Clinical response to 2 protocols of aerosolized gentamicin in 46 dogs with Bordetella bronchiseptica infection (2012‐2018)

Abstract Background Bordetella bronchiseptica (Bb) infection commonly causes respiratory disease in dogs. Gentamicin delivered by aerosol maximizes local drug delivery without systemic absorption but clinical response to protocols remains undetermined. Objectives To compare the clinical response to 2 protocols of aerosolized delivery of gentamicin in bordetellosis. Animals Forty‐six dogs with Bb infection confirmed by culture or quantitative polymerase chain reaction on bronchoalveolar lavage. Methods Retrospective study. Administration of aerosolized gentamicin for ≥10 minutes q12h for ≥3 weeks using 4 mg/kg diluted with saline (group 1) or undiluted 5% solution (group 2). Clinical response firstly assessed after 3‐4 weeks and treatment pursued by 3‐weeks increments if cure not reached. Cure defined as absence of cough persisting at least a week after treatment interruption. Results Demographic data were similar between both groups. Clinical cure at 3‐4 weeks was more frequently observed with the use of undiluted solution (19/33 vs 3/13 dogs, P = .03) in association with a shorter median duration of treatment (4 vs 6 weeks, P = .01). Dogs from group 2 having less than 1000 cells/μL in lavage were also more likely to be cured at 3‐4 weeks than dogs with more than 1000 cells/μL [9/9 vs 10/19, P = .006] and median duration of treatment in that subgroup of animals was reduced (3 vs 5 weeks, P = .02). Conclusion and Clinical Importance Aerosolized delivery of gentamicin seems effective for inducing clinical cure in Bb infection. Clinical response appears better using undiluted 5% solution, particularly in the subgroup of dogs having less than 1000 cells/μL in lavage.


| INTRODUCTION
Despite the widespread availability of vaccines, Bordetella bronchiseptica (Bb) is still commonly documented as one of the primary causative agents of canine infectious respiratory disease complex (CIRD-C) affecting mainly young dogs, especially in boarding kennels. [1][2][3][4][5][6][7][8] This contagious disease can be often self-limiting, but a wide range of respiratory signs have been described, from mild illness to severe pneumonia and potential death. [8][9][10][11] As both symptomatic and carrier dogs can be a source of infection for other dogs and immune-compromised owners, [12][13][14]  While acute illness is the most common presentation, chronic and refractory Bb infection can also develop in dogs as a result of different mechanisms including adherence of Bb to cilia, induction of ciliostasis, biofilms and local immunosuppression. [19][20][21][22][23][24] Therapeutic concentrations of antimicrobial drugs could not be reached on the apical surface of bronchial epithelium, even if isolates are known to be in vitro susceptible to systemic drugs such as doxycycline or enrofloxacin. 25 In humans, aerosolized delivery of aminoglycosides is now considered as one of cornerstones for treatment of bacterial infections in patients with various bronchopneumopathies [26][27][28] with nebulizers producing adequate size of droplets able to reach lower airways. 29,30 Although nebulization of gentamicin is largely used in practice for respiratory infections in dogs, a single study about aerosolized delivery of gentamicin for Bb infection in dogs has been published 3 decades ago. 31 Bacterial load decreased immediately after first administrations but increased within days of stopping treatment; protocols also varied between dogs with treatment and follow-up periods being unfortunately short. 31 The first objective of our study was to describe the clinical response to 2 protocols of gentamicin nebulizations in dogs with Bb infection that showed only poor improvement with previous oral antimicrobial treatment. Secondly, possible factors that could be associated with clinical response to aerosolized delivery of gentamicin were evaluated including age, breed, acute presentation (cough for less than 2 weeks), presence of alveolar radiographic lesions, previous steroid treatment, bacterial co-infections, total cellularity and severity of neutrophilic inflammation of the lavage, concomitant oral antimicrobial treatment, use of cage vs facial mask, concomitant abnormal airways conformation and type of protocol.

| Aerosolized delivery of gentamicin treatment
Immediately after Bb infection confirmation, aerosolized delivery of gentamicin was prescribed twice daily. Two protocols were successively used over the study period. Initially (from 2012 to the beginning of 2014), a dose 4 mg/kg was used twice daily with systematic dilution in saline to obtain in toto 5 mL of reconditioned solution nebulized until the cupule was empty (=protocol 1). The protocol was then modified since 2015 to a fixed concentration of gentamicin solution (gentamicin 5%) (2-4 mL volume of undiluted gentamicin) nebulized over a minimum of 10 minutes independently of the dog size (=protocol 2). The time needed before the cupule was empty varied according to the type of nebulizer; if the cupule was empty before 10 minutes of nebulization, additional volume of undiluted gentamicin was added.
A pediatric nebulizer was used. Owners of dogs of less than 10 kg of body weight were instructed to place their dog in a feline transport cage adapted to the dog's size covered by a thick towel or blanket.
Owners were also advised to wait a few minutes after the end of the nebulization before taking the dog out of the cage. For dogs of higher weight (>10 kg), the tube of the nebulizer was connected to a facial mask adapted to the size of the nose. Care was also taken to minimize owner's direct exposition to nebulized drug using distance from the dog during nebulization, limited space between mask and muzzle in >10 kg dogs or blanket over the transport cage in <10 kg dogs. For both protocols, nebulizations were initially prescribed twice daily for at least 3-4 weeks and then pursued by additional 3 weeks-increments until complete resolution of cough.
No cough suppressant was prescribed during the period study at both centers. Concomitant oral antibiotic treatment was initiated or pursued at the clinicians' discretion. At time of follow-up, all owners were asked for clinical tolerance of nebulization at home.
Clinical cure was defined as the resolution of clinical signs (absence of cough, resolution of dyspnea, pyrexia, inappetence, and lethargy) without recurrence within 1 month after the interruption of nebulization. Follow-up was carried out by both control visits and phone calls. After the last control visit, owners were instructed to inform the teaching hospital if cough re-occurred thereafter, specifically within the first months after discontinuation of aerosols.

| Statistical analysis
Statistical analyses were performed with a commercially available software (XLstat software). Data were expressed as median and range for continuous variables and as proportion for categorical variables.
Clinical response to aerosolized delivery of gentamicin was studied on 2 criteria: the proportion of dogs cured after 4 weeks and the median duration of treatment until clinical cure. Both criteria were studied depending on several factors including: age less than 6 months, breed (brachycephalic vs nonbrachycephalic dogs), acute presentation (cough for less than 2 weeks), presence or absence of co-infection by M. cynos or other bacterial species, presence or absence of alveolar pattern on thoracic radiographs, history of previous steroid treatment, concomitant oral antimicrobial treatment during gentamicin protocol, use of cage vs facial mask, concomitant abnormal airways conformation at diagnosis, total cellularity of BALF (TCC less or more than 1000 cells/μL), severity of neutrophilic inflammation (less or more than 50% neutrophils in BALF) and finally, type of protocol (use of diluted or undiluted gentamicin).
Associations between proportion of dogs cured after 4 weeks and studied criteria were tested by Chi-square test (with n > 5) and associations between median duration of treatment and studied criteria were tested by Mann-Whitney test. Values of P ≤ .05 were considered significant. Eight dogs were previously vaccinated; only parenteral vaccine was used for all these 8 dogs and was administered more than 2 months before diagnosis.

Forty
All dogs had cough of 4 days to 2 years duration (median 2 months). Acute presentation (cough less than 2 weeks) was observed for 6 dogs, whereas for 36 dogs, cough was reported since adoption from kennels, pet shops or shelters. Twenty cases experienced moderate or severe dyspnea requiring hospitalization for oxygen treatment in the half of them. In 19 cases, mucous or purulent nasal discharge was observed associated with sneezing or reverse sneezing in 2 dogs. Ten dogs presented recurrent pyrexia of moderate severity (from 39.2 to 39.9 C) before diagnosis. All included dogs had been unsuccessfully treated with antimicrobial drugs including amoxicillin or amoxicillin/clavulanic acid (n = 37), doxycycline (n = 20), marbofloxacin or enrofloxacin (n = 13), cephalexin (n = 1), cefovecin (n = 1), spiramycin (n = 1), metronidazole (n = 1), and azithromycin When dogs from both groups were compared (Tables 1 and 2), cases treated with protocol 2 were more frequently pretreated with steroids before inclusion (12/33, 36% vs 0/13, 0%, P = .01). When only dogs treated with protocol 2 were considered (data not shown), similar results were obtained: association between proportion of dogs cured at 3-4 weeks and TCC from BALF remained significative (9/9 vs 9/19, P = .006) and median duration of treatment was also shorter for dogs having TCC less than 1000 cells/μL in BALF (3 weeks vs 5 weeks, P = .02).

T A B L E 3
Number and proportion of dogs with clinical cure at 3-4 weeks and median duration of treatment depending on studied criteria  We also observed that dogs having less than 1000 cells/μL in BALF were more frequently cured at 3-4 weeks than dogs with more than 1000 cells/μL in lavage. Moreover, for dogs receiving undiluted solution, median duration of treatment was also significantly shorter with cellularity of lavage less than 1000 cells/μL. This observation seems relevant as a standardized BAL procedure was used. We can hypothesize that cellularity of BALF could be associated with disease severity, although validated clinical scoring was not available in our study population to confirm this hypothesis. High cellularity (more than 1000 cells/μL) could also suggest significant accumulation of suppurative secretions, which might limit optimal diffusion of gentamicin. Such finding poses TCC as an interesting clue of severity carrying prognostic value.
Interestingly, we failed to show a significant association of other  were used over the study period, leading to potential variations in droplets size, power of the aerosolization and the deepness of the drug deposition. In addition, for ethical reason, our study did not include any group treated with saline only. Moreover, definitive cure was only defined based on clinical response, that is, absence of cough persisting at least 1 month after the interruption of aerosols. Even if clinical follow-up was long, a second analysis of BALF collected at the end of treatment would have been more helpful to most adequately qualify definitive cure. However, such procedure requires anesthesia and most owners were reluctant to allow it while their pet appeared cured; thus, it was not considered in the study design for the inclusion. Additional studies including sequential BALF analysis after treatment is warranted to ascertain the efficacity of aerosolized administration of gentamicin. Lastly, although presence of brachycephalic obstructive airway syndrome would be expected to compromise optimal aerosolized drug delivery to the lower airways (particles colliding into upper airway mucosa due narrowed airways), we failed to observe a statistical difference in clinical response between brachycephalic and nonbrachycephalic dogs. We cannot rule out that this negative finding was related to the small size of the population in the present study (underpowered study).
In conclusion, while being more time-consuming than oral antimi-