Introduction

Bacterial pneumonia in dogs occurs most commonly due to aspiration or secondary to community-acquired infectious agents. Treatment of bacterial pneumonia consists of antimicrobial therapy that is ideally based on the results of bacterial culture and appropriate antimicrobial susceptibility testing from airway lavage samples. In 2017, the International Society of Companion Animal Infectious Disease published guidelines for the treatment of canine pneumonia based on available evidence or the consensus of the authors.¹ These guidelines included recommendations for empiric antibiotic choice; however, definitive recommendations for the duration of antimicrobial treatment were not made due to lack of evidence. Appropriate use of antimicrobials is of particular interest, as the prolonged or inappropriate use of these drugs can lead to antimicrobial resistance,^2,3 disruption of the gut microbiome,⁴ increased or unnecessary treatment costs, and decreased pet owner compliance with administration.⁵

Canine pneumonia treatment has traditionally included antimicrobial administration for 4 to 6 weeks and/or at least 1 to 2 weeks beyond resolution of clinical and/or radiographic signs of disease.^6–8 This recommendation is based on expert opinion without sound scientific evidence, and many have questioned this practice.^1,8 In a prospective, noncontrolled observational study⁹ of a heterogenous cohort of dogs diagnosed with pneumonia, there was no significant difference in outcome in 24 dogs treated with a shorter course of antimicrobials (median, 11 days) compared to 23 dogs treated with a longer course of antimicrobials (median, 23 days). In a different study¹⁰ investigating C-reactive protein, thoracic radiographs, and lung ultrasound to guide the timing of antibiotic discontinuation in 17 dogs with aspiration pneumonia, 12 dogs were treated with antimicrobials for 1 week, whereas the remaining 5 dogs had antibiotics discontinued after 3 weeks. Although these studies provide some initial evidence that a shorter duration of antimicrobial therapy may be sufficient in some dogs, additional controlled studies are warranted. The aim of this randomized, double-masked, placebo-controlled clinical trial was therefore to describe outcomes in dogs with pneumonia that were treated with a shorter or longer duration of antimicrobials.

Methods

Client-owned dogs with clinical signs and physical examination and radiographic findings consistent with uncomplicated pneumonia at the Matthew J. Ryan Veterinary Hospital at the University of Pennsylvania from July 2020 to May 2023 were eligible for the study. Clinical and physical examination findings associated with pneumonia included but were not limited to fever, cough, tachypnea, and abnormal lung sounds (ie, increased bronchovesicular sounds, crackles) on pulmonary auscultation. Pneumonia was considered uncomplicated if the dog was otherwise previously healthy and had no preexisting lung pathology. Pneumonia was characterized as aspiration pneumonia, community-acquired pneumonia, or undetermined status. Aspiration pneumonia was based on a clinical history of vomiting preceding the onset of respiratory signs. Community-acquired pneumonia was suspected in dogs that had exposure to possible infectious disease, such as a history of being housed at a boarding facility or interactions with dogs at dog parks, and no history of vomiting. Undetermined status was assigned when dogs met neither of the previously described criteria or met criteria for both aspiration and community-acquired pneumonia. Dogs were treated with antimicrobials based on clinician preference guided by the International Society of Companion Animal Infectious Disease guidelines¹ or airway sampling culture and susceptibility testing results, when available. Results from PCR testing for respiratory pathogens were recorded when available. To minimize the potential for confounding by indication, dogs that died or had conditions that could predispose them to recurrent pneumonia, such as brachycephalic breeds, laryngeal paralysis, tracheal stent, megaesophagus, and/or receiving immunosuppressant medications, were excluded. Additionally, dogs that were PCR-positive for carbapenem-resistant Enterobacterales were excluded due to infection prevention protocols as part of a hospital biosecurity program. Finally, dogs that were already receiving antimicrobials for pneumonia prior to being presented to our hospital were excluded to limit the potential added effects of previous antimicrobials. The study protocol was reviewed and approved by the Matthew J. Ryan Veterinary Hospital Privately Owned Animal Protocol and University of Pennsylvania IACUC (No. 585). Informed owner consent for study enrollment was obtained.

After hospital discharge following the initial diagnosis of pneumonia, dogs were scheduled to return to the hospital at 12 ± 2 days. At this first follow-up visit (“first study visit”), an improvement in clinical signs was assessed through an owner interview and a complete physical examination of the dog was performed by one of the study investigators. The following was recorded for each dog: lethargy (present or absent), appetite (normal, no change, or decreased), cough (present or absent at home and present or absent during the examination of the dog), nasal discharge (present or absent), respiratory effort (eupneic, tachypneic, or dyspneic), and lung sounds (normal or abnormal).

Three-view thoracic radiographs were obtained at each study visit. A board-certified radiologist (AS) who was masked to the treatment group evaluated all thoracic radiographs using previously established radiographic severity scoring.¹¹ In short, 7 lung segments were evaluated and a score was assigned for each lung segment: 0, normal; 1, mild (0% to 25% lung lobe affected); 2, moderate (26% to 75% lung lobe affected); and 3, severe (76% to 100% lung lobe affected). A global severity score from 0 to 21 was calculated by adding the severity score for each lung lobe affected. The number of affected lung lobe segments and the global severity score assigned at this initial study visit were compared with the thoracic radiographs from the initial diagnosis of pneumonia, which were also evaluated and scored by the same radiologist (AS).

Dogs with persistent clinical signs associated with pneumonia and progression or no improvement regarding the number of affected lung lobe segments and radiographic global severity score from the initial diagnosis of pneumonia to the first study visit were excluded and offered additional diagnostics and alternate care. Dogs that had an improvement in their clinical signs with normal physical examination findings and either improvement or no change in the radiographic global severity score were randomized to either continued treatment with antimicrobials (4-week group) or placebo medication (2-week group). When prescribing placebo medication, the quantity and frequency matched those of the originally prescribed medication. For example, if the dog was receiving 2 different antimicrobials PO every 12 hours, the dog’s caretaker would administer 2 different placebo medication capsules PO every 12 hours to ensure anonymization. Parallel group randomization was predetermined by use of a random number generator in blocks of 10, with an equal proportion of dogs in each group prior to study commencement. The randomization list was kept in the veterinary clinical investigation center. The group assignment of the dog to either the 2-week group or 4-week group was done at the first screening visit by a veterinary clinical investigation center trial nurse. All study investigators and dog caretakers were masked to the treatment group.

Dogs returned to the hospital for the second follow-up assessment (“second study visit”) at 12 ± 2 days following the first study visit. Dog caretakers were interviewed again on the presence or absence of clinical signs, and a complete physical examination was performed by one of the study investigators. Three-view thoracic radiographs were obtained, and radiographs were evaluated and scored by the same radiologist listed previously. Owner compliance with medication administration was assessed by the trial nurse by counting the remaining doses of antimicrobials or placebo medication at each visit.

For each follow-up visit, the following outcomes were evaluated: (1) the persistence of clinical signs of pneumonia, as defined by the total number of clinical signs (cough, increased respiratory rate, lethargy) present at the visit, and (2) the lack of improvement of radiographic evidence of pneumonia, as assessed by the number of lung lobes affected and the global radiographic severity scoring system.

Results

During the study period from July 2020 to May 2023, 523 dogs presented to the Matthew J. Ryan Veterinary Hospital and were diagnosed with pneumonia. A total of 439 dogs were determined to be ineligible for the study after application of the exclusion criteria, leaving 84 dogs with uncomplicated pneumonia eligible for the study. There were 49 dogs for which the dog’s caretakers declined study participation and 35 dogs for which informed consent was obtained and that subsequently had the first study visit. At the first study visit, 5 dogs were excluded from further study participation due to progression or no improvement in radiographic pneumonia and lack of improvement of clinical signs (3 dogs), persistent megaesophagus (1 dog), or no evidence of pneumonia on initial thoracic radiographs (misdiagnosis; 1 dog). Two of the 3 dogs that were excluded from the study due to lack of improvement in clinical and radiographic signs had their antimicrobial treatment changed (1 dog had an endotracheal wash performed and was positive for Bordetella bronchiseptica), and 1 dog continued the prescribed antimicrobials and did not return to the hospital. Thirty dogs were randomized: 15 dogs to the 4-week group, in which antimicrobials were continued, and 15 dogs to the 2-week group, in which antimicrobials were discontinued and a placebo medication was given for another 2 weeks (Figure 1). Additional enrollment in the study was terminated at this point due to difficulty of enrollment and the long duration of the clinical trial. All enrolled dogs completed the trial, and no dogs were excluded from the analysis.

View Full Size

Flowchart illustrating study dog enrollment. CRE+ = Carbapenem resistant Enterobacterales.

Citation: Journal of the American Veterinary Medical Association 2024; 10.2460/javma.24.04.0259

• Download Figure
• Download figure as PowerPoint slide

Patient information is presented in Table 1. There was no significant difference in the breed, age, sex, or weight of the dogs in the 2-week versus 4-week group (P = .48; P = .64; P = .63; and P = .35, respectively). The median number of lung lobe segments affected on thoracic radiographs at the initial diagnosis of pneumonia was not significantly different in the dogs in the 4-week group (median, 4 segments; IQR, 2 to 4 segments) compared to the 2-week group (median, 4 segments; IQR, 2 to 5 segments) (P = .43). Similarly, the median radiographic global severity score was not significantly different between the dogs in the 4-week group (median score, 5; IQR, 2 to 7) and 2-week group (median score, 5; IQR, 4 to 7) (P = .66).

Only 2 dogs (both in the 4-week group) had an endotracheal wash culture performed and were diagnosed with Escherichia coli. One dog in the PL group had a pyothorax secondary to pneumonia that was managed with thoracostomy tubes and was diagnosed with Pasteurella canis. This dog was noted to be a deviation from study protocol, as a pyothorax is not considered a form of uncomplicated pneumonia. Three dogs (two 4-week dogs and one 2-week dog) had PCR testing performed, of which only 1 dog (2-week group) was positive for Streptococcus zooepidemicus. The most common antimicrobials prescribed included amoxicillin and enrofloxacin (10 dogs), amoxicillin-clavulanate potassium and enrofloxacin (8 dogs), and amoxicillin-clavulanate potassium (6 dogs). The remaining 5 dogs were prescribed doxycycline, azithromycin, azithromycin and enrofloxacin, clindamycin and cefpodoxime, and metronidazole and cefpodoxime. 1 dog each received these antimicrobials. Neither the cause of the pneumonia nor the antimicrobials prescribed were significantly different between the dogs in the 2-week and 4-week groups (P = .13; P = .25). One dog that was prescribed 2 antimicrobials was mistakenly given only 1 of the antimicrobials by the owner. Excluding this outlier, the median number of missed doses was 1 (IQR, 0 to 3 doses) for the first medication and 1 (IQR, 0 to 2 doses) for the second medication. There was no difference in the number of missed doses by treatment group (P = .35 for the first medication and P = .24 for the second medication).

The first study visit occurred within 12 ± 2 days for most dogs (22 of 30 [73.3%]), although 5, 2, and 1 dogs were seen 9, 15, and 16 days after the original visit, respectively. Most dog caretakers noted no change or improved energy levels or appetite (93% [27 of 29 dogs] or 92% [24 of 26 dogs], respectively). No coughing was noted by the caretaker in 83% (24 of 29) of dogs. For the dogs in which coughing at home was reported, there were 3 dogs in the 2-week group (3 of 14 [21%]) and 2 dogs (2 of 15 [13%]) in the 4-week group (Table 2). At the first study visit, only 1 dog in each group was noted to be coughing on physical examination, and no study dogs had nasal discharge, tachypnea or dyspnea, or abnormal lungs sounds noted. 18 dogs (60%) had complete resolution of radiographic lesions. The remaining 12 dogs (40%)all had improvement in their radiographic lesions. The median number of lung lobe segments affected was 0 (IQR, 0 to 1) in both the 4-week and 2-week groups, and the median global severity score was also 0 (IQR, 0 to 1) in both the 4-week and 2-week dogs.

The second study visit occurred within 12 ± 2 days of the first study visit for 26 of 30 dogs (86.7%). Four dogs came 15 days after the first study visit. Most owners noted no change in their dog’s energy levels or appetite (96.4% or 96.6%, respectively). There was only 1 dog (1 of 30 [3%]) in the 4-week group in which the caretaker reported coughing at home (noted also at the first study visit), but the coughing was noted to be decreased (Table 2). At the second study visit, 1 dog in the 2-week group experienced coughing when a sandbag was placed over the dog’s neck to facilitate obtaining the thoracic radiographs, but the dog’s caretaker did not report any coughing at home. No dogs in either the 4-week or 2-week group were noted to have coughing, nasal discharge, dyspnea or tachypnea, or abnormal lung sounds during the physical examination.

At the second study visit, 25 of 30 dogs (83%) had complete resolution of radiographic lesions, including twelve 4-week dogs (12 of 15 [80%]) and thirteen 2-week dogs (13 of 15 [87%]). The remaining 5 of 30 dogs (17%) had either stable (4 dogs) or continued (1 dog) improvement in radiographic lesions. These dogs had severity scores of 1 (mild) in 4 lung lobes (1 dog), 2 lung lobes (2 dogs), and 1 lung lobe (2 dogs). The median number of lung lobes affected was 0 (IQR, 0 to 0) in both groups, and the median global severity score was 0 (IQR, 0 to 0) in both groups (Table 2).

Discussion

Our study investigated clinical and radiographic signs of pneumonia during a 1-month study period in dogs that were randomized to receive antimicrobials for 4 weeks and dogs that received antimicrobials for 2 weeks and a placebo for 2 weeks. None of the 30 dogs in the study had recurrent clinical signs or radiographic worsening of pneumonia during the study period. This finding suggested that dogs receiving a shorter or longer duration of antimicrobials follow a similar clinical and radiographic course of recovery. At this point, it is unknown whether our findings were due to a true equivalence of the 2 treatments or because our study was not powered to detect smaller differences in efficacy. It is notable that some guidelines, although not based on specific studies, suggest that “7 days of antibiotic therapy is likely adequate in most cases [of uncomplicated pneumonia],” even suggesting that the 3 to 5 days of therapy recommended in human guidelines for community-associated pneumonia was “reasonable to consider in dogs.”¹² Taken together, our findings and those of others indicate that 4 weeks of antimicrobial therapy in dogs experiencing clinical improvement is likely unnecessary.

Serial thoracic radiographs are currently recommended in dogs with pneumonia to assess clinical response and to guide the timing of antimicrobial discontinuation. However, in people, this strategy is no longer recommended, as clinical improvement typically precedes the radiographic resolution of lesions.^13,14 In our study, only 60% of the study dogs (53% in the 2-week group) had resolution of radiographic lesions at the first study visit, but all dogs were clinically improved, with only 1 dog in each group having a cough and no dogs in either group experiencing a relapse in pneumonia. At the end of the study, there were 5 dogs (17%) that had persistently mild signs on thoracic radiographs. A similar result was identified in 17 dogs with aspiration pneumonia, in which clinical improvement and normalization of C-reactive protein occurred prior to the resolution of radiographic signs and lung ultrasound abnormalities in most dogs.¹⁰ We did not measure C-reactive protein in our study dogs, and therefore it is unknown whether C-reactive protein levels were normal at the time of antimicrobial discontinuation. However, both of our study results confirm what has already been established in people with pneumonia: clinical improvement occurs prior to radiographic resolution of pneumonia in most patients. Therefore, based on currently available evidence, clinical improvement should be used to guide antimicrobial treatment duration rather than serial thoracic radiographs. The continued reliance on thoracic radiographs to guide antimicrobial treatment duration will lead to longer and unnecessary antimicrobial treatment and increased pet owner healthcare costs.

Due to difficulties recruiting dogs (lack of interest in participating from dog caretakers [49 dogs] and a preponderance of complicated pneumonia patients [439 dogs] at our referral facility, which led to a large number of exclusions), enrollment over 3 years did not achieve our original target study population of 70 patients. This low patient accrual is not dissimilar to what has been reported in human clinical trials, with 55% of clinical trials being terminated due to low recruitment globally and over 80% of trials in the US failing to achieve targeted patient enrollment.¹⁵ As a result of not achieving our target enrollment, it is likely that the current study is underpowered to detect modest differences in relapse rates between dogs treated with different durations of antimicrobial therapy; therefore, comparison on outcome measures (clinical and radiographic relapse of pneumonia) was not performed. A post hoc sample size calculation indicates that our sample size of 30 patients would have been able to detect a 1-sided difference in relapse rates of 50% or more (power, 80%; α = .05). True pneumonia relapse rates in dogs are unknown, but another small study⁹ identified a 17% long-term relapse rate in dogs with uncomplicated pneumonia and a higher relapse rate (38%) in dogs with conditions that predispose pneumonia, such as megaesophagus. While larger controlled prospective studies are needed to identify whether antimicrobial duration affects relapse rates of pneumonia, our experience highlights the difficulty of performing timely controlled prospective studies in veterinary medicine, which is likely one of the factors that has limited the development of evidence-based guidelines for durations of antimicrobial use in companion animals.

Another potential limitation of our study was that we did not require airway sampling to confirm the presence of bacterial pneumonia and assess for antimicrobial susceptibility; only 2 study dogs had airway sampling performed, and 2 additional dogs had positive bacterial culture of pleural effusion or positive PCR testing. It is possible that dogs enrolled in our study may not have required antimicrobials (ie, viral pneumonia or pneumonitis) or that they may have been infected with antimicrobial-resistant bacteria. Airway fluid sampling for bacterial culture and susceptibility is invasive, may not always be recommended for dogs with uncomplicated pneumonia, and may be declined by the dog’s caregivers. However, our study did reflect clinical practice, where most dogs with radiographic signs and a clinical history consistent with pneumonia are treated empirically with antimicrobials. Interestingly, we did exclude 3 dogs from the study at the first study visit due to lack of improvement of radiographic and clinical signs. Two of these dogs had an improved clinical response following a change in antimicrobial therapy, and these dogs potentially represent initial empiric antibiotic treatment failure. These dogs were excluded from our study, as our goal was to investigate the clinical course of dogs with pneumonia that were treated with a shorter course of antibiotics versus a longer course of antibiotics, not to investigate the incidence of empiric antimicrobial treatment failure that can occur. Finally, as it is not always easy to differentiate between different causes of pneumonia and there are no current recommendations for a difference in duration of antimicrobial treatment based on cause, we elected to include all uncomplicated pneumonia dogs. There is a possibility that different causes of pneumonia may require different durations of antimicrobial treatment. We did have 1 study protocol deviation with the inclusion of the dog in the study with a pyothorax secondary to pneumonia. This dog was randomized to the shorter course antimicrobial treatment group and ultimately did not experience relapse of clinical or radiographic signs. The duration of antimicrobial treatment in dogs with pyothorax and other forms of complicated pneumonia requires further investigation.

The similarity in outcomes observed in this study and in other observational studies suggests that shorter durations of antimicrobial therapy may be sufficient in dogs with uncomplicated pneumonia. Our results also suggest that clinical signs are more useful for guiding discontinuation of antimicrobial therapy for pneumonia than radiographic signs. To demonstrate the true noninferiority of different durations of antimicrobial therapy, very large sample sizes would be necessary, which appears to be difficult to achieve in veterinary settings. However, multiple layers of evidence appear to support the idea that shorter durations of antimicrobials can be used in this patient population.

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.

Funding

Grant funding was provided by the Companion Animal Research Fund of the University of Pennsylvania School of Veterinary Medicine.

ORCID

E. L. Reineke https://orcid.org/0000-0001-9911-8698