Arcanobacterium bernardiae Bacteremia in a Patient with Deep Soft Tissue Infection

To the Editor:

We read with great interest the review by Clarke et al. highlighting the uncertainties regarding the pathogenic potential of gram-positive bacilli, so-called “coryneforms,” in patients with complicated soft tissue infections [1]. The article describes a case of necrotizing fasciitis caused by Arcanobacterium bernardiae and Morganella morganii coinfection and gives a comprehensive review of the isolation of coryneforms from various infections.

There are several factors that complicate the placement of these microorganisms in the category of human pathogens: (1) They frequently are members of the normal flora of human skin and oropharynx and, therefore, are considered typical contaminants of microbiological specimens [2]; (2) they frequently are isolated in polymicrobial infections [3], so the virulence of the individual species remains obscure; (3) their taxonomic classification is unclear and has undergone various changes during the last decades [2]; and (4) the taxonomic identification to species or even genus level is impossible or at least unreliable by the routine methods of a bacteriology laboratory [4]. As Clarke et al. point out, a broader application of molecular biology or mass spectrometry for bacterial identification will help to clarify some of these issues [1].

We contribute to the discussion of the clinical relevance of coryneforms by reporting on a patient who suffered a bacteremic infection with A. bernardiae. The 72-year-old woman was admitted to the Military Hospital in Santiago, Chile, with a two-day history of fever, chills, anorexia, and progressive prostration. The patient, who suffered Alzheimer disease, lived with her family and had been hospitalized six months earlier because of an infection of her chronic pressure ulcers. On admission, she was in severely compromised general and mental condition, afebrile but with an arterial blood pressure of 93/66 mm Hg, tachycardia of 150 beats/min, and tachypnea of 40 beats/min. Physical examination revealed a large stage 4 pressure ulcer of the sacral region with indications of infection (purulent exudate, foul odor, surrounding inflammation), as well as signs of dehydration. Laboratory examinations showed a leukocyte count of 28,000/microliter (96% neutrophils), an elevated C-reactive protein (CRP) concentration of 209 mg/L (normal value < 10 mg/L), and hypernatremia of 157 mEq/L. Urine sediment and chest radiography were unremarkable. Blood and urine cultures were taken, and empiric antibiotic therapy with intravenous ceftriaxone (2 g qd) and metronidazole (500 mg q8h) was initiated. With antibiotic treatment and intravenous fluid replacement, the patient stabilized rapidly. During surgical debridement of the sacral ulcer that extended to the underlying bone, further microbiological samples were harvested.

After 48 h of incubation, both blood culture bottles were positive for gram-positive irregular (“coryneform”) bacilli, whereas urine cultures stayed negative. From the sample of the infected pressure lesion (taken two days after the initiation of antibiotic therapy), methicillin-resistant Staphylococcus aureus and Citrobacter freundii were isolated. Given the favorable clinical response that was accompanied by a marked reduction of the leukocyte and CRP values, the intravenous antibiotic schedule was changed after six days to oral treatment with amoxicillin/clavulanic acid 875/125 mg bid. Radiographs of the sacrum showed regional osteopenia compatible with osteitis. Pelvic magnetic resonance imaging to confirm the probable bone infection was not available. After 10 days, the patient was discharged with local therapy and—because of the probable bone involvement—a six-week course of oral antibiotics.

The strain isolated in blood cultures during the patient's acute illness grew on blood and chocolate but not on MacConkey agar. After 48 h of aerobic incubation at 37°C, small whitish pinpoint colonies resembling corynebacteria were visible. Colonies were non-hemolytic and catalase negative. Examination by the API-Coryne system (BioMérieux, Marcy l'Etoile, France) resulted in identification code 2410763, which is 98.3% compatible with Gardnerella vaginalis. The strain was sent to the Molecular Biology Laboratory of the Catholic University of Chile in Santiago for 16S rRNA gene sequencing. The DNA was extracted using the QI Amp DNA Mini Kit (Quiagen, Valencia, CA). Polymerase chain reaction (PCR) was performed using two universal primers for the amplification of the 16S rRNA gene. The PCR product was electrophoresed on 1.5% agarose gel and purified with a QIAquick purification kit (Quiagen). Purified DNA was sequenced directly using a Big Dye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Foster City, CA) on an ABI 310 sequencer (Applied Biosystems). The sequences detected were compared with sequences in the GenBank NCBI database (National Center for Biotechnology Information; http://www.ncbi.nlm.nih.gov) using BLAST software (Blast Internet Services, Pittboro, NC). By comparison of 500 basepairs of our strain with known sequences, a species match of > 99% with A. bernardiae was achieved.

Arcanobacterium bernardiae is the third species of the genus Arcanobacterium causing human infections, the other two being A. haemolyticum and A. pyogenes. Phylogenetically, it is closely related to A. pyogenes, but the colonies are smaller and do not produce hemolysis on sheep blood agar [5]. Little is known about the epidemiology and virulence of A. bernardiae. Its role as a probable opportunistic pathogen was first described in 1987 (as U.S. Centers for Disease Control and Prevention [CDC] group 2 coryneform bacteria [6]). Since then, apart from the case of necrotizing fasciitis described by Clarke et al. [1], only four more clinical cases have been reported. Two were complicated urinary tract infections, both occurring in patients with chronic bladder dysfunction and anatomic disruption of the urinary system [7, 8]. One patient also suffered from perirenal abscesses and bacteremia [7]. Furthermore, two infections of the osteoarticular system were published: One in an immunocompromised patient (systemic lupus erythematosus) suffering from septic arthritis [9] and a recent report of chronic osteitis of the knee (coinfection with S. aureus) [10].

The true incidence of infections with A. bernardiae and other gram-positive bacilli is unknown. When isolated in clinical samples, these bacteria often are classified as contaminants and not identified further. Our case clearly demonstrates the potential virulence of this Arcanobacterium species. Although our patient had a soft tissue co-infection with S. aureus and C. freundii, only A. bernardiae was isolated in blood cultures during the patient's sepsis. This result underscores the value of blood cultures in the microbiological diagnosis of severe soft tissue infections. Apart from our case, there is only one other clinical description of A. bernardiae bacteremia [7]. Furthermore, three strains derived from human blood samples have been included in biochemical and molecular studies, but clinical details of these strains are not available [5]. Another interesting factor of our case is the therapeutic response to beta-lactam treatment despite the fact that an MRSA strain was isolated from our patient's chronically infected ulcer. This is another point to demonstrate that, indeed, A. bernardiae and not S. aureus was the “leading pathogen” of our infection.

The identification of A. bernardiae with common routine tools is difficult and often misleading. Although this species is documented in the database of the API CORYNE system, it is misidentified frequently as Gardnerella vaginalis [8, 10]. A thicker inoculum might overcome this problem [10]. Because of their slow growth and unusual nutritional requirements, susceptibility testing of A. bernardiae isolates is difficult. Data are scarce, but according to Funke et al. [2], most strains are sensitive to clindamycin, erythromycin, beta-lactams, rifampin, tetracycline, and vancomycin; ciprofloxacin and gentamicin are less active.

Our report gives further proof of the ability of a less known Arcanobacterium species, A. bernardiae, to cause severe soft tissue infections that might be complicated by bacteremia and sepsis. Although it might not be easy to realize in daily microbiological routine, it seems important to keep a closer eye on all those “coryneforms” to clarify the conundrum of the gram-positive bacilli in skin and soft tissue infections, infections of bone and joints, and complicated urinary tract infections.