Erysipelas of the right arm due to Bordetella trematum: a case report

The genus Bordetella belongs to the Alcaligenaceae family and includes dozens of species [1, 2] mostly responsible for bronchopulmonary infections in mammals. The agent of whooping cough, Bordetella pertussis, is the most famous of Bordetella species. Bordetella trematum was identified in 1996 from chronic media otitis and chronic ulcer [3]. This germ is difficult to isolate, and its pathogenicity remains debated in humans. An increasing number of infections due to B. trematum are described in the literature as a potential emerging pathogen. In this context, we report the case of erysipelas due to B. trematum associated with bacteremia.

An 88-year-old Caucasian male patient had a history of chronic kidney failure, stress angina with quadruple coronary bypass surgeries, removal of multiple skin carcinomas, and chronic lymphocytic leukemia treated for a year by chloraminophene in 2012. This former carpenter, who normally lived alone, presented himself to emergency unit for repeated falls and erythematous edema of the right arm since 3 days suggesting erysipelas (Fig. 1). On admission, his temperature was 39.7 °C, his pulse was at 107 beats per minute, and blood pressure was 131/54 mmHg. He presented erysipelas of the right arm with edema and without any adenopathy. There was no other clinical sign at admission. The blood test found leukocytes at 156 × 10⁹/L consisting of 12.8 × 10⁹/L neutrophils and 140 × 10⁹/L lymphocytes, serum biochemical analysis with creatinine level at 160 µmol/L, hyperkalemia at 5.3 mmol/L, and elevated C-reactive protein (CRP) at 271 mg/L. One set of blood culture was collected from a peripheral access prior to administration of amoxicillin, and the patient was transferred to the geriatric unit. Aerobic blood culture grew with small, gray, shiny, rounded colonies in polyvitex and blood agar plate (Fig. 2a, b), which were Gram-negative coccobacilli (Fig. 2c) after 18 h at 35° CO₂ on conventional medium. B. trematum was identified using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) and confirmed by 16S ribosomal RNA sequencing (Additional file 1). Another blood culture was performed 2 days after the first one and while taking antibiotic therapy and was negative. No additional tests were carried out. Diagnosis of thrombosis was eliminated by venous ultrasonography and Doppler of his right arm.

After a multidisciplinary meeting involving geriatricians, infectious diseases specialists, and microbiologists, antibiotic therapy was changed to ceftazidime 1 g per day for 10 days after a charging dose of 2 g, slow intravenous, according to susceptibility testing (Table 2) and renal function at that time. The clinicobiological evolution was rapidly favorable. The patient was discharged from hospital and referred to hematologists on the basis of white blood cell count, and finally, the chemotherapy by chloraminophene was resumed. He was rehospitalized for depression the next month, at which point his arm was totally healed. At the last follow-up, 6 months after the first symptoms, the patient was in good health and has not had any recurrence of infection.

Bordetella trematum is a small, mobile, capsulated, nonsporulating, and nonfermentative Gram-negative bacillus. The term trema refers to something pierced or penetrated like wounds, and inspired its name [1]. Little is known about its pathophysiology because it is scarce and quite recently identified. It is genetically close to B. avium and frequently mistaken for it [4]. The supposed ecological reservoir of B. trematum is soil [5], and most cases involve skin and soft-tissue infections. Chang et al. have revealed a cytolethal distending toxin (CDT) [6]. This genotoxin has only been described in B. trematum, which may partly explain the ineffectiveness of antibiotic treatment in some patients. It could be interesting to better characterize the role of this genotoxin in the specific virulence of B. trematum, and evaluate the specific effect of antitoxin antibiotic drugs such as macrolides.

The interpretation of a positive blood culture of B. trematum remains difficult and, in the context of sepsis, should prompt a suitable antibiotic therapy. Thanks to new diagnostic tools such as MALDI-TOF and molecular biology, it is now easier to identify [7‐12], though there are still mistakes in its recognition.

Bordetella trematum infections have been identified in 14 case reports in the literature (12 are summarized in Table 1). Most cases were found in diabetic patients, particularly from ulcers and chronic ear infections [1, 4, 7, 9, 12]. Several authors consider that this germ is not very pathogenic and that its natural evolution does not require specific antibiotherapy [4, 7].

In contrast, as reported with our patient, invasive infection seems to occur mostly in immunocompromised patient [8, 13, 14]. Thereby, B. trematum could be considered as an opportunistic agent. However, some cases are reported in immunocompetent patients such as in a 7-month-old child [11] or in a 54-year-old patient with peritonitis [10].

Once the bacteria have been identified, treatment is not consensual. Indeed, there are no data about the antibiotic susceptibility of B. trematum in European guidelines (European Committee on Antimicrobial Susceptibility Testing, EUCAST 2018). We have reported in Table 2 the different antibiotic susceptibility tests described in the literature. There were interpreted according to minimal inhibitory concentration (MIC) interpretative standards of closely related species (other non-Enterobacteriaceae and Enterobacteriaceae). The choice of ceftazidime in our case was made according to the sensitivity testing. Despite the piperacillin–tazobactam sensitivity in vitro in all precedent case reports, patients died despite piperacillin–tazobactam therapy in two of them [12, 13] and had an unfavorable outcome in another one [11], whereas, in the case of Hernandez-Porto et al., the patient had also renal failure and the treatment was a success. Given these discordant results, the place of piperacillin–tazobactam in the therapeutic strategy for the treatment of B. trematum infections should be reconsidered.

Furthermore, clinicians should remain cautious, especially in immunocompromised patient, when presented with atypical skin infection.