Bordetella spp. are aerobic coccobacilli known to be present in the upper respiratory tract of many animals . B. bronchiseptica infections are uncommon in humans . The literature on this subject is subsequently poor. Two comprehensive reviews have been published in the last two decades. The first included 25 cases from 1911 to 1990 . However, the presence of B. bronchiseptica was microbiologically confirmed in only ten patients. The other study, published in 1995, included 52 patients but no details were provided about the microbiological identification of B. bronchiseptica . Another study published in 2005 focused on the pathogenesis but gave little information on human infections . Recently, several case reports addressed the problem in cystic fibrosis patients , HIV patients [2,6] or children with lung transplants . Facing this lack of recent information, we decided to review all cases detected at our institution from 1993 to 2008.
The present study was undertaken at the Geneva University Hospitals, a 2220-bed tertiary care centre. We performed a retrospective case review approved by the institutional ethics review board and based on computerized laboratory log files, examining the pattern of disease caused by B. bronchiseptica, addressing underlying conditions and exposures, and examining antimicrobial treatment in relation to patient outcome. To avoid misclassification bias, we included only patients with verified B. bronchiseptica isolates because automated identification systems can lead to misclassification as Acinetobacter spp. or other nonfermentative Gram-negative rods. Most cases were first identified by API 20NE gallery or VITEK2 (bioMerieux SA, Marcy-l’Etoile, France) phenotypic identification systems, and then supported by a positive oxidase reaction. All available stored isolates (4/8) were verified as B. bronchiseptica by 16s rDNA sequencing and a positive oxidase reaction, confirming that our phenotypic identification was reliable. Because the method of previous antimicrobial susceptibility testing (AST) could not be verified in this retrospective study, we performed new AST [MIC using the E-test (AB Biodisk, Solna, Sweden)] on all available isolates, despite the absence of CSLI guidelines for the E-test AST of B. bronchiseptica . Therefore, E-tests were performed as previously recommended . Considering the fastidious character of the microorganism, a second MIC determination was performed at 48 h, but this did not reveal any relevant change in AST.
During a 15-year period, eight patients had confirmed B. bronchiseptica infection (n = 3), co-infection (n = 2) or colonization (n = 3). Table 1 summarizes the key features of these cases. All specimens except one were retrieved from the airways. Cases were equally distributed among gender and age groups. Amongst the eight patients, seven had significant underlying disease, including four patients with severe lung disease, two patients with AIDS and one patient with autoimmune neutropenia. Three patients had documented contact with cats prior to infection. In one case, the cat itself had respiratory symptoms.
Table 1. Characteristics of patients
| ||Age, sex||Underlying disease||Diagnosis||Specimen||Other microorganisms||Year||Method of identification||Pathogenic role of Bordetella bronchiseptica||Animal exposure||Antimicrobial treatment||Short-term outcome|
|1||65, F||Severe COPD||Colitis with septic shock||Tracheal aspirate||Enterobacter aerogenes||2009||Vitek2 (99%)a and MALDI-TOF MS||Colonizer||Unknown||Imipenem||Death, unrelated cause|
|2||34, F||None||Asymptomatic||Cervical smear||None||2008||Vitek2 (99%) and oxidase||Colonizer||Cat||Amoxicilin-clavulanate||Cured|
|3||65, M||Peripheral neutropenia||Pneumonia with abscess||Sputum||None||2008||Vitek2 (99%) and oxidase||Infectious agent||Cat||Imipenem||Cured|
|4||21, M||Cystic fibrosis||Asymptomatic||Sputum||Pseudomonas aeruginosa||2008||Vitek2 (99%) and oxidase||Colonizer||Unknown||Trimethoprim-sulfamethoxazole and azithromycin||Cured|
|5||17, M||Cystic fibrosis||Bronchitis||Throat swab||β-hemolytic streptococcus group G||2006||16s rDNA sequencing and oxidase||Infectious agent||Catb||Ciprofloxacin (S)c and tobramycin (S)||Cured|
|6||42, F||AIDS||Pneumonia||Broncho-alveolar lavage||Pneumocystis jiroveci||2001||16s rDNA sequencing and oxidase||Co-pathogen||Unknown||Trimethoprim-sulfamethoxazole (S)||Cured|
|7||68, M||Severe COPD||Pneumonia||Broncho-alveolar lavage||None||1999||16s rDNA sequencing and oxidase||Infectious agent||Unknown||Amoxicilin-clavulanate (S)||Cured|
|8||35, M||AIDS||Bronchitis||Broncho-alveolar lavage||Pneumocystis jirovici||1993||16s rDNA sequencing and oxidase||Co-pathogen||Unknown||No data available||Cured|
The disease pattern related to B. bronchiseptica was not uniform. Except for one chronic obstructive pulmonary disease (COPD) patient who presented with colitis and septic shock unrelated to B. bronchiseptica airway colonization, three clinical presentations were observed: (i) two patients were asymptomatic carriers of B. bronchiseptica; (ii) two patients had symptoms of bronchitis; and (iii) three patients presented with severe pneumonia. In two of these latter cases, the pathogenic role of B. bronchiseptica was established (e.g. lung abscess). The remaining patient with pneumonia was co-infected with Pneumocystis jiroveci.
All available isolates (n = 4) were susceptible to amoxicillin-clavulanate, piperacilin-tazobactam, imipenem, amikacin, gentamicin, tobramycin, ciprofloxacin, and tigecyclin. Trimethoprim-sulfamethoxazole was susceptible in two of four cases only and all four isolates were resistant to erythromycin. Patients received various antibiotic treatment regimens (Table 1). In two patients with pneumonia and one patient with bronchitis, microbiological persistence of B. bronchiseptica was documented despite antimicrobial treatment.
Despite the inherent limitations of the retrospective design of the present study and its small sample size, several features emerge from this series of eight human B. bronchiseptica cases, which is one of the most comprehensive in the recent literature. Although we retrieved three cases in 2008, the overall small number of cases does not enable us to conclude an increasing incidence of human B. bronchiseptica cases because of possible random variation. The majority of our patients presented with a significant underlying disease, mainly COPD or AIDS. The only patient without an underlying condition presented an unusual location not previously described: B. bronchiseptica was identified on the maternal side of the placenta of an asymptomatic woman who was admitted for delivery by Caesarean section. A laboratory contamination could be reasonably excluded.
Our analysis confirms the tropism of B. bronchiseptica for the respiratory tract, leading to occasional upper airway colonization, as previously described . The only case of death observed was not caused by B. bronchiseptica because the patient was diagnosed with septic shock due to colitis.
Immunocompromised patients who own cats should be aware of this small but non-negligible threat of airway contamination with B. bronchiseptica. Although an animal vaccine for B. bronchiseptica exists , its efficacy remains controversial  and thus the vaccination of pets might not confer protection. Because of cross immunity, immunocompromised patients might, however, benefit from B. pertusssis vaccination, as shown in a mouse model .
Different antibiotic regimens were administered to our patients. This variety reflects the absence of reliable treatment recommendations. As a consequence, we cannot provide any firm conclusion on the optimal therapeutic approach. However, B. bronchiseptica are usually susceptible to anti-pseudomonal penicillins, carbapenems, flouroquinolones and aminoglycosides, but not to erythromycin, in contrast to other Bordetella species.
In summary, it remains difficult to clearly establish the pathogenic role of B. bronchiseptica in human disease as a result of the rare occurrence and challenging microbiological diagnosis. In the present study, we could reasonably establish the microbiological identification and pathogenic role of B. bronchiseptica and therefore explore its related clinical symptoms. On the basis of these cases, B. bronchiseptica should be considered as potentially pathogenic when found in the lower airways of immunocompromised patients.