Clinical Features of Epistaxis in Dogs: A Retrospective Study of 35 Cases (1999–2002)
Epistaxis was retrospectively evaluated in 35 dogs. Systemic disease was diagnosed in seven dogs and intranasal disease in 29. Nineteen dogs with intranasal disease had neoplasia. Dogs with neoplasia were older (mean 10.0 years) than dogs with nonneoplastic intranasal disease (mean 5.6 years). Signs persisting for >1 month occurred more often in dogs with intranasal than systemic disease. Unilateral epistaxis did not distinguish intranasal from systemic disease. Only dogs with intranasal disease had facial deformity, decreased airflow, or regional sub-mandibular lymphadenopathy. Dogs with systemic disease had a lower packed cell volume (mean 31.8%) than dogs with intranasal disease (mean 42.7%).
Introduction
Epistaxis is defined as hemorrhage from the nose.1,2 The hemorrhage can result from local disease within the nasal cavity or from systemic disorders resulting in hemostatic abnormalities. Diseases of the nasal cavity that can result in epistaxis include neoplasia, fungal infections, trauma, foreign bodies, nasal parasites (Pneumonyssoides and Capillaria spp.), inflammatory rhinitis, and advanced dental disease.3–10 Systemic disorders that can cause epistaxis include thrombocytopenia from infection, immune-mediated disorders, myelophthisis, or consumptive processes; thrombocytopathies from drugs (e.g., aspirin), uremia, or von Willebrand’s disease; inherited clotting factor abnormalities; acquired clotting factor disorders (e.g., liver failure, rodenticide toxicity); vasculitides; and hyperviscosity syndrome.1,9 Systemic hypertension has been implicated as a potential cause of epistaxis in people, although its direct role is controversial.11–15
Because of this long and varied list of potential causes, the diagnostic evaluation of dogs with epistaxis can be extensive and costly. Testing for systemic etiologies often includes a complete blood count (CBC), platelet count, serum biochemical panel, urinalysis, funduscopic examination, systemic blood pressure (bp) measurement, coagulation profile, von Willebrand’s factor assay, buccal mucosa bleeding time, and serology for tick-borne organisms. Testing for nasal cavity diseases can include nasal radiography, computed tomography (CT), magnetic resonance imaging, rhinoscopy, nasal biopsy, nasal culture, serology for fungal infections, and fecal examinations for parasites.9,16
The purposes of this study were to retrospectively review records of dogs that were presented with the complaint of epistaxis to determine the frequency of different etiologies, to evaluate the presenting signs, and to identify abnormalities in basic laboratory data that were useful in prioritizing the differential diagnoses of epistaxis.
Materials and Methods
Medical records were reviewed from dogs presented to the North Carolina State University Veterinary Teaching Hospital (NCSU-VTH) with a primary complaint of epistaxis, nasal discharge, or sneezing from 1999 through 2002. Dogs were included in the study if at least one episode of epistaxis was documented in the medical record and if 1) a final diagnosis of a specific systemic disorder was obtained that was consistent with epistaxis, or 2) a diagnostic evaluation for intranasal disease was performed that included nasal radiography or CT, rhinoscopy, and nasal biopsy.
Diagnostic Criteria
For inclusion in the study, specific diagnostic criteria were used to confirm the presence of certain systemic diseases. A diagnosis of ehrlichiosis or bartonellosis required a serum indirect fluorescent antibody titer of ≥1:64. A diagnosis of Rocky Mountain spotted fever required rising serum antibody titers, with at least a fourfold increase in titer between acute and convalescent samples. A diagnosis of immune-mediated thrombocytopenia (IMT) required a platelet count <50,000/μL without evidence of underlying diseases (e.g., neoplasia, infection). A diagnosis of rodenticide toxicity required prolongation of the prothrombin time (PT) and partial thromboplastin time (PTT) and a positive response to treatment with vitamin K. A diagnosis of hypertension required a systolic bp >180 mm Hg. A diagnosis of von Willebrand’s disease (VWD) required confirmation by von Willebrand factor antigen assay.
Diagnostic criteria were also created for classifying intranasal disease. A diagnosis of neoplasia was confirmed by cytological or histopathological examination of specimens from the nasal cavity or submandibular lymph nodes, or by CT abnormalities considered by a board-certified radiologist at the time of initial evaluation to be indicative of neoplasia.17–19 A diagnosis of fungal rhinitis was made by visualization of fungal plaque(s) and/or intranasal erosion upon rhinoscopy and supportive cytology, histopathology, culture, or CT results. A diagnosis of parasitism was made by visualization of parasites upon rhinoscopy or identification of ova in the feces. A diagnosis of idiopathic rhinitis was based on histopathological findings of inflammation and the absence of evidence of other etiologies on nasal imaging or rhinoscopy. The diagnosis was undefined when there was histopathological evidence of inflammatory rhinitis; but radiographic, CT, or rhinoscopic findings were suggestive of a more aggressive disease (e.g., neoplasia, fungal rhinitis).
Data Reviewed
Data retrieved from the medical records included signalment, history, physical examination findings, diagnostic tests, and diagnosis. Specific attention was paid to the duration of epistaxis or other nasal signs; the character of any previous or concurrent nasal discharge; whether the epistaxis was unilateral or bilateral; the detection of decreased airflow through the nares; the presence of facial or palate deformities, submandibular lymphadenopathy, or dental disease; and the presence of funduscopic abnormalities. Other signs of nasal disease and evidence of hemorrhage elsewhere in the body were also noted. For the purposes of this study, clinical signs of ≤1 month in duration were defined as “acute,” and signs >1 month in duration were defined as “chronic.” Results of all CBCs, serum biochemical profiles, urinalyses, and tests related to hemostasis were also scrutinized.
Statistical Analysis
Statistical comparisons were performed for numerical data using a Mann-Whitney rank sum test (M-W test) for data that failed tests for normality or equal variance or a t-test (α=0.050). Analysis of tabular data was performed using Fisher’s exact test or chi-square analysis. Values of P≤0.05 were considered significant.a
Results
Signalment
Thirty-five dogs met the criteria for inclusion in the study. The dogs ranged in age from 2.5 to 13.5 years (mean 8.0 years; median 8.0 years). Thirty (86%) dogs were ≥5 years of age. Body weights ranged from 3.6 to 67 kg (mean 29.2 kg). There were 17 females (all spayed) and 18 males (12 castrated) in the study. Breeds represented included the Labrador retriever (n=3), German shepherd dog (n=2), golden retriever (n=2), and one each of the following breeds: Boston terrier, Treeing Walker coonhound, Doberman pinscher, Old English sheepdog, German wirehaired pointer, Kuvasz, Pomeranian, Chinese shar pei, spitz, Saint Bernard, and Walker hound. Additionally, there were 17 mixed-breed dogs.
Diagnoses
Systemic diseases were diagnosed in seven (20%) dogs. Serological evidence of infection with tick-borne pathogens was identified in three dogs. Two dogs were seroreactive to both Ehrlichia canis (E. canis) and Bartonella (B.) vinsonii (berkhoffii). One dog was seroreactive to B. vinsonii (berkhoffii) but also had a nasal carcinoma. This dog was considered to have both systemic and intranasal disease. One dog was diagnosed with IMT and one with VWD. One dog had neutrophilic vasculitis, presumably secondary to sepsis, based on systemic signs and histopathology of a skin lesion. One dog was diagnosed with bacterial endocarditis and septicemia based on blood culture and echocardiography. Systemic hypertension (systolic bp=187 mm Hg) was found in one dog that had a histopathological diagnosis of inflammatory rhinitis with ulceration. Because the ulcerative rhinitis was more likely the cause of epistaxis than the modest hypertension, the dog was included in the group with intranasal diseases.
Intranasal diseases were diagnosed in 29 (83%) dogs, one of which was previously described as seroreactive to B. vinsonii (berkhoffii). Neoplasia was diagnosed in 19 (66%) of these dogs. Nonneoplastic nasal disease was diagnosed in nine (31%) of these dogs. The diagnosis was undefined in one dog. Of the dogs with neoplasia, 12 had carcinomas (carcinoma [n=7], adenocarcinoma [n=4], and squamous cell carcinoma [n=1]), and four had sarcomas (fibrosarcoma [n=1], chondrosarcoma [n=1], myxomatous sarcoma [n=1], and poorly differentiated sarcoma [n=1]). Three dogs were presumptively diagnosed with neoplasia based on CT findings. Of the dogs with nonneoplastic intranasal disease, five had idiopathic rhinitis, two had fungal rhinitis, and two had capillariasis.
The dog with both nasal carcinoma and bartonellosis was eliminated from further comparisons because of the inability to attribute the clinical abnormalities to specific disease. At the time of referral, the affected dog was a 6-year-old, spayed female, Old English sheepdog that had intermittent bouts of epistaxis since 1 year of age. Seroreactivity to B. vinsonii (berkhoffii) antigens was documented 1 month prior to referral (titer 1:256) and at the time of referral (1:128), when a histological diagnosis of nasal carcinoma was also obtained.
Influence of Signalment
All genders were represented in both the systemic and intranasal disease groups, and no statistical differences were found between the groups. Of the dogs with systemic disease, two were spayed females and four were males (two were castrated). Of the dogs with intranasal disease, 14 were females (all spayed) and 14 were males (10 were castrated). There were no obvious associations of systemic versus intranasal disease with breed or body weight. Dogs with systemic disease tended to be younger than dogs with intranasal disease, but this finding was not statistically significant (t-test, P=0.058). Dogs with systemic disease ranged in age from 2.5 to 10 years (mean 5.7 years), while dogs with intranasal disease ranged in age from 2.5 to 13.5 years (mean 8.6 years).
Clinical Signs
Intranasal disease was associated with a chronic history of epistaxis. Epistaxis was chronic in 16 of 28 dogs with intranasal disease, but in only one of six dogs with systemic disease (Fisher’s exact test, P=0.022). An additional six dogs with intranasal disease had chronic nasal signs (including sneeze, snorting or stertor, and nasolacrimal duct obstruction), but their epistaxis was <1 month in duration. Of eight dogs with previous or concurrent mucoid or mucopurulent nasal discharge, one dog had systemic disease (IMT), and seven dogs had intranasal disease. A greenish black discharge was reported in an additional dog with intranasal disease.
Of 31 dogs in which the side of epistaxis was recorded, 29 had a history of either unilateral involvement (n=22) or unilateral involvement that progressed to bilateral involvement (n=7). No statistically significant differences could be found between dogs with systemic disease and dogs with intranasal disease with regard to the epistaxis being unilateral, bilateral, or unilateral progressing to bilateral. Among dogs with systemic disease, the epistaxis was described as unilateral (n=3), unilateral progressing to bilateral (n=1), bilateral (n=1), and not defined (n=1). Among dogs with intranasal disease, the epistaxis was described as unilateral (n=19), unilateral progressing to bilateral (n=6), bilateral (n=1), and/or not defined (n=2). All dogs with a history of sneezing (n=11) or reverse sneezing (n=2) had intranasal disease.
Physical Findings
On physical examination, there was evidence of hemorrhage at body sites other than the nose in two dogs with systemic disease. One dog with intranasal disease had melena, but the melena was presumed to be secondary to swallowing nasal blood. Five dogs had facial or palate deformities, and eight had decreased airflow from one or both nostrils. Of these dogs, two had both a deformity and decreased airflow. All of the dogs with a deformity or decreased airflow had intranasal disease. Generalized lymphadenopathy was noted in three dogs, two with systemic disease and one with intranasal disease. Regional submandibular lymphadenopathy was noted in nine dogs, all with intranasal disease. Dental disease was found in 14 dogs, one with systemic disease and 13 with intranasal disease. Funduscopic examinations were normal in 15 dogs (four with systemic disease, 11 with intranasal disease) and abnormal in two dogs. A focal area of retinal hemorrhage was seen in the dog with vasculitis, and mild papilledema was noted in a dog with intranasal neoplasia. Additional physical examination abnormalities included depigmentation and erosion of the nasal planum in one dog with fungal rhinitis.
Laboratory Findings
Analysis of systemic clotting tests showed no statistical differences in platelet counts between dogs with systemic disease and dogs with intranasal disease (M-W test). However, two dogs with systemic disease (one with IMT, one with vasculitis) had platelet counts <50,000/μL. The lowest platelet count from dogs with intranasal disease was 142,000/μL. Coagulation profiles which included PT and PTT, fibrin degradation products (FDPs), and fibrinogen were measured in 21 dogs. Prothrombin times and PTTs were normal in all 21 dogs (six with systemic disease, 15 with intranasal disease). Fibrin degradation products were detected (≥5 μg/mL) in six dogs, two with systemic disease and four with intranasal disease. Fibrinogen was increased in five dogs, three with systemic disease and two with intranasal disease. Buccal mucosa bleeding time was performed in nine dogs and found to be normal in eight dogs (three with systemic disease and five with intranasal disease) and abnormal (22 minutes; normal <4 minutes) only in the dog with VWD.20
Analysis of basic laboratory test results revealed that dogs with systemic disease had a lower packed cell volume (PCV) than dogs with intranasal disease (t-test, P<0.001). The mean PCV of dogs with systemic diseases was 31.8% (range 20% to 39%; reference range 33% to 58%). The mean PCV from dogs with intranasal diseases was 42.7% (range 32% to 52%). Globulin concentrations were not statistically different between the two disease groups (M-W test). One dog with ehrlichiosis and bartonellosis was hyperglobulinemic (8.5 g/dL; reference range 2.6 to 4.4 g/dL). One dog with idiopathic rhinitis was mildly hyper-globulinemic (4.6 g/dL).
Antibodies were not detected to E. canis or B. vinsonii (berkhoffii) antigens in seven of 10 dogs tested (three with systemic disease, four with intranasal disease). Systolic bp was normal in 11 dogs (four with systemic disease, seven with intranasal disease). Hypertension was diagnosed in one dog with idiopathic rhinitis (systolic bp 187 mm Hg).
Intranasal Disease
Records from dogs with intranasal disease were further evaluated for presenting signs or abnormalities in laboratory data that would assist in the discrimination of neoplastic from nonneoplastic disease. The dog with neoplasia and bartonellosis and the dog with undefined intranasal disease were omitted from this analysis. Data from 27 dogs with intranasal disease (18 with neoplasia, nine with nonneoplastic disease) were analyzed. All genders were represented in both neoplastic and nonneoplastic groups, and no statistical difference was found between the groups. Of the dogs with neoplastic disease, 11 were females (all were spayed) and seven were males (five were castrated). Of the dogs with nonneoplastic disease, three were females (all were spayed) and six were males (four were castrated). No obvious associations between disease groups and breed or body weight were noted. Dogs with neoplastic disease were significantly older than dogs with nonneoplastic disease (t-test, P<0.001). Dogs with neoplastic disease ranged in age from 2.5 to 13.5 years (mean 10.0 years). Dogs with nonneoplastic disease ranged in age from 2.5 to 9 years (mean 5.6 years).
No statistically significant differences were found between the dogs with neoplasia and dogs with nonneoplastic disease with respect to chronicity of signs, presence of mucoid or mucopurulent nasal discharge, sidedness of epistaxis, or sneezing. Of the 18 dogs with neoplasia, 12 had chronic epistaxis and three had acute epistaxis with other chronic nasal signs. Of the nine dogs with nonneoplastic disease, three had chronic epistaxis and two had acute epistaxis with other chronic nasal signs. Previous or concurrent mucoid or mucopurulent nasal discharge was noted in three dogs with neoplasia and in four dogs with nonneoplastic disease. A greenish-black discharge was reported in one additional dog with nonneoplastic disease. Of the dogs with neoplasia, the epistaxis was described as unilateral in 14 dogs and as unilateral progressing to bilateral in four dogs. Of the dogs with nonneoplastic disease, the epistaxis was described as unilateral in four dogs, unilateral progressing to bilateral in two dogs, bilateral in one dog, and undefined in two dogs. Sneezing was reported in seven dogs with neoplasia and in four dogs with nonneoplastic disease. Reverse sneezing was noted in one dog from each group.
On physical examination, the five dogs with facial or palate deformities all had intranasal neoplasia. Of the eight dogs with decreased airflow from one or both nostrils, seven had neoplasia and one had nonneoplastic disease. Submandibular lymphadenopathy was present in seven dogs with neoplasia and in two dogs with fungal rhinitis.
There was no statistically significant difference in platelet counts between the neoplastic and nonneoplastic groups (t-test). Interestingly, four dogs with neoplasia had platelet counts in excess of 500,000/μL (551,000/μL; 647,000/μL; 700,000/μL; 715,000/μL). The highest platelet count from the dogs with nonneoplastic disease was 466,000/μL. Detectable FDPs (≥5 μg/mL) were found in two dogs with neoplasia and in two dogs with nonneoplastic disease. Fibrinogen was increased in one dog with neoplasia and in one dog with nonneoplastic disease. There was also no statistically significant difference in PCV or globulin concentrations between the two groups (t-test). At the time of presentation, the mean PCVs were 43.5% in dogs with neoplasia (range 32% to 52%) and 42.4% in dogs without neoplasia (range 33% to 47%). At the time of presentation, the mean globulin concentrations were 3.3 g/dL (range 2.6 to 3.9 g/dL) in dogs with neoplasia and 3.5 g/dL (range 2.8 to 4.6 g/dL) in dogs without neoplasia. Of the four dogs with intranasal disease and negative titers for E. canis and B. vinsonii (berkhoffii), one had neoplasia and three had idiopathic rhinitis.
Discussion
In the study reported here, systemic disorders were diagnosed in only 20% of the epistaxis cases. The animal population at NCSU-VTH and the study design likely resulted in systemic disorders being underrepresented, however, and may also have contributed to the variety of systemic diseases that were identified. The animal population at NCSU-VTH is largely referral based, and primary care veterinarians may be more likely to diagnose systemic disorders associated with epistaxis, such as IMT, rodenticide toxicity, or ehrlichiosis. Referral may be less necessary to assist with the diagnosis of systemic disorders associated with epistaxis compared with intranasal diseases, for which specialized equipment and expertise are often required. Further, dogs with systemic diseases are not always referred for nasal signs, and these dogs would have been missed by the search criteria. Although not representative of an animal population at a primary care facility, the study population did allow comparisons between dogs with a variety of systemic and intranasal diseases. The population also allowed comparisons to be made between dogs with neoplastic and nonneoplastic intranasal disease.
Three of the dogs with systemic disorders had positive serology for B. vinsonii (berkhoffii). The role of Bartonella spp. in canine diseases appears to be complex and is still being investigated.21 All three dogs with bartonellosis had other diseases as well. Two had concurrent ehrlichiosis. An association between ehrlichiosis and epistaxis has been established historically since the Vietnam war.22 Coinfection with E. canis and B. vinsonii (berkhoffii) is a frequent occurrence in dogs with extensive tick exposure, even in dogs from southeast Asia.23 Concurrent infection with B. vinsonii (berkhoffii) may have been an exacerbating factor or an incidental finding. A role for B. vinsonii (berkhoffii) in epistaxis was most compelling in the third case that had a diagnosis of concurrent carcinoma. This dog had a history of intermittent epistaxis beginning at 1 year of age. The carcinoma was not identified in this dog until 5 years later. This long duration of epistaxis was more consistent with a disease such as bartonellosis than malignant neoplasia. It has been postulated that some chronic bacterial infections may cause cancer or facilitate tumor development, but the exact relationship of the two diseases in this dog was not defined.24
The majority (66%) of dogs with epistaxis from intranasal disease had neoplasia. This proportion was higher than that in previous reports of dogs with intranasal disease, and it may have indicated a stronger association between neoplasia and epistaxis than between neoplasia and other nasal signs.7,25 In a study by Forbes, et al., 35 (32%) of 109 dogs with rhinoscopy-assisted biopsy for nasal signs had neoplasia.25 In a report by Tasker, et al., 14 (36%) of 39 dogs with a definitive diagnosis had neoplasia.7 Other explanations for the strong association found between neoplasia and epistaxis in the study reported here could be sampling errors arising from relatively small study populations and differences in animal populations between the reports. The diagnosis of neoplasia in three of the dogs in this study was based on CT findings alone, and these results added to the overall number of neoplastic cases. A recent study demonstrated specificities of 100% and 98% for CT in the diagnosis of nasal neoplasia.19 Even if these three cases were excluded from consideration, the majority of dogs (16 [62%] of 26) in this study presenting for epistaxis with intranasal disease would still have been diagnosed with neoplasia. The histological types of neoplasia seen in this study were similar to those previously reported.26,27 Approximately 60% to 70% of reported canine nasal tumors are of epithelial origin.26,27
As expected, neoplasia was more likely to occur in older dogs when compared with nonneoplastic intranasal disease. Young age cannot be used to completely rule out a diagnosis of neoplasia, however. One of the youngest dogs in the study was diagnosed with carcinoma at 2.5 years of age. In a report of 285 dogs with sinonasal neoplasia, 5% were 1 to 4 years of age, while 63% were 9 to 16 years of age.26
In the study reported here, dogs with systemic disease were usually presented with epistaxis of ≤1 month in duration. Duration of signs was more chronic in dogs with intranasal disease. It should be noted, however, that three of 18 dogs with neoplasia had nasal signs of ≤1 month in duration. This finding was consistent with a report by MacEwen, et al., in which 43 dogs with nasal tumors had a duration of signs ranging from a few weeks to 6 months (median 3 months).27
Dogs with systemic disease in this study rarely had a history of mucoid or mucopurulent nasal discharge. Surprisingly, the one dog with mucopurulent discharge did not have systemic inflammatory disease, but had IMT. No dog with systemic disease had a history of sneezing or a reverse sneeze. The majority of dogs with systemic and intranasal disease (both neoplastic and nonneoplastic) had a history of either unilateral epistaxis alone or unilateral epistaxis that became bilateral. Based on the results of this study, a history of unilateral epistaxis cannot be used to rule out systemic disease.
Dogs with systemic disease did not have facial or palate deformity or obstruction of airflow; however, caution is needed when using obstruction of airflow to identify intranasal disease. One of the authors (Hawkins) has treated a dog with epistaxis associated with ehrlichiosis that had complete obstruction of airflow from one naris, presumably from an intranasal blood clot.
Dental disease was more often noted in dogs with intranasal disease than in the dogs with systemic disease, even though dental disease was not implicated as the cause of epistaxis in any of these cases. It is possible that particular attention was paid to the teeth and oral cavity when intranasal disease was suspected, because there is a known association between dental and/or oral cavity disease and nasal signs.7,9,10
Funduscopic examinations were generally normal in this study. Nevertheless, funduscopic examination should be performed in all dogs with unexplained epistaxis because of the great potential to detect lesions, such as retinal detachments, vasculitis, hemorrhages, or papilledema. Although there were no statistical differences detected in platelet counts in the dogs of this study, platelet counts are clearly indicated in these animals to identify marked thrombocytopenia as a cause for epistaxis.
The only laboratory parameter that was associated with a particular disease group was the PCV, which was lower in dogs with systemic diseases. Unfortunately, there was a great deal of overlap between disease groups, making the clinical usefulness of this parameter doubtful for an individual animal.
No evidence was found in this study to support an association between hypertension and epistaxis. The one dog with increased systolic bp also had ulcerative changes on nasal biopsy that more readily explained the epistaxis, although a contribution of hypertension cannot be ruled out. In people, the association between systemic hypertension and epistaxis is somewhat controversial.11–15,28
Conclusion
In this study of 35 dogs with epistaxis, epistaxis of acute duration and a lower PCV were associated with systemic diseases when compared with intranasal diseases. Dogs with neoplastic intranasal disease were significantly older than dogs with nonneoplastic intranasal disease. A majority of dogs with systemic diseases had a history of unilateral epistaxis. Facial or palate deformities were seen only in dogs with neoplasia; obstruction of airflow was seen only in dogs with intranasal disease; and regionally enlarged sub-mandibular lymph nodes occurred in dogs with neoplasia or fungal rhinitis. Additional studies are warranted to examine results in a larger study population and in dogs that are presented to primary care facilities. Screening tests for both E. canis and B. vinsonii (berkhoffii) infections should be considered in dogs with epistaxis until additional data are available.
SigmaStat for Windows, version 2.03, 1992–1997; SPSS, Inc., Chicago, IL 60606
Contributor Notes
