Background

Vector-borne infections pose significant health risks to humans, domestic species, and wildlife. Domestic dogs (Canis lupus familiaris) are susceptible to many vector-borne pathogens (VBP) of veterinary and public health concern [1,2,3,4,5,6], with Dirofilaria immitis (heartworm), Borrelia burgdorferi (Lyme disease), Anaplasma spp. (anaplasmosis), and Ehrlichia spp. (ehrlichiosis) being some of the most studied pathogens [7, 8]. Importantly, several of these pathogens are significant pathogens of domestic dogs [9]. Finally, because several of these pathogens are zoonotic, dogs can serve as sentinel hosts. Using Lyme disease as an example, seroprevalence of B. burgdorferi infections in dogs is associated with counties where human infections occur [6, 10]. For these reasons, understanding the epidemiology of vector-borne infections in dogs is important for both human and veterinary health.

The origin and history of individual dogs is key to effective surveillance of canine VBP. Generally, the domestic dog population can be divided into groups such as owned dogs with veterinary care, owned dogs without veterinary care, stray dogs, and shelter dogs. The assumption is that many owned dogs under veterinary care would receive a combination of preventatives specific to VBP and intestinal parasites [9, 11, 12]. Dogs without veterinary care (owned or not) and those that enter shelters (generally > 50% of which are strays) are likely to be at a higher risk of infection with VBP through increased environmental exposure and/or lack of preventatives [7, 13, 14]. Other potential risk factors associated with an increased likelihood of VBP infection include age and body condition [15]. Older dogs will have a higher probability of being exposed during their lifetime, and dogs in poor body condition may be immunologically compromised, increasing infection risk.

Co-infections of vertebrate hosts are common and may complicate diagnosis and treatment or increase the risk of severe disease [4, 16,17,18]. In general, co-infections occur with pathogens with a common vector and/or overlapping geographical ranges [4, 7, 9, 11, 12]. For the pathogens of interest in our study, D. immitis is a mosquito-borne parasite, while the most common vector of B. burgdorferi and Anaplasma phagocytophilum is Ixodes scapularis, the most common vector of Ehrlichia canis is Rhipicephalus sanguineus and presumptively Anaplasma platys, and Amblyomma americanum is a common vector of Ehrlichia chaffeensis and Ehrlichia ewingii [7, 19].

The geographical distribution of tick species is changing, and thus the risk of co-infections is changing. This is reflected in the changes in the distribution and prevalence of canine VBP, highlighting the need for contemporary data [5, 8, 20,21,22,23]. In the Southeastern United States, co-infections between D. immitis and Ehrlichia spp. have been common while co-infections with B. burgdorferi and A. phagocytophilum are more common in the Upper Midwest and Northeast [1, 9, 11, 12]. The increasing detection of A. americanum in northeastern states could increase the risk of Ehrlichia infections in dogs in that region [24, 25]. Infection with many tick-borne pathogens causes similar clinical signs and presentations, but may require different treatments (e.g., Babesia spp.); thus, knowledge regarding the risk of co-infections is important. The purpose of this study is to evaluate the seroprevalence of select VBP (D. immitis, Ehrlichia spp., Anaplasma spp., and B. burgdorferi) in shelter dog populations in the Eastern United States and evaluate risk factors associated with seroprevalence and co-infections.

Methods

Sample collections

From 2016 to 2020, we recruited shelters in 19 states to participate in the study. Focal areas included 97 counties in the eastern half of the United States and were selected due to known geographical ranges of the focal pathogens (Fig. 1). Participating shelters were either asked to test dogs that met the inclusion criteria or provide data that was collected using the same methods and inclusion criteria. To be included, dogs had to be 6 months of age or older and had to originate from the county or neighboring counties around the shelter. Dogs that were too young or under bite quarantine were excluded. Dog blood samples were tested using SNAP® 4Dx® Plus tests which were provided by IDEXX Laboratories and shipped directly to the shelter or to the University of Georgia (UGA) for sample testing. We requested that shelters test between 50 and 200 dogs. Furthermore, SNAP® 4Dx® Plus data were provided by investigators from a related shelter study in Mississippi, which included data collected from June 2016 through February 2017 [15]. The approximate age, location, date the sample was collected, estimated breed and the corresponding American Kennel Club (AKC) breed group, sex, and intact status were collected for each dog.

Fig. 1
figure 1

Map of the counties included in the study and seroprevalence maps of Borrelia burgdorferi, Anaplasma spp., Ehrlichia spp., and Dirofilaria immitis. For the county map, gray counties are where dogs originated from and black counties are where the shelters are located. For the seroprevalence map, the circles denote the approximate number of dogs tested in the state, and states in gray were not included in the study. The maps were created in R statistical software (R Core Team)

Full size image

Data analysis

Variables assessed as risk factors included age (< 1 year old or ≥ 1 year old), location, breed group, intact status, and sex. Dogs were classified into AKC breed groups based on dominant breed features because many of the dogs were of mixed breed. Each pathogen and co-infection combination were evaluated using binomial generalized linear models (GLMs) to assess the relationship between the risk factor and the SNAP® 4Dx® Plus test outcome (positive, negative). Significant results from the GLMs were further analyzed through pairwise comparison to calculate the odds ratio (OR), confidence interval (CI), and P-value using the lsmeans package [26]. P-values were adjusted to correct for multiple comparisons using Tukey’s honestly significant difference test. All pathogens were assessed on an individual basis per pathogen and further analyzed through observed co-infection combinations, excluding dogs with no applicable values for the variable being analyzed. Statistical analyses were performed in R version 4.1.1 [27] and factors with P ≥ 0.05 were considered significant.

Results

Population data

The study included 3750 dogs from 19 states and 97 counties with dogs per county ranging from two to 226. The demographic data of the dogs in our study population can be found in Table 1. Most dogs were ≥ 1 year and intact; similar numbers of males and females were sampled (Table 1). The AKC breed group with the largest representation was the terrier breed group (n = 1185/3750, 31.6%) (Table 1). Nearly half (n = 2036/3750, 54.3%) of the dogs were from the Southeast with the remaining were from the Midwest (n = 974/3750, 26.0%) and Northeast (n = 740/3750, 19.7%) (Table 1).

Table 1 Demographic data and seroprevalence of infection with vector-borne pathogens among 3750 dogs from shelters in 19 states in the Eastern United States
Full size table

Dirofilaria immitis antigen seroprevalence

A total of 419 (11.2%, n = 3750) dogs were positive for D. immitis antigens (Table 1, Fig. 2). Dogs < 1 year of age had a significantly lower seroprevalence and were 12.2 times less likely to have D. immitis antigens detected compared with dogs ≥ 1 year of age (Tables 1 and 2). A higher seroprevalence was noted for males than for females (Tables 1 and 2). Dogs that were intact were twice as likely to be positive as non-intact dogs (Tables 1 and 2). The seroprevalence of D. immitis antigens was lower in the toy breed group than in the herding, hound, non-sporting, sporting, and terrier breed groups (Tables 1, 2).

Fig. 2
figure 2

Bar plot of single seroprevalence data and co-infection data. The main bar plot represents the dogs that either had only one pathogen detected or had co-infections detected. The side bar plot represents the single seroprevalence data without the subtraction of the co-infections

Full size image
Table 2 Significant results of bivariable generalized linear model and pairwise comparison analysis of potential risk factors for seropositive status for vector-borne pathogens among 3750 dogs from shelters in 19 states in the Eastern United States
Full size table

Dirofilaria immitis antigen detection was highest in dogs from the Southeast (17.4%, n = 355/2036) followed by dogs from the Midwest (5.3%, n = 52/974) and dogs from the Northeast (1.6%, n = 12/740) (Table 1, Fig. 1). Dogs from the Midwest were 3.4 times more likely to have D. immitis antigen detection than dogs from the Northeast (Table 2). Dogs in the Southeast were 3.7 times more likely to have D. immitis antigen detection than dogs from the Midwest and 12.7 times more likely than dogs from the Northeast (Table 2). An analysis of state vs. state comparisons is provided in Additional file 1: Text S1 and Table S1.

Ehrlichia spp. antibody seroprevalence

Ehrlichia spp. antibodies were detected in 8.0% (n = 299/3750) of the dogs (Table 1, Fig. 2). Analyzed risk factors in relation to detection of Ehrlichia spp. antibodies of significance were age, breed group, and location (Table 2). Dogs ≥ 1 year of age were 3.7 times more likely to have Ehrlichia spp. antibody seroprevalence than dogs < 1 year of age (Table 2). The hound breed group had the highest seroprevalence of Ehrlichia spp. antibodies (Table 1). The hound group was at an increased risk of Ehrlichia spp. antibody detection compared to the other analyzed breed groups (Table 2). The herding breed group was found to be 3.2 times more likely to have Ehrlichia spp. antibody detected than the toy breed group (Table 2).

The highest seroprevalence of Ehrlichia spp. antibodies was documented in dogs from the Southeast (10.7%, n = 217/2036) (Table 1, Fig. 1). The lowest seroprevalence was documented in dogs from the Northeast (4.6%, n = 34/740) and dogs from the Midwest had a seroprevalence of 4.9% (n = 48/974) (Table 1). Dogs from the Southeast had a statistically higher risk of Ehrlichia spp. antibody seroprevalence (Table 4). Dogs from the Southeast were 2.5 times more likely to have Ehrlichia spp. antibody seroprevalence than dogs from the Northeast and 2.3 times more likely than dogs from the Midwest (Table 2). An analysis of state vs. state comparisons is provided in Additional file 1: Text S2 and Table S2.

Anaplasma spp. antibody seroprevalence

Anaplasma spp. antibodies were detected in 2.4% (n = 90/3750) of the dogs (Table 1, Fig. 2). Risk factors that were significant in relation to Anaplasma spp. antibody seroprevalence included breed group and location (Table 2). The sporting breed group had the highest seroprevalence of Anaplasma spp. antibodies of 5.6% (n = 26/466) (Table 1). The sporting breed group was at an increased risk of Anaplasma spp. antibody seroprevalence compared to the terrier and toy breed groups (Table 2). The sporting breed group was 4.1 times more likely to have Anaplasma spp. antibodies detected than the terrier breed group (Table 2). The sporting breed group was 24.5 times more likely to have Anaplasma spp. antibodies detected than the toy breed group (Table 2).

The highest seroprevalence of Anaplasma spp. antibodies was documented in dogs from the Northeast (5.7%, n = 42/740) (Table 1, Fig. 1) followed by dogs from the Midwest (2.0%, n = 19/974) and dogs from the Southeast (1.4%, n = 29/2036) (Table 1, Fig. 1). Dogs from the Northeast were 4.2 times and 3.1 times more likely to have Anaplasma spp. antibodies than dogs from the Southeast and Midwest, respectively (Table 2). An analysis of state versus state comparisons is provided in Additional file 1: Text S3 and Table S3.

Borrelia burgdorferi antibody seroprevalence

Borrelia burgdorferi antibodies were detected in 8.9% (n = 332/3750) of dogs (Table 1, Fig. 2). Risk factors found to be significantly related to detection of B. burgdorferi antibodies were age, breed group, and location (Table 2). Dogs ≥ 1 year of age were 3.5 times more likely to have B. burgdorferi antibodies than dogs < 1 year of age (Tables 1 and 2). The hound breed group had the highest seroprevalence of B. burgdorferi antibodies (15.5%, n = 78/503) (Table 1). The herding, hound, sporting, terrier, and working breed groups had significantly higher seroprevalence of B. burgdorferi compared to the other analyzed breed groups (Table 2). Additionally, the toy breed group (2.9%, n = 12/418) had a decreased risk of B. burgdorferi antibody detection (Table 2).

The highest B. burgdorferi antibody seroprevalence was documented in dogs from the Northeast (19.3%, n = 143/740) followed by dogs from the Midwest (6.8%, n = 66/974) and dogs from the Southeast (6.0%, n = 123/2036) (Table 1, Fig. 1). Dogs from the Northeast had an increased risk of B. burgdorferi antibody detection which was 3.3 times more likely than dogs from the Midwest and 3.8 times more likely than dogs from the Southeast (Table 2, Fig. 1). An analysis of state versus state comparisons is provided in Additional file 1: Text S4 and Table S4.

Co-infections

There were nine different co-infection combinations observed in 179 dogs (Table 3, Fig. 2). The three most prevalent co-infections were B. burgdorferi + Anaplasma spp. (1.47%, n = 55/3750), B. burgdorferi + Ehrlichia spp. (1.23%, n = 46/3750), and D. immitis + Ehrlichia spp. (1.57%, n = 59/3750) (Tables 3, 4, Figs. 2, 3).

Table 3 Co-infections among 3750 dogs from shelters in 19 states in the Eastern United States
Full size table
Table 4 Demographic data and seroprevalence of infection with the three most prevalent co-infection combinations (Borrelia burgdorferi + Anaplasma spp., B. burgdorferi + Ehrlichia spp., and Dirofilaria immitis + Ehrlichia spp.) pathogens among 3750 dogs from shelters in 19 states in the Eastern United States
Full size table
Fig. 3
figure 3

Map of co-infections included in the study. Each co-infection combination is represented with a unique color and the included states are shaded gray. The points are representative of a single co-infection and are randomly placed within the state of origin. The map was created in R statistical software (R Core Team)

Full size image

For all three of these co-infection pairs, location was a significant risk factor (Table 5). The B. burgdorferi + Anaplasma spp. pair was more likely to be detected in the Northeast than in the Southeast (Tables 4 and 5, Fig. 3). The co-infection combination between B. burgdorferi + Ehrlichia spp. was mostly observed in the Northeast and was significantly more likely to be detected in the Northeast than in the Midwest, although all of the Northeast positives were from Maryland (Table 5, Fig. 3). The co-infection pair of D. immitis + Ehrlichia spp. was significantly more likely to occur in the Southeast than the Midwest, and no cases were noted in the Northeast (Tables 4 and 5, Fig. 3).

Table 5 Significant results of the binomial generalized linear model and pairwise comparison analysis of potential risk factors for positive status for the three most common co-infections noted in this study among 3750 dogs from shelters in 19 states in the Eastern United States
Full size table

There were also significant associations for breed group for co-infections B. burgdorferi + Anaplasma spp. and B. burgdorferi + Ehrlichia spp. (Table 5). The sporting breed group was 3.8 times more likely to have co-infections with B. burgdorferi + Anaplasma spp. than the terrier breed group (Table 5). The hound breed group was 7.0 times more likely to have co-infections with B. burgdorferi + Ehrlichia spp. than the sporting breed group and 4.1 times more likely than the terrier breed group (Table 5).

Additional risk factors for co-infections with D. immitis + Ehrlichia spp. were age group and intact status (Table 5). Dogs that were ≥ 1 year of age were 9.3 times more likely to have a co-infection with D. immitis + Ehrlichia spp. than dogs < 1 year of age (Table 5). Dogs that were intact were 2.1 times more likely to have co-infections with D. immitis + Ehrlichia spp. than dogs that were not intact (Table 5).

Discussion

In the present study, we investigated the seroprevalence of VBP (D. immitis, Ehrlichia spp., Anaplasma spp., and B. burgdorferi) in shelter dogs in the Eastern United States from 2016 to 2020. We found regional variation in seroprevalence of all pathogens and several risk factors (age, sex, breed group, and intact status) were associated with infection. We also observed several different co-infection combinations with B. burgdorferi + Anaplasma spp., B. burgdorferi + Ehrlichia spp., and D. immitis + Ehrlichia spp. being the most frequently detected. This study provides contemporary data on the seroprevalence of these pathogens in a group of dogs that are expected to have limited veterinary care or preventative use. High seroprevalence and detection outside of known endemic regions highlight the need for continued monitoring.

Knowledge regarding the accurate distribution of pathogens is critically important for veterinarians and clients to gauge the risk of disease in dogs and other possible hosts. In general, the geographical distributions for the VBP were consistent with the known geographical ranges reported in previous studies for both the pathogens and vectors [1, 4, 5, 9, 11, 12, 15, 16, 22, 28]. However, there were some notable findings. We detected a low seroprevalence (1.1%, n = 2/189) of D. immitis in Maine which historically has few heartworm detections (< 0.5%) [11]. Detection of D. immitis outside the known endemic range is often assumed to be related to travel or translocated dogs from heartworm-endemic regions. However, the inclusion criteria of this study should have excluded most translocated dogs, and our findings are supported by increasing heartworm prevalence trends in the far Northeast and other regions (e.g., Colorado) [11, 21, 29, 30]. The potential for local transmission in non-endemic regions highlights the need for heartworm preventative use. In addition, we noted a higher seroprevalence of B. burgdorferi infection in dogs from Virginia compared to past studies using similar methods [11], which corresponds with reported changes in the distribution of this pathogen in dogs and people and its vector in Virginia [8, 31,32,33,34]. In addition, a low number of B. burgdorferi-positive dogs were detected in states where Lyme disease risk is low (e.g., Missouri, Florida, and Georgia). Our inclusion criteria should have excluded translocated dogs, but we do not know whether all shelters were 100% compliant (although few dogs from the Northeast/Upper Midwest where Lyme disease is more common are moved south). Additional studies are needed to investigate the possible transmission of B. burgdorferi in these areas. Similarly, a few states had notable detections of Ehrlichia (e.g., Maine, New Hampshire, Minnesota, and Wisconsin). Although the species involved is unknown, these detections may be due to the well-documented northern expansion of A. americanum in the Northeast or E. muris eauclairensis in Minnesota and Wisconsin [24, 25, 35]. Continued studies on the distribution of VBP are warranted, as changes in the distribution and density of vectors and their associated pathogens have been noted in recent years, which may be related to several factors such as climate or habitat changes [5, 8, 20,21,22,23, 36]. Additionally, novel vectors (e.g., Asian longhorned tick, Haemaphysalis longicornis) have been introduced into the United States, and this tick may alter the native pathogen transmission dynamics [37,38,39,40,41].

Consistent with previous studies, dogs that were ≥ 1 year of age had an increased risk of being positive [7, 15]. Tick-borne pathogen infection was assessed with the detection of antibodies which may be persistent; older dogs have an increased time at risk of exposure, increasing the likelihood of infection to the vector and pathogen. Heartworm infections also typically occur in older dogs as this parasite has a long life cycle and older dogs have increased time for mosquito exposure [42]. We also found that intact dogs were more likely to be infected with D. immitis compared with non-intact dogs, who may be more likely to have previous access to veterinary care or preventative medications. Other studies have also noted that intact dogs had higher rates of VBP infection [43, 44] and tick exposure [45].

Interestingly, we found that breed group was associated with changes in infection risk. Our findings indicate that the toy breed group specifically had a lower risk of infection for all pathogens, which is consistent with a previous study on ticks [45]. The decreased risk of infection for the toy breed group may be due to the popularity of these dogs in urban settings or that smaller dogs likely spend less time outdoors. Similarly, larger breed groups (e.g., hound, herding, sporting) had higher seroprevalence of infection, which may be related to more time spent outdoors. There may also be spatial differences in breed distribution because of different trends in owner popularity and dog utility. For example, both D. immitis and Ehrlichia were more common in large breed groups that are popular outdoor dogs in Southern states. Although this study focused on VBP, similar results (increased risk of parasitic infections for large breed groups) have been noted in other parasite systems (e.g., Dracunculus) [46, 47].

In general, the seroprevalence of pathogens included in this study was comparable to previous shelter-based studies conducted in the same regions using similar methods (e.g., 16.0% D. immitis seroprevalence in Texas) [7]. However, our seroprevalence rates for D. immitis, B. burgdorferi, and Ehrlichia spp. were higher than those in other studies primarily conducted on owned dogs and public data available on the Companion Animal Parasite Council (CAPC) website [11, 28]. Our observed higher seroprevalence rates most likely were because our sample collection was from shelter dogs, and > 50% of the dogs that enter a shelter are considered strays [13]. These dogs are expected to have decreased access to veterinary care and therefore preventative medications. Further analysis of this association is reported separately [48].

Dogs are not exposed to vectors or pathogens in isolation, and most geographical regions have multiple vectors and pathogens co-circulating. Many large serosurvey studies on dogs and pathogen infection are unable to examine co-infections because testing data are not associated with individual cases [11, 21]. However, prospective studies, such as this one, provide opportunities to examine patterns of co-infections [7]. Although we detected nine co-infection groups, including 11 dogs with infection with three pathogens, the most common pairs were B. burgdorferi + Anaplasma spp., B. burgdorferi + Ehrlichia spp., and D. immitis + Ehrlichia spp. The most likely explanation for the increased frequency between these three combinations is because of the pathogens' shared ranges and vectors. Anaplasma phagocytophilum and B. burgdorferi have overlapping ranges in the Northeast, which was the region of highest seroprevalence documented in our study, and share the same vector (I. scapularis) [9, 49]. Co-infections between B. burgdorferi and Ehrlichia spp., and D. immitis and Ehrlichia spp. were most likely more common due to overlapping ranges. The Ehrlichia spp. most commonly detected in our study is likely E. canis, due to its documented prevalence in the South; however, other Ehrlichia spp. are possible [9].

Conclusions

In this study, we analyzed seroprevalence data for VBP in 3750 dogs sampled in shelters from 97 counties in 19 states in the Eastern United States. In general, we found D. immitis and Ehrlichia spp. seroprevalence to be highest in the Southeast and Anaplasma spp. and B. burgdorferi seroprevalence to be highest in the Northeast, which is consistent with previous studies and expected vector ranges. Co-infections that were most common were between B. burgdorferi + Anaplasma spp., B. burgdorferi + Ehrlichia spp., and D. immitis + Ehrlichia spp. We found decreased risk of infection in dogs that were less than 1 year of age, in the toy breed group, and in dogs that had been spayed or neutered. In general, we found increased risk of infection in dogs that were more than 1 year of age and in the hound breed group. However, some pathogens were detected outside their typical range, so these data support previous studies that show an expanding range for these pathogens and/or vector species, highlighting the need for continued surveillance and assessment of risk factors in both owned and unowned dogs.