Background

Capnocytophaga canimorsus is a gram-negative bacillus found in saliva of healthy dogs and cats and is transmitted to humans principally through animal bites [1]. It can cause sepsis and other forms of infection. Here, we report a patient with sepsis and infective endocarditis (IE) caused by C. canimorsus. As C. canimorsus IE is rare and this microbe is difficult to culture, drug susceptibility is often unclear and its standard treatment regimen remains unestablished.

Case presentation

A 46-year-old man with a history of dog-bite in his left hand 3 months ago, developed fever (body temperature > 38 °C), visual disturbance, and dyspnea at rest. He had been otherwise healthy without significant medical history. He was tachycardic, and coarse crackle and diastolic heart murmur (Levine III) was audible. Laboratory test results were as follows: white blood cell count, 10,500/μL (59.8% neutrophils); hemoglobin level, 11.7 g/dL; brain natriuretic protein level, 689.2 pg/mL; and C-reactive protein level, 9.0 mg/dL (Table 1). Chest X ray showed pulmonary congestion and bilateral pleural effusion. Brain magnetic resonance imaging revealed no lesion in optic nerve and brain. Transthoracic echocardiography revealed moderate-to-severe aortic valve regurgitation and vegetation of 17-mm in size (Fig. 1). Seven days later, blood culture yielded Coagulase-negative staphylococci in one of four culture bottles. Although diagnosis of IE was not definitive according to Duke criteria [2], history of dog bites, his clinical course, and imaging studies suggested Staphylococcal IE. Following administration of cefazolin 6 g/day and gentamycin 3 mg/kg/day for a week, the patient underwent aortic valve replacement and resected aortic valve was negative for Staphylococci. A week following surgery, however, microorganism grew in two bottles of preoperative blood culture. This microorganism was cultured on blood agar, and gram staining of the colonies showed Capnocytophaga-like gram-negative bacilli (Fig. 2). 16S ribosome DNA sequencing both from blood and from resected heart valve identified C. canimorsus. Disk diffusion test revealed that the isolate was susceptible to almost all antimicrobial agents and did not produce β-lactamase (Table 2). The protocol of the disk diffusion test was as follows: A Brucella HK agar plate was seeded with a lawn of C. canimorsus using sterile cotton swabs. For the plate, antibiotic disks containing 10 IU of penicillin G, 10 μg/10 μg of sulbactam/ampicillin, 10 μg/100 μg of tazobactam/piperacillin, 30 μg of ceftriaxone, 10 μg of meropenem, 10 μg of gentamycin, 30 μg of amikacin, 5 μg of levofloxacin, 30 μg of minocycline, 250 μg of sulfamethoxazole/trimethoprim, 15 μg of clarithromycin, 2 μg of clindamycin were used with BD Sensi-Disc (BD Bioscience Co., USA) and dispensed on the agar surface. Both plates were incubated at 30 °C overnight and the diameter of each zone was measured in millimeters to evaluate susceptibility or resistance using the comparative standard method.

Table 1 Laboratory data on admission
Fig. 1
figure 1

Echocardiogram showing moderate-to-severe aortic valve regurgitation and vegetation of 17-mm in size

Fig. 2
figure 2

Capnocytophaga-like gram-negative bacilli on the aortic valve (× 1000)

Table 2 Drug susceptibility shown by disk diffusion method

Based on these results and symptoms, empirically selected combination of gentamycin and cefazolin was converted to ceftriaxone 4 g/day. The patient completed a total of 4 weeks of ceftriaxone. The patient has been doing well for 12 months after hospital discharge.

Discussion and conclusion

C. canimorsus is a less virulent pathogen. IE accounts for less than 2% of C. canimorsus bloodstream infection and is extremely rare [3]. Only 18 cases have been reported in the literatures since 1977 (Table 3) [4,5,6,7,8,9,10]. Patients were 52.8 years of age (range 24 to 73 years) on average and were predominantly male (80.0%). Affected valves were aortic in 11 (61.1%), tricuspid in six (33.3%), and mitral in four (22.2%). Nine patients (50.0%) were surgically treated, mostly using mechanical valves. Penicillin was given in eight (44.4%), and Cephalosporin in four (22.2%). Four patients (22.2%) had underlying cardiac diseases, and five (27.7%) were vulnerable to infection; alcohol abuse in four and chronic lymphocytic leukemia undergoing chemotherapy in one. Twelve of 18 patients (66.6%) had dog bite or close contact with dogs.

Table 3 Infective endocarditis caused by Capnocytophaga canimorsus in literature

C. canimorsus is a facultative anaerobe and grows slowly in blood culture bottles and on agar plates. It has fastidious requirements for growth (5-10% CO2) and efficient culture method has not yet been established. Diagnosis of C. canimorsus IE generally requires high indices of suspicion because clinical symptoms are non-specific and routine blood cultures are often negative. If a pet owner or an immunocompromised host develops IE and blood culture is initially negative, therefore, longer incubation or terminal subculture should be considered. In addition, Polymerase chain reaction and sequencing for 16S rDNA is useful to identify C. canimorsus [11, 12].

Since IE is a life threatening illness, antibiotic treatment often needs to be commenced before causative organism is identified. Aminoglycosides and/or β-lactam antibiotics are common empirical drugs of choice. However, almost all strains of C. canimorsus are resistant to aminoglycosides [13]. Decades ago, β-lactamase-producing Capnocytophaga was less than 2% [14], but recent papers suggest such strains have remarkably increased and account for 32% [15] or 79% [16]. So far, prognosis of C. canimorsus IE is poor chiefly due to delay in diagnosis and suboptimal drug choice. During treatment for C. canimorsus IE, therefore, addition of β-lactamase-inhibitor might be beneficial. In the present case, we chose Ceftriaxone soon after extended culture yielded gram negative bacilli. As disk diffusion test showed the strain was sensitive to β-lactam antibiotics, Ceftriaxone was continued until completion.

In conclusion, C. canimorsus is a fastidious and slow-growing microbe. C. canimorsus IE shows no specific findings but this pathogen should be kept in mind especially when pet owners show fever of unknown origin. Longer incubation along with some molecular biological diagnostic methods should be considered. Because diagnosis of C. canimorsus IE is often delayed and β-lactam tolerance is relatively common, its prognosis is not good. Prompt antimicrobial susceptibility test is essential.