in Mexicali. This variation likely results from a combination of factors that result in differences in tick and pathogen exposures between locations—including dog density, presence of free roaming dogs, and socioeconomic factors—and are consistent with previous studies in the region that have shown high seroprevalence in areas of known RMSF outbreaks.^12,13,15,33 The very high seropositivity in Mexicali may indicate a high level of population level immunity that protects against emergence of cases and spillover to humans, a hypothesis that is supported by a recent decline in cases in the city of Mexicali.³⁰ The lower levels of seropositivity seen in San Diego and Imperial counties suggest that under conditions of adequate dog density and tick presence, rapid disease emergence could occur.

Serology also detected circulation of E canis and A platys at all locations, confirming widespread exposure to tick-borne pathogens and to brown dog ticks. In addition, E canis exposure was strongly associated with rickettsial exposure. Anaplasma platys exposure was not significantly associated with rickettsial exposure, unlike previous research in Imperial County, which found a significant relationship between co-exposure of Anaplasma and Rickettsia and not between Ehrlichia and Rickettsia, which may be attributable to the wider geographic coverage and different source population used in this study.³³ The relatively large proportion of A platys infections that were classified as potentially acute due to being PCR-positive and seronegative suggests that in-shelter transmission was occurring, though these cases occurred in all 4 shelters where samples were collected in Tijuana and Mexicali and not just at a single location.

However, there are several limitations to using serology for surveillance in such a context. First, the lag in time to seroconversion means that active infections may be missed. Second, serology is not specific for any of the 3 target pathogens in this study. The SNAP 4dx Plus test does not discriminate between exposure to A phagocytophilum and A platys or between E canis and E ewingii, though E ewingii is not thought to occur in western North America, and A phagocytophilum is transmitted by Ixodes species ticks.²⁴ Ixodes pacificus occurs very sparsely in temperate areas of San Diego and northern Baja California,³⁵ so it is possible that Anaplasma seropositive dogs had been exposed to A phagocytophilum, although the majority were likely exposed to A platys, supported by the findings of PCR-positive dogs for that pathogen.

Similarly, rickettsial seropositivity could be due to a variety of cross-reacting pathogens.⁴³ While some subset is almost certainly in response to R rickettsii, other circulating rickettsial organisms may be responsible for seropositivity, particularly in San Diego, where nearly 30% of dogs were seropositive despite no reports of RMSF so far in the region. There is less data on rickettsial exposure of dogs in San Diego County than the other locations included in our study, so it is unclear whether the relatively high seroprevalence we found is in line with previous research. Rickettsia felis is an emerging zoonotic pathogen most commonly transmitted by fleas and is increasingly associated with murine typhus cases in the United States and Mexico.⁴⁴ Rickettsia felis has been identified in the brown dog tick previously, and in this study it was identified in a single tick, but the tick’s role in transmission is unknown.⁴⁵ The human pathogens R massiliae and R parkeri, both present in the study region, cause milder illness than RMSF but cross-react serologically.^46,47 Surveillance for these pathogens is warranted for both their impacts on human and canine health, and is also important for their confounding role in detection and diagnosis of RMSF.

We acknowledge several study limitations including that turnover of the canine populations in Mexican shelters typically was higher than in California shelters, and dogs in Mexicali and Tijuana were being sampled closer to their date of intake. Use of ectoparasite treatment also differed between shelters and time of year, but the impact could not be assessed here because data on treatment history was not available for individual dogs. Because sampling was cross sectional, seasonality may have impacted the prevalence of infection, particularly because shelters were not all sampled at the same time. Brown dog tick populations and rickettsial transmission are known to vary by time of year and weather, and increase in hot weather,¹⁶ while the R rickettsii positive dogs were sampled in December, a cooler time of year in Tijuana; because this was a cross-sectional study, it is unclear whether time of year would affect these findings. In the Mexicali epidemic, human cases typically peak in the summer, but this data is not yet available for Ensenada.³⁰ Repeated sampling through the course of the year would help determine whether there is temporal variation that would impact surveillance. Data were not available on whether dogs were strays or relinquished to shelters by owners; this may have been unequal among regions, and owned versus stray dogs typically differ in rates of exposure to ticks and tick borne pathogens as well as overall health.³¹ Finally, travel history—and therefore whether exposure occurred outside the study area or in one of the other quadrants—could not be ascertained.

Despite these limitations, this study demonstrates value in shelter-based surveillance of zoonoses and highlights the importance of shelter staff awareness of RMSF and tick-borne disease. Dogs entering the shelter fulfill the criteria necessary to be effective RMSF sentinels for both humans and the dog population at large: they are an observed population that can be tested for both infection and exposure, and represent the exposure of a source population.³ In addition, at the shelter level, dogs infected with RMSF who have ticks pose a risk of transmission to staff and volunteers as well as co-housed dogs via contact with infected ticks, even if they do not appear unwell. Given the high prevalence of infection of dogs in this study with A platys and E canis in Tijuana and Mexicali, respectively, these pathogens may pose an underrecognized risk to humans in the region as well.^21,22

This study contributes evidence that risk of diseases transmitted by the brown dog tick is increasing in the border region studied here. A 1994 study found no evidence of E canis in 200 shelter dogs in southern California,⁴⁸ and the historical extent of A platys in Mexico is unclear as it was only detected there in the last decade.⁴⁹ Data from private commercial laboratories show a marked, steady increase in seroprevalence for Anaplasma and Ehrlichia antibodies in dogs from both San Diego and Imperial counties since 2012.⁵⁰ The presence of pathogens that are vectored by brown dog ticks, combined with seropositivity to rickettsial organisms and proximity to Mexican cities experiencing epidemic disease, suggests that both Imperial and San Diego counties are at risk of RMSF emergence. Sheltered dogs represent a wide geographic cross section, likely overlapping with human tick-borne disease risk, and increased surveillance serves to benefit both canine and human health while acting as a valuable source of sentinel data on tick-borne pathogens. To quantify the value of shelter dogs as sentinels for human cases, data collected simultaneously for dogs and human cases should be compared, and represents an important next step in validating this method for surveillance so that interventions to mitigate human disease risk can be implemented accordingly.