Pneumonic Plague in a Dog and Widespread Potential Human Exposure in a Veterinary Hospital, United States

Paula A. Schaffer; Stephanie A. Brault; Connor Hershkowitz; Lauren Harris; Kristy Dowers; Jennifer House; Tawfik A. Aboellail; Paul S. Morley; Joshua B. Daniels


Emerging Infectious Diseases. 2019;25(4):800-803. 

In This Article

The Study

A 3-year-old mixed-breed dog was brought to a veterinarian in Colorado, USA, during December 2017 for evaluation of lethargy and fever 4 days after the dog was observed sniffing a dead prairie dog. Treatment with amoxicillin/clavulanic acid was initiated before referral the next day to the Colorado State University Veterinary Teaching Hospital (CSU-VTH; Fort Collins, CO, USA) because of progressive illness and development of hemoptysis. Imaging demonstrated unilateral lung lobar consolidation and small foci of parenchymal density in other lobes and intrathoracic lymphadenopathy (Figure 1).

Figure 1.

Transverse computed tomography of dog with pneumonic plague on day 2 of hospitalization, Colorado, USA. Image shows accessory lung lobar consolidation.

Plague was considered unlikely because of the animal species, season, lack of peripheral lymphadenopathy, and unilateral lobar imaging pattern consistent with an aspirated foreign body, which is common in dogs.[3] Treatment with ampicillin/sulbactam and enrofloxacin was initiated, and accessory lung lobectomy was performed to remove the presumed source of sepsis. Consolidation of the accessory lobe and scattered dark red foci in other lung lobes were noted intraoperatively. Histologically, the excised lobe was effaced by severe necrosuppurative pneumonia with hemorrhage and fibrinous pleuritis but no intralesional bacteria (Figure 2).

Figure 2.

Histopathologic analysis of accessory lung lobe of dog with pneumonic plague (hematoxylin and eosin stain), Colorado, USA. A) Parenchyma, which is diffusely effaced by necrohemorrhagic pneumonia. Scale bar indicates 500 μm. B) Alveolar detail, which is obscured by necrosis, hemorrhage, and suppurative inflammation without intralesional bacteria. Scale bar indicates 20 μm.

After 48 hours of aerobic incubation, a swab specimen of lung parenchyma yielded light and pure growth of bacteria that we identified by using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Vitek-MS, with 91.1% confidence as Yersinia pseudotuberculosis, although the database of the instrument contained mass spectra for 5 Y. pestis strains. Because signs were not consistent with Y. pseudotuberculosis, there was concern about misidentification. We performed PCR of the isolate the next day (5 days after admission) by using a Centers for Disease Control and Prevention (Atlanta, GA, USA) Laboratory Response Network protocol for Y. pestis (

The dog was humanely killed the same day because of progression of pneumonia and poor prognosis. A limited necropsy was performed by informed personnel and found diffuse necrosuppurative and hemorrhagic pneumonia and severe necrotizing tonsillitis. Only liver tissue was positive for Y. pestis by PCR.

The dog had been transported throughout the hospital and housed in an oxygen cage vented to the room, potentially exposing personnel from multiple clinical services. Those handling specimens in the diagnostic laboratory were also considered exposed to Y. pestis. Exposures during the first 2 days of hospitalization were considered most critical because fluoroquinolone treatment was initiated after admission, and guidelines from the Colorado Department of Public Health and Environment (CDPHE) state that veterinary patients with Y. pestis have limited contagious risk after 48 hours of appropriate therapy.[4]

While PCR results were pending, paper sheets were circulated to personnel to record contact with the dog. After the positive PCR result, emails were sent to these persons, followed by emails to all personnel. The delay between suspicion and diagnosis of Y. pestis resulted in word of mouth traveling faster than official communication, which caused anxiety among personnel. Many expressed frustration that suspicion and diagnosis of plague did not occur earlier. Two hospital-wide meetings were held for questions, discussion, and feedback. An online postincident survey was conducted, and 52 respondents indicated that they were aware of their potential exposure within 48 hours of the diagnosis.

We found 116 documented potential human exposures (Table 1). CDPHE recommendations were based on risk assessments, and interventions were decided by potentially exposed persons in consultation with healthcare providers (Table 2). In addition, 46 hospitalized animals co-housed in the same room were classified as potentially exposed. Prophylactic antimicrobial drugs were recommended because most of these animals were critically ill and had decreased immune status. To our knowledge, there were no cases of Y. pestis infection in potentially exposed humans or animals. A fever developed in 1 person, but this fever was not determined to be caused by Y. pestis infection. One survey respondent reported adverse effects from antimicrobial drugs.