A retrospective study of vector‐borne disease prevalence in dogs with proteinuria: Southeastern United States

Abstract Background Proteinuria is a risk factor for progressive kidney injury in dogs. Enhanced understanding of potential associations between canine vector‐borne diseases (CVBD) and proteinuria is needed. Objectives To determine the proportion of evaluated proteinuric dogs exposed to ≥1 CVBD, including Babesia spp., Ehrlichia spp., spotted‐fever group Rickettsia, Bartonella spp., Anaplasma spp., hemotropic Mycoplasma spp., Borrelia burgdorferi, and Dirofilaria immitis, and to determine if demographic or clinicopathologic differences exist between proteinuric dogs exposed to CVBD versus proteinuric dogs with no evidence of CVBD exposure. Animals Two‐hundred nine proteinuric dogs, concurrently tested for CVBD, which were examined at a single academic veterinary hospital between January 2008 and December 2015. Methods Retrospective cross‐sectional study. Demographic, clinicopathologic, and CVBD test results were extracted from medical records. A multivariable logistic regression model was used to assess associations between CVBD and selected variables. Results Based on serology and polymerase chain reaction testing, 34% of proteinuric dogs (72/209) were exposed to ≥1 CVBD. Exposure to Rickettsia spp. (19%), Ehrlichia spp. (12%), and B. burgdorferi (9%) were most common. The CVBD exposure was lower in dogs tested in autumn or spring, higher in intact dogs, and higher in dogs with lower serum albumin and higher serum creatinine concentrations. Conclusions and Clinical Importance Exposure to CVBD, particularly exposure to Rickettsia spp., Ehrlichia spp., and B. burgdorferi was found in proteinuric dogs from the southeast United States. Additional controlled prospective studies examining a potential causal relationship between CVBD and proteinuria are warranted.


| INTRODUCTION
Proteinuria is an important clinical finding and is not only a marker of nephropathy but also a prognostic factor for progression of chronic kidney disease. [1][2][3][4][5][6][7][8] Proteinuria is related to decreased survival in azotemic and non-azotemic dogs, and decreasing proteinuria improves survival. 1,2,4,9 Because a variety of disease processes can cause or contribute to proteinuria, not all of which can be treated, it is important to identify opportunities for clinical interventions that may prevent progression of kidney injury in proteinuric dogs. Because of the possibility for chronic persistence of canine vector-borne disease (CVBD) pathogens within the bloodstream or tissues, the potential role of CVBD as a cause of, or contributor to, proteinuria in dogs requires investigation.
Clinicians attempting to prevent disease progression may elect to treat proteinuric dogs with suspected or confirmed glomerular disease based on indirect evidence of comorbidities that may be leading to immune complex formation, such as positive serology for CVBD.
When searching for an underlying infectious disease as a component of the evaluation for proteinuric dogs, clinicians are advised to use their clinical judgment and be cognizant of potential CVBD exposures that may be unique for each patient. 10,18 Although associations between IMGN and certain CVBDs have been reported, the prevalence of, and risk factors for, infection with these organisms in dogs with proteinuria have remained largely unstudied. Epidemiologic studies of CVBD exposure in proteinuric dogs can provide important information on the prevalence of diseases in certain demographic groups and locations, and the seasons when transmission or disease expression are most common.
Additionally, although coinfection with multiple vector-borne pathogens can cause more severe disease manifestations, this association has not been investigated specifically for proteinuric dogs. 19,20 The purpose of our retrospective study was to describe the CVBD exposures among proteinuric dogs. The first aim was to determine the proportion of evaluated proteinuric dogs exposed to each CVBD of interest, including B canis, Babesia gibsoni, E canis, spotted-fever group Rickettsia, Bartonella vinsonii spp. berkhoffii, Bartonella henselae, Bartonella koehlerae, Anaplasma spp., hemotropic Mycoplasma spp., B burgdorferi (Lyme disease), and Dirofilaria immitis (heartworm disease). The second aim was to determine if demographic or clinicopathologic differences existed between proteinuric dogs exposed to CVBD (P-CVBD+) and proteinuric dogs that lacked serological or polymerase chain reaction (PCR) evidence of CVBD exposure (P-CVBD−). Our hypothesis was that there would be no significant differences in demographics or clinicopathologic variables between these groups.

| Variables and data sources
When available, the following information was extracted from the medical records of the 209 proteinuric dogs included in the study: age, sex, breed, clinicopathologic data (CBC, serum biochemistry profile, urinalysis, and UPC ratio) and CVBD serology and PCR test results. Complete blood count, serum biochemistry profiles, urinalysis, and UPC ratios were performed by the clinical pathology laboratory at NCSU. For the purpose of statistical analyses, dogs that met the inclusion criteria were separated into 2 groups: proteinuric dogs with ≥1 serological or PCR-positive CVBD test results indicating CVBD exposure (P-CVBD+) and proteinuric dogs with no serological or PCRpositive CVBD test results (ie, every CVBD test result was negative), indicating no CVBD exposure (P-CVBD−).

| Statistical analysis
Descriptive statistics were obtained for proteinuric dogs with and without CVBD exposure as defined by the previously described testing methods. The Shapiro-Wilk test for normality was used, and continuous variables with normal distribution were described using mean and standard deviation, whereas nonnormally distributed continuous variables were described using median and interquartile range. Differences between P-CVBD+ and P-CVBD− groups were calculated using On the basis of serology or PCR testing, 34.4% of proteinuric dogs (72/209) were exposed to ≥1 CVBD ( Figure 2,    There were 23 proteinuric dogs exposed to ≥1 CVBD pathogen  ) seroreactive against B burgdorferi, by breed. Error bars represent 95% confidence intervals for percent estimates test results. Therefore, some dogs in the P-CVBD− group may have been misclassified as not exposed to CVBD pathogens. Despite the substantial lack of Bartonella spp. IFA and PCR diagnostic sensitivities, 5 proteinuric dogs were exposed to a Bartonella spp. 40 A potential role for Bartonella spp. in the pathogenesis of proteinuria has not been investigated in dogs or humans, but glomerulonephritis associated with Bartonella spp. endocarditis has been documented in human case reports of humans. 41,42 Among the 72 proteinuric dogs with CVBD exposure, 31.9%, had been co-exposed to ≥2 CVBD pathogens. Some arthropods are competent vectors for transmission for >1 CVBD pathogen, and it is not unusual for dogs to be exposed to several different arthropod species simultaneously or sequentially. 16 When compared to infection with individual pathogens, coinfections with multiple vector-borne pathogens can contribute to complex disease expression resulting in diagnostic, therapeutic, and medical management challenges. 27 The extent to which coinfections might contribute to proteinuria has not been systematically studied, but the high prevalence of co-exposures and the trend toward more severe disease manifestations in the co- In recent years, access to rapid, in-hospital screening tests has allowed veterinarians to assess antibodies to selected CVBD pathogens.
In our study, 172 of 204 dogs were SNAP 4DX-negative, of which 39 were determined to have CVBD exposure based on IFA or PCR testing. If only the rapid assay ELISA had been used to screen the dogs in our study, only 32 dogs would have been identified as CVBD-positive (rather than 71), missing 55% of truly CVBD-exposed dogs. These findings support the recommendation for more comprehensive serological and PCR testing rather than relying solely on the SNAP 4DX ELISA to facilitate an accurate microbiological diagnosis of CVBD pathogen exposure in proteinuric dogs. 32 Previous research has shown that concurrent use of PCR and serological assays in parallel is likely to increase the detection or exposure of CVBD pathogens. 46 Retesting, particularly once immunosuppressive treatment has been instituted, also can increase diagnostic PCR documentation of CVBD infection. 47 Failure to diagnose a CVBD pathogen in a dog with proteinuria could result in a diagnosis of idiopathic IMGN, immunosuppressive drug treatment, and potentially an incomplete treatment response with an adverse clinical outcome.
Our study was subject to the limitations inherent in any retrospective analysis of patient medical records and results. There was no standardization of data collection or recording at the time of patient evaluations. Also, there was variability in diagnostic testing among individual dogs based on test request selection by the attending clinician. Variability in requested CVBD tests may be associated with potential clinician biases in test selection, as well as laboratory changes in test offerings over the study time frame and diagnostic improvements in assay sensitivity and specificity. For example, many seroreactive dogs did not have PCR testing requested. Therefore, the associations between CVBD and demographic and clinicopathologic variables found here reflect only associations with exposure, not necessarily active infection, with these pathogens. Because of regional variations among tick species and CVBD pathogens, the results of our study should not be extrapolated to all regions of the United States or elsewhere.
We attempted to exclude cases of renal tubular proteinuria (ie, dogs lacking serum albumin concentrations or having normal serum albumin concentration with a UPC ratio <1). This decision was based on general acceptance that UPC ratio >0.5 indicates renal proteinuria when urine lacks evidence of inflammation or macroscopic hematuria, 1-3 and UPC ratio ≥2.0 is considered a magnitude of proteinuria sufficiently high to support pathologic proteinuria. 3 Furthermore, all patients were evaluated by a minimum of 2 faculty clinicians and 1 resident at the time of initial diagnostic evaluation. Each case then was reviewed further by a senior internal medicine resident for inclusion in the study. Despite these attempts to include only patients with glomerular proteinuria, it is possible that we captured a subset of dogs with UPC ratio between 0.5 and 2.0 in which proteinuria was not caused by glomerular injury but was by another cause of renal proteinuria such as tubulointerstitial disease. Ultimately, without matching renal biopsy samples for each patient included, the location of renal injury leading to proteinuria cannot be definitively identified. 39 In addition, IFA and PCR diagnostic testing for CVBD both have limitations. Both antibodies and infection can persist for variable intervals depending upon the pathogen and variations in host immune response.
In addition, serology does not confirm active or persistent infection. 10 Conversely, molecular techniques such as PCR detect active infection, but low levels of intravascular infection are common among CVBD pathogens. Therefore, PCR testing at a single time point may produce falsenegative PCR results for a dog infected with ≥1 CVBD pathogens. 10 Because treatment history in these proteinuric dogs was unknown, there is likely also a subset of proteinuric dogs that received antibiotics before CBVD testing, potentially decreasing PCR sensitivity.
Overall, CBVD exposure, particularly exposure to Rickettsia and Ehrlichia species and B burgdorferi, was frequent among proteinuric dogs in our study. Establishing a direct cause-and-effect relationship between positive CVBD test results and proteinuria will require both controlled, prospective epidemiologic studies as well as experimental studies. 8 With recent standardizations in renal histopathology, interpretation of renal biopsy results obtained from proteinuric dogs may allow for more precise interpretation of CVBD test results. 48 Collectively, information derived from our study should improve the current understanding of CVBD exposure in proteinuric dogs and facilitate the design of future prospective studies targeting individual CVBD pathogens or CVBD coinfections in the development or progression of proteinuria.