Introduction

There is mixed anecdotal evidence and reported literature that ehrlichiosis can lead to the development of both acute and chronic kidney disease (CKD).^1–3 Canine ehrlichiosis may be caused by Ehrlichia canis, E chaffeensis, and/or E ewingii.⁴ Rhipicephalus sanguineus, the brown dog tick, is the primary vector for E canis and has a wide geographic distribution, including the South and Southwest regions of the United States. Symptomatic dogs that have been infected with Ehrlichia spp. have an increased risk of multisystemic disease.⁵ E canis is the causative agent for canine monocytotropic ehrlichiosis and continues to be an important pathogen responsible for severe, life-threatening illness.⁶ Canine monocytotropic ehrlichiosis affects multiple body systems with acute, subclinical, or chronic disease states characterized by lethargy, anorexia, myalgia, splenomegaly, lymphadenopathy, and bleeding disorders.⁶ Clinical signs in dogs chronically infected with E canis may include fever, anorexia, bleeding tendencies, secondary infections due to neutropenia, neurologic and kidney damage secondary to hyperviscosity syndrome, and gastrointestinal signs secondary to azotemia.^7,8 In addition, chronic infection with E canis may have an association with the development of kidney disease in dogs.² Many dogs recover from acute disease with appropriate therapy, although untreated or inappropriately treated dogs may develop subclinical disease that can persist for months to years.⁹ Histologic evidence of kidney disease is associated with E canis infection in several studies.^2,3 In acute experimental infections with E canis, dogs developed transient proteinuria with minimal glomerulopathy and without histologic evidence of glomerular disease.¹ In contrast, a separate study documented renal vasculitis with acute experimental canine ehrlichiosis,^1,3 and a further study of naturally infected dogs with undefined stages of disease showed dogs had both glomerular and tubulointerstitial lesions despite normal creatinine (Cr) concentrations.² E ewingii can cause disease in dogs and people.⁴ Clinical signs of E ewingii may be mild, brief, and self-limiting with acute fever, lethargy, and occasionally neurologic signs present.⁴ Although previously undocumented, a recent retrospective case study in dogs naturally infected with E ewingii identified a relatively high proportion of dogs diagnosed with clinicopathologic markers of kidney disease and immune-mediated hemolytic anemia, medical conditions not previously associated with E ewingii infection in dogs, which supports further investigation of E ewingii as a causative agent of kidney disease secondary to infection in dogs.¹⁰

Understanding the risk of exposure to ticks infected with E canis could result in earlier detection of CKD and improved compliance with use of preventives and indicated therapies. Because patients with vector-borne disease (VBD) may have other systemic issues or chronic lameness resulting in decreased muscle mass, Cr concentrations that are dependent on muscle mass may underestimate glomerular filtration rate, especially in chronic forms of disease.^11–13 Traditional biomarkers for kidney function such as Cr and blood urea nitrogen (BUN) in some cases delay identification of an early decline in glomerular filtration rate (GFR) and early diagnosis of kidney disease.¹¹ Symmetric dimethylarginine (SDMA) has been shown to be well correlated to GFR and helps diagnose CKD earlier in dogs.^12,14 In this study we used both Cr and SDMA to define CKD, relying on the combination of biomarkers to better describe early decline in kidney function.

The purpose of the study was to evaluate whether Ehrlichia spp. antibody status is associated with an increased incidence of development of CKD. A retrospective matched cohort study was performed to determine if dogs with detectable antibodies to Ehrlichia spp. in E canis–endemic regions had an increased incidence of CKD over those without exposure to Ehrlichia spp.

Materials and Methods

A commercial laboratory’s patient result databases^a,b were used to source data for this retrospective study. Each sample was obtained and submitted to a commercial reference laboratory or analyzed in-house by a practicing veterinarian during the normal diagnostic workup and monitoring of clinically well and ill dogs in his or her care. All samples were obtained with the consent of the pet owner. To ensure privacy, additional demographic information on the pet, pet owner, or veterinarian who submitted the sample was not collected. Patient selection included dogs tested in the United States, having a minimum of one SNAP 3Dx^c, SNAP 4Dx^d, or SNAP 4Dx Plus^e test result or 4Dx Plus enzyme-linked immunosorbent assay for Ehrlichia, at least one IDEXX SDMA Test^f result, and one Cr result. SDMA was determined using a commercially available high-throughput immunoassay. The SNAP 3Dx, SNAP 4Dx, or SNAP 4Dx Plus tests determine the presence of antibody to Ehrlichia spp. using enzyme-linked immunosorbent assay technology for immunoglobulin M and immunoglobulin G antibodies to the p30/p30-1 region of the outer membrane. This test cannot differentiate between antibodies to E canis, E ewingii, or E chaffeensis. When further species identification is needed, blood smears for organism morphology and cell line affected and/or polymerase chain reaction (PCR) can be performed. Cr concentrations were determined by a colorimetric method, Jaffe’s reaction using picrate at alkaline pH^g. Urine specific gravity (USG) results were also collected but not required for study inclusion. Data used in the study to determine patient exposure to tick-borne disease were obtained from a combination of databases between January 1, 2003, and January 1, 2018. Data used to define kidney disease were collected from the reference laboratories database between July 13, 2015, and January 1, 2018. Date ranges for both data sets were determined by availability of historical data.

Regions for observation were established using geography, number of veterinary practices, and volume of lab work submitted. This allowed for the inclusion of a spectrum of case densities in each region, which ensured that overall cohort populations remained constant across all regions. Patients exposed to Ehrlichia spp.–infected ticks were defined as having a minimum of one antibody-positive Ehrlichia spp. test result in their available history. Patients defined as unexposed had only antibody-negative Ehrlichia spp. results in their available history, with a minimum of one result. Exposures and outcomes were required to occur sequentially. CKD was defined as concurrent increased SDMA (>14 μg/dL) and Cr (>1.5 mg/dL) for a minimum of 25 days with inappropriate USG (<1.030) during that time. Neither SDMA, Cr, nor USG levels could return to normal ranges once increased in the available patient history. Regardless of SDMA or Cr results, the absence of a USG result or USG ≥1.030 resulted in classifying the patient as CKD negative.

The E canis–endemic regions of the United States included in this study were identified using real-time PCR results^a from January 2011 to June 2017 (Figure 1). These geographic regions had a greater than 8:1 ratio of E canis to E ewingii PCR-positive samples and were known to have R sanguineus present as a vector for E canis.¹⁵

View Figure

• Download as Powerpoint
• Download Figure

Each patient meeting the definition of Ehrlichia spp. exposure within the endemic region was matched with four unexposed patients. A 1:4 matching scheme was used to increase the statistical efficiency of the study as the overall positive rate of Ehrlichia spp. results was low (<5%) within the region. Suitable matches were found through propensity score matching¹⁶ using caliper nearest neighbor matching without replacement. Covariates used to estimate propensity scores, and therefore controlled for in the study, were age, breed, and region. Covariate distributions of exposed and unexposed groups were compared after matching to validate match quality using the χ² test (breed and region) and the Kolmogorov-Smirnov test (age). The total number of patients included in the study after matching was 22,440. The ratio of CKD risk in exposed patients to CKD risk in unexposed patients, known as relative risk, was then calculated. Statistical significance of this estimate was defined by P value < 0.05, and all analyses were performed using the R statistical computing software^h. For clarity, a diagram of the study population appears in Figure 2.

View Figure

• Download as Powerpoint
• Download Figure

Results

The canine population in this study ranged in age from <1 to 25 yr and included all sexes and breeds (Figure 3). A Kolmogorov-Smirnov test did not find a significant difference (P = 1.00) in the age distributions of the exposed and unexposed groups. Mixed/unknown breed was the most common dog in the study, making up 37.22% of the study population. There was a total of 183 named breeds, with the 5 most common breeds in the study population being Labrador retriever (8.4%), Chihuahua (4.1%), German shepherd dog (3.8%), golden retriever (3.0%), and standard poodle (2.6%). Other named breeds made up 33.4% of the population (Table 1). A χ² test on the distribution of breeds in the exposed and unexposed groups did not detect a significant difference (P = 1.00). The distribution of dogs by state within the endemic region is found in Table 2. A χ² test did not detect a significant difference (P = 0.95) in the distributions of regions between the exposed and unexposed groups. The overall prevalence of CKD in the overall study population was 0.45% (101/22,440); in the Ehrilichia-exposed group was 0.78% (35/4488); and in the Ehrlichia-unexposed group was 0.37% (66/17,952). The relative risk of CKD for patients within E canis–endemic areas with serological evidence of Ehrlichia infection was estimated to be 2.12 (95% confidence interval [1.35–3.15], P = 0.0006); a 112% greater risk over those in the unexposed cohort (Table 3).

View Figure

• Download as Powerpoint
• Download Figure

TABLE 1 Distribution of Dog Breeds, After Matching, in Total Study Cohort

View Fullscreen

TABLE 2 Distribution of Dogs by State, After Matching, in Total Study Cohort

View Fullscreen

TABLE 3 Contingency Table

View Fullscreen

Discussion

This study supports the association of Ehrlichia spp. infection and kidney disease. Dogs that test antibody positive for Ehrlichia spp. in E canis–endemic areas have a statistically significant increased risk of developing CKD. The reported prevalence of CKD within the study cohort is small relative to previously reported estimates of 0.5–1.0% prevalence of CKD in dogs.¹⁷ This is likely due to inherent study design factors such as the stringency of the CKD definition and the stipulation that the outcome of CKD occurs within a two-and-a-half-year period for which data were available. Although this prevalence is not representative of the broader population, it is equivalent for both exposure groups. It should be noted that although E canis is found in higher frequency in certain geographic regions of the United States (Figure 1), dogs infected with E canis can be found throughout the United States.

This study was designed to determine if there was an association between ehrlichiosis and CKD. Although this single study does not conclude causality, the significant association identifies an at-risk population and justifies further investigation into a potential causal relationship. Several design decisions were made to avoid overestimation of risk between the exposed group and the CKD outcome; however, they may contribute to underestimation of the strength of the relationship in this study. The stringency of the CKD definition may foster survival bias in the unexposed group, if dogs are more likely to be excluded from the study owing to acute disease before they meet the CKD definition. This would result in their inclusion in the non-CKD outcome group and cause underestimation of the relative risk of CKD. The criteria used in this study to define CKD may underestimate the patients with CKD given that the absence of a USG result or USG ≥1.030 resulted in classifying the patient as CKD negative. The relatively short time frame used to define CKD coupled with the strict requirements for CKD classification likely underrepresents the number of CKD patients in both groups. The stringent inclusion criteria surrounding result history biases the study toward well-cared-for pets. The benefit of the stringent definition of CKD was increased confidence in diagnosis of CKD versus an acute kidney injury or single spurious result; thus, there is strong confidence in classification of CKD in this study. Patient movement into and out of the E canis region could also bias these results, because the tick exposure that resulted in the Ehrlichia spp.–positive result could have occurred outside the region defined in the study (Figure 1).^18–20 Additionally, the endemic region’s climate, geology, or strain of E canis may limit the extrapolation of conclusions. Follow-up studies could mitigate this concern.^5,21 Although a retrospective matched cohort study of this design is extremely powerful for identifying an exposure and measuring associated risk factors for an outcome, there are several limitations inherent to the study. Misclassification of exposure or outcome status as a result of missing data, loss to follow-up, or patient misidentification may introduce bias. In an ideal study, lifetime outcomes for all dogs would be available.

This relative risk of 2.12 (95% confidence interval [1.35–3.15]) reveals a population of dogs that are at risk of developing CKD. Accordingly, any patient that tests positive for Ehrlichia spp. should be considered for comprehensive evaluation and extended and more frequent biochemical monitoring. This monitoring should have a strong focus on kidney health. Historically, markers such as Cr, BUN, and USG have best estimated a patient’s kidney function. In patients with ehrlichiosis, systemic disease and chronic lameness have the potential to result in decreased muscle mass and may reduce confidence in Cr concentrations as an accurate estimator of renal function.¹¹ In this study we used Cr, SDMA, and USG to define CKD, relying on the combination of biomarkers to better describe early decline in kidney function with less concern for extrarenal factors. SDMA has been shown to be well correlated to GFR in dogs and an earlier indicator of renal disfunction than Cr for chronic kidney disease.¹² A persistent increase in SDMA, with or without concurrent increases in Cr, in a patient seropositive for Ehrlichia spp. warrants further investigation to evaluate for ongoing kidney injury. The authors recommend that dogs with a newly positive Ehrlichia spp. antibody test should receive a complete laboratory assessment consisting of a complete blood count, biochemical evaluation with SDMA, and urinalysis with sediment examination to be instituted with a regular preventive care protocol.

Conclusion

This study identified an association between dogs with positive Ehrlichia spp. test results in endemic areas and a higher incidence for CKD. Dogs with a positive Ehrlichia spp. antibody test result in E canis–endemic areas had a 112% greater risk of developing CKD.

Screening patients exposed to ticks with a biochemical panel that includes SDMA, Cr, and BUN and performing a urinalysis can help identify kidney disease that may be a secondary complication of VBD, such as ehrlichiosis. In addition to annual VBD screening, these seropositive patients should receive a physical examination, a complete blood count and biochemical panel with SDMA, and a complete urinalysis with sediment to monitor for multisystemic disease.