Clinical features and radiographic findings in cats with eosinophilic, neutrophilic, and mixed airway inflammation (2011-2018).

Abstract Background Idiopathic inflammatory airway disease (IAD) in cats often is described as asthmatic (eosinophilic) or bronchitic (neutrophilic), but this designation requires collection of airway fluid and it fails to consider cats with mixed airway inflammation. Objective To identify clinical features that would differentiate inflammatory disease types. Animals Forty‐nine cats with nonspecific airway inflammation identified by bronchoscopic bronchoalveolar lavage (BAL) between 2011 and 2018 were evaluated. Methods This is a retrospective study. Cats were categorized by BAL differential cytology as having eosinophilic (eosinophils >20% with neutrophils <14%, or eosinophils >50%), mixed (eosinophils 20%‐50% and neutrophils >14% or discordant inflammation from 2 BAL sites), or neutrophilic (neutrophils >14% and eosinophils <20%) inflammation. Type and duration of presenting complaints, signalment, body condition score, respiratory rate, CBC results, bronchoscopy, BAL results (% recovery, total nucleated cell count, differential cell count), and radiographic findings were compared among groups. Results Idiopathic IAD was diagnosed in 49 cats, with BAL eosinophilic inflammation in 23, mixed inflammation in 14, and neutrophilic inflammation in 12. Cough was the predominant presenting complaint with no difference in duration of signs among groups (median, 5.5 months). Respiratory rate and effort also did not differ. Cats with eosinophilic inflammation were significantly younger (4.4 ± 3.3 years) than those with neutrophilic (8.0 ±5.6 years) or mixed inflammation (7.5 ± 4.0 years; P = .03). Results of CBC and interpretation of radiographic findings did not differ among groups. Conclusions and Clinical Importance Substantial overlap exists in clinical and radiographic findings in cats with various forms of idiopathic airway inflammation.

K E Y W O R D S asthma, endoscopy, eosinophilic inflammation, neutrophilic inflammation, parenchymal disease, respiratory tract

| INTRODUCTION
Inflammatory airway disease (IAD) is common in cats and is an important differential diagnosis for cats presenting with a history of cough, wheezing, or increased expiratory effort. [1][2][3][4][5] Inflammatory airway disease has been proposed to encompass 2 distinct syndromes: asthma and chronic bronchitis. 4,6,7 Asthma in cats is considered to be similar to asthma in humans and is characterized by increased airway responsiveness and bronchoconstriction, manifested by acute wheezing or respiratory difficulty, which can be reversed by treatment with a bronchodilator. Similar to some forms of the syndrome in humans, asthma in cats has been suggested to result from exposure to allergens, and a disease syndrome typified by eosinophilic airway inflammation can be created experimentally by allergen or antigen challenge. [2][3][4]7 In contrast, chronic bronchitis in cats is defined by the presence of cough, and bronchoconstriction is not part of the clinical picture. 2,8 Chronic bronchitis in cats has been defined similarly to the disease in dogs 9 with development of nonseptic suppurative airway inflammation in response to an unknown insult. Currently, the etiopathogenesis of the naturally occurring diseases remains poorly understood, and it is unclear whether asthma and chronic bronchitis in cats are separate diseases or part of a spectrum of diseases resulting from inflammation.
In people, differentiation of asthma from bronchitis relies on pulmonary function testing, 10 which is not readily applicable to cats.
Diagnosis of IAD is based on exclusion of other etiologies such as parasitic, infectious, or neoplastic disease in conjunction with a nonseptic inflammatory infiltrate on airway lavage sampling. Asthma in cats is considered to result in predominantly eosinophilic inflammation and chronic bronchitis in predominantly nonseptic neutrophilic inflammation. 4,6,7 However, marked airway eosinophilia has been reported in clinically normal cats, 11,12 and techniques used to collect and evaluate airway samples have been inconsistent across studies. [1][2][3][4][5]7,9,[12][13][14] Previous studies evaluating clinical findings in cats with naturally occurring disease have failed to identify clinically relevant differences between cats with asthma and those with chronic bronchitis, with the exception of some investigators noting younger age in cats with airway eosinophilia. 1,5,8,13,15 The purpose of our retrospective study was to investigate clinical variables that could aid in differentiation between inflammatory airway syndromes in a large group of cats with naturally occurring disease in which standardized diagnostic testing, including bronchoscopic bronchoalveolar lavage (BAL), had been performed. We hypothesized that cats with eosinophilic inflammation would be younger, have a shorter duration of clinical signs, display less severe radiographic changes, and have higher circulating eosinophil counts compared to those with neutrophilic or mixed airway inflammation. Cats were premedicated using a balanced anesthetic plan determined by board-certified anesthesiologists, and bronchoscopy was performed under IV propofol anesthesia by using an induction dosage of 4 mg/kg followed by 0.1 to 0.4 mg/kg/min as a continuous rate infusion. All cats were preoxygenated and treated with SC terbutaline (0.01 mg/kg) before the procedure, and oxygenation was maintained by jet ventilation at 180 breaths/min. Pulse oximetry, ECG, and blood pressure were monitored throughout the procedure in all cats. Bronchoscopy was performed using either a 2. 8  Bronchoscopy was performed in standard fashion by faculty and house officers trained by 1 of the authors (L.R.J.). All bronchi were evaluated sequentially, and then the bronchoscope was removed from the airways to irrigate the channel with sterile saline and wipe debris from the outer surface. The bronchoscope was reinserted to complete BAL at specific site(s) chosen by the endoscopist. Bronchoalveolar lavage was performed by instilling 3 to 5 mL of warmed, sterile saline through the biopsy channel of the endoscope, flushing the channel with 2 to 3 mL of air, and immediately applying hand suction to recover fluid that had been in contact with the bronchoalveolar space.
Single-aliquot lavage was performed at 1 to 3 sites and fluid was sub- Reference intervals used by our laboratory for BAL total and differential cell counts are 300 to 400 cells/μL comprised of up to 7% neutrophils or lymphocytes, up to 18% eosinophils, and 65% to 85% macrophages. 11,12 Cats were categorized as having eosinophilic, mixed, or neutrophilic inflammation based on BAL fluid differential cytology.
Previous reports have used eosinophil percentages of >17 to 20 8,[15][16][17][18] to define eosinophilic inflammation, and neutrophil percentage exceeding the reference range to describe noneosinophilic or neutrophilic inflammation. Other studies define inflammation based on the prominent cell type. 1,3,19,20 Rigorous categorizations were used here, and eosinophilic inflammation was defined as BAL fluid cytology containing eosinophils >20% with neutrophils <14% (mean + twice the SD to ensure that cytological inflammation was consistent with clinical disease 11 ), or by eosinophils >50% in 1 to 2 BAL sites. Neutrophilic inflammation was defined by neutrophils >14% and eosinophils <20%. Mixed inflammation was defined by eosinophils 20% to 50% and neutrophils >14% on cytology or as discordant inflammation from 2 BAL sites. 16 When only a single site was lavaged, cases were categorized based on interpretation of a single lavage fluid cytology.
Data collected from the medical record included type and duration of presenting clinical signs (e.g., cough, sneeze, nasal discharge, increased respiratory effort) as reported by the owner, age, sex, indoor or outdoor status, medications used within the last 7 days, weight, body condition score (BCS), physical examination findings (temperature, heart rate, respiratory rate, and effort), bronchoscopy findings, BAL fluid recovery, BAL fluid cell count and cytology, and BAL culture. Fecal and heartworm test results were recorded where available and contrasted with results from the hospital population. Radiographic features and CBC data were included only when performed at the University of California Davis School of Veterinary Medicine for consistency of technique, interpretation, and reference ranges. The CBC data were evaluated for total leukocyte, neutrophil, and eosinophil counts. Thoracic radiographs were reviewed in masked fashion by a board-certified radiologist (E.G.J.) using a 0 to 9 scoring system adapted from previously described studies. 1,4,15 Bronchial patterns were scored as follows: 0, absence of a bronchial pattern; 1, mild, first-generation bronchi visible; 2, moderate, second-generation bronchi visible; and 3, severe, third-generation bronchi visible. Interstitial patterns were scored as follows: 0, absence of interstitial markings; 1, mild interstitial framework visible; 2, moderate, interstitial framework distinguishable from bronchiolar pattern; and 3, severe, overt interstitial framework. Alveolar patters were scored as follows: 0, absence of alveolar infiltrates; 1, mild, focal unilateral alveolar pattern; 2, moderate, focal bilateral alveolar pattern; and 3, severe, multifocal bilateral alveolar pattern or lobar collapse. Total possible severity score was 9. Additional features recorded included topographic distribution of disease, presence of bronchiectasis or broncholithiasis, pulmonary artery enlargement, and hyperinflation or hypoinflation.
Data were assessed for normality using the D'Agostino and Pearson Omnibus test (GraphPad Prism version 5.0f, San Diego, California).
Parametric data are reported as mean with SD and nonparametric data as median with range. Normally distributed data (age, heart rate, radiographic score, neutrophil count, BAL fluid recovery) were compared among groups using a 1-way analysis of variance with Tukey's multiple comparison test for post hoc analysis. Nonparametric data (weight, duration of signs, temperature, respiratory rate, white blood cell count, eosinophil count, BAL fluid TNCC/μL) were assessed using a Kruskal-Wallis test and post hoc Dunns multiple comparison test. Presence of cough, sneeze, or nasal discharge, increased respiratory effort, and radiographic findings were assessed by using chi-square analysis. The correlation between duration of cough and total radiographic score was assessed using linear regression. Significance was defined as P < .05.

| RESULTS
Clinical information of cats with inflammatory airway disease evaluated in this study    or bronchitis in 24 cats. 13 No differences in history, signalment, or radiographic findings were reported in that study, other than the presence of nasal discharge in cats with chronic bronchitis. In contrast, our study found that cats with BAL fluid eosinophilia were younger than cats with other types of airway inflammation, which is consistent with some 8,13,15,20 but not all 1,6 investigations of lower airway disease in cats. Interestingly, despite younger age, no differences in duration of signs were found among cats with different types of airway inflammation, which was contrary to our hypothesis and not specifically reported in a previous study, 13 which required signs >3 months in duration for a diagnosis of bronchitis. Peripheral eosinophilia was relatively uncommon in our group of cats (14%) and although the highest counts tended to be in cats with respiratory eosinophilia, this result was not statistically significant. In comparison, the previous study reported peripheral eosinophilia in 40% of cats with asthma and in 27% of those with chronic bronchitis. 13 Finally, contrary to our hypothesis, radiographic changes were of equal severity in all groups of cats and, overall, no clinical variables were found that could differentiate among cats with different types of lower airway inflammation.
Cats with eosinophilic airway disease were younger (mean age, 4.4 years) than cats with mixed or neutrophilic inflammation, which is consistent with previous reports, 8,13,15,20 although only 1 other study found this finding to be statistically significant. 15 This tendency toward a younger age of onset parallels the development of asthma in humans, most commonly a childhood disease where the allergic or hypersensitivity response is related to specific development of neonatal and juvenile immunity. [21][22][23] In neonatal murine models of asthma, increased diversity of lung or gut microbiota promotes a more favorable environment for tolerance against allergic disease by inducing T-regulatory cells, lower concentrations of IgE and altered balance of Th1/Th2 immune responses. [24][25][26] In studies of cats, a similar relationship has been suggested between gut and lung mucosal microbiomes. 27,28 If mucosal immunity development is the same in the cat as in the mouse, this observation could have implications for specific treatment or prevention of eosinophilic lung disease.
Airway disease in the cat also shares several corollaries with asthma syndrome in horses, which encompasses IAD, a relatively mild condition of younger horses, and recurrent airway obstruction. Both of these disorders in horses are associated with increased airway mucus and airway hyperresponsiveness, and they respond to strict control of exposure to dust and allergens. Recurrent airway obstruction is characterized by more severe airway obstruction and marked airway neutrophilia, in comparison with the mild mixed airway inflammation usually present in IAD. 29,30 Neutrophilia predominates in the allergic disease process in horses partially because of IL-4-mediated release of neutrophilic chemotactic factors, 31 and the severity of disease correlates to some extent with the severity of airway neutrophilia. 32 The cytokine milieu in the respiratory tract of cats with naturally occurring airway disease has not been evaluated, but 1 study did associate the magnitude of airway inflammation with the clinical and radiographic severity of disease, 3 and airway responsiveness has been correlated with the severity of airway inflammation. 8,15 Airway eosinophilia was the most common IAD (47%) in our group of cats, which is similar to recent studies, 8,11,13,15,20 but other, primarily older, publications 1-3 have reported eosinophilic inflammation in a smaller proportion (24%-32%) of cats evaluated. This difference might suggest an increasing trend toward eosinophilic disease in cats, which is interesting considering that allergic and asthmatic diseases have been increasing in people. 21 This phenomenon has been attributed to multiple environmental and cultural shifts sometimes referred to as the hygiene hypothesis. It has been reported previously that exposure to rural environments has a protective effect on development of allergy and atopy in people. 33 reported nasal discharge more often in cats with chronic bronchitis than in those with asthma. An association of asthma with upper respiratory viruses has been noted in children, 37,38 and nasal airways in cats with experimentally induced asthma were noted to have eosinophilic infiltration in the absence of clinical signs. 39 Whether this finding was related to the method of hypersensitization via aerosolized allergen in the experimental study or represented a distinct response of the upper respiratory epithelium in those cats is unclear. It seems likely that sneezing or nasal discharge noted in cats in our study was not related to the same stimulus causing lower airway inflammation.
Diffuse bronchial and interstitial patterns were the most common radiographic findings in our study, with 9% of cats having normal thoracic radiographic findings. Similar findings have been reported previously, along with the fact that radiographic findings did not differ among inflammatory categories or even between infectious and noninfectious IADs. 1,2,5,11,13,15,40,41 In our study, the severity of radiographic pulmonary infiltration was not associated with duration of clinical signs. Although differentiation of lower airway disease in cats into asthma or chronic bronchitis often is based on detection of eosinophilic versus neutrophilic airway lavage cytology, there is no standardized definition.
In addition, multisegment BAL cytology can detect discordant inflammation in separate lung lobes of a single cat in up to 48% of cases. 16 Given the similarities in clinical complaints, physical examination findings, clinicopathologic variables, and radiographic results among groups identified here, the previous finding of discordant BAL cytology among lobar lavage sites, and the occurrence of mixed inflammation in many cases, it is possible that asthma and chronic bronchitis in cats are different manifestations of a disease process based on inflammation.
Whether this disease process has the same etiology or is multifactorial remains to be determined.
Limitations of our study are those typically encountered in a retrospective study, including lack of standardization in history collection, failure to exclude animals based on prior administration of medications, and lack of fecal or heartworm testing in all patients. Despite being in a nonendemic region for parasitic airway diseases, 46