Vertebral Osteomyelitis Caused by Enterococcus raffinosus

Abstract

Enterococcus raffinosus is a rare isolate in clinical specimens. A case of vertebral osteomyelitis caused by E. raffinosus in an elderly patient is described and confirms this organism to be an opportunistic human pathogen.

CASE REPORT

A 73-year-old woman presented with severe back pain and bilateral weakness and numbness in the legs. In addition, the patient had fecal incontinence for 24 h and had not passed urine for 2 days. During the preceding 4 weeks, she had experienced midthoracic back pain requiring increasing amounts of analgesia but had no difficulty walking and no altered sensation in the legs. No other symptoms of note were elicited. There was a history of rheumatoid arthritis, osteoarthritis (requiring bilateral knee replacements), ischemic heart disease, and a crush fracture of a thoracic vertebra (T5) following a fall 2 years previously. Among other medications, the patient was taking long-term low-dose prednisolone (5 mg once daily).

On examination, the patient was febrile (38.2°C) and had flaccid paralysis of the legs, a palpable bladder, and a grade 3 ejection systolic murmur. There were no other symptoms of endocarditis, and the patient was hemodynamically stable. An initial diagnosis of osteoporotic vertebral collapse and spinal cord compression was made. In view of the fever, blood and, later, urine cultures were taken. Hematological investigations revealed a hemoglobin level of 10.9 g/dl, a platelet count of 201 × 10⁹/liter, and a white cell count of 7.6 × 10⁹/liter. C-reactive protein was at 334 mg/liter. Plain roentgenograms of the spine showed wedging of thoracic vertebrae T6 and T7. Magnetic resonance imaging confirmed partial collapse of the T6 and T7 vertebral bodies, causing kyphosis and cord compression consistent with osteoporotic collapse. The patient was taken quickly to the operating room on the day of admission to decompress the spinal cord. During surgery, foul-smelling material was removed from the T6 and T7 vertebral bodies and sent for microscopy and culture. Anterior excision of the vertebral bodies was performed, and rib strut grafting (using the patient's own ribs, rather than a metal prosthesis) was used to bridge the bone defect, because of the clinical suspicion of infection. Gram-positive cocci were seen in stained tissue, and cephradine and fusidic acid were commenced as empirical antibiotic therapy. On the first postoperative day, blood cultures became positive with an Enterococcus species. Antibiotic susceptibilities were determined by E-test (AB Biodisk, Solna, Sweden), and the results are shown in Table 1. A history of penicillin allergy was given, and the therapy was therefore changed to intravenous vancomycin and metronidazole. On the fifth postoperative day, the vertebral body tissue specimens also grew an Enterococcus species. Metronidazole was discontinued when anaerobic cultures were negative. Urine cultures were negative on two occasions. Postoperative progress was slow: an intermittent low-grade fever and shivering episodes persisted, and the C-reactive protein failed to return to normal levels. The patient suffered episodes of cardiac failure and pulmonary embolism and died on the 45th postoperative day, in spite of apparently appropriate antibiotic therapy.

TABLE 1.

E-test MICs for standard E. faecalis (ATCC 29212) and E. raffinosus isolated from the bloodstream and from vertebral body pus

Antibiotic	MIC (mg/liter) for:
	E. faecalis ATCC 29212	E. raffinosus from the bloodstream	E. raffinosus from a vertebral body
Benzylpenicillin	1.0	0.75	0.75
Dalfopristin-quinupristin	6.0	3.0	3.0
Rifampin	0.38	0.5	0.5
Teicoplanin	0.19	0.19	0.19
Vancomycin	3.0	1.0	1.0

Specimens of bone tissue and swabs taken during the operation were plated onto MacConkey agar (Oxoid, Basingstoke, United Kingdom) incubated at 37°C in air, chocolate agar (Oxoid) incubated at 37°C in air supplemented with 5% CO₂, and horse blood agar (Oxoid) incubated at 37°C both in air and anaerobically. In addition, brain heart infusion broth (Oxoid) was inoculated and incubated at 37°C in air. After 24 h of incubation, grey, smooth, and entire nonhemolytic colonies, 1 mm in diameter, were evident on all plates and subsequently also isolated from the enrichment broth. On Gram stain, the isolate was found to be a gram-positive coccus, which formed pairs, singles, and short chains. Lancefield group D antiserum produced a positive reaction, while the catalase test was negative. Results of further biochemical characterization as described previously (4) are shown in Table 2. The identity of the isolate as Enterococcus raffinosus was subsequently confirmed by the Public Health Laboratory Service (Colindale, United Kingdom). Interestingly, the API 20 Strep identification strip (bioMérieux, Marcy l'Etoile, France) profile obtained (5146750) identified the isolate as E. avium with 99.9% probability. PCR amplification of the intraribosomal DNA variable spacer region was carried out as previously described (7). The PCR ribotype profiles obtained from both clinical isolates and the E. raffinosus standard (NCTC12192) were identical and distinct from the profiles produced by E. dispar, E. casseliflavus, E. durans, E. hirae, E. faecalis, and E. faecium (data not shown). Pulsed-field gel electrophoresis of the blood and bone isolates and the E. raffinosus standard was carried out essentially as previously described (5). Electrophoresis confirmed that the clinical isolates from blood and bone were indistinguishable (data not shown).

TABLE 2.

Biochemical characteristics of E. raffinosus isolated from vertebral body pus and a standard E. raffinosus isolate (NCTC12192) and API Strep (bioMérieux) reactions (profile obtained, 5146750)^a

Characteristic or test	Result for:		API 20 Strep result
	E. raffinosus vertebral body pus isolate	E. raffinosus (NCTC12192)
Growth
In NaCl-BHIb	+	+	ND
In bile-BHIc	+	+	ND
At 10°C in BHI	+	+	ND
At 45°C in BHI	+	+	ND
Utilization of:
Arabinose	+	+	+
Inulin	ND	ND	−
Glycogen	ND	ND	−
Mannitol	+	+	+
Methyl glucopyranoside	−	−	ND
Pyruvate	+	+	+
Raffinose	+	+	+
Ribose	ND	ND	+
Sorbitol	+	+	+
Sorbose	+	+	ND
Starch	ND	ND	−
Sucrose	+	+	ND
Trehalose	ND	ND	+
Hydrolysis of:
Esculin	+	+	+
Hippurate	ND	ND	−
Enzyme activity
Alkaline phosphatase	ND	ND	−
Arginine dihydrolase	−	−	−
Leucine arylamidase	ND	ND	+
Pyrrolidonyl arylamidase	+	+	+
α-Galactosidase	ND	ND	−
β-Galactosidase	ND	ND	−
β-Glucuronidase	ND	ND	−
Beta-haemolytic	−	−	−
Motile	−	−	ND
Pigmented	−	−	ND

^aND, not done; +, positive; −, negative. 

^bNaCl-BHI, brain heart infusion plus 6.5% NaCl at 37°C. 

^cBile-BHI, brain heart infusion plus 40% bovine bile salts at 37°C. 

E. raffinosus is a recently described species that is occasionally isolated from human clinical specimens (2). The natural habitat of E. raffinosus is unknown, but the organism has been described to occur among the oropharyngeal flora of domestic cats, second only to E. faecalis (3). E. raffinosus has been implicated as a cause of endocarditis and has been recovered from a urine specimen and wound swabs, but clinical and microbiological details confirming the ability of this bacterium to cause serious invasive disease have not been published (1, 4, 6). This report describes a case of vertebral osteomyelitis (VO) caused by E. raffinosus and the microbiological and diagnostic features.

Infections of bone caused by enterococci are rarely reported. Infective discitis due to enterococci has been reported on six previous occasions since 1969, each one associated with advanced age (8). The patient described herein was also elderly, had underlying medical problems, and may have been immunosuppressed secondary to steroid therapy. VO is almost invariably the result of hematogenous seeding and as such is frequently associated with underlying endocarditis. It is possible that the patient described in this report had underlying E. raffinosus endocarditis; however, there were no peripheral clinical features of endocarditis, and echocardiography was not performed. Negative urine cultures on two occasions make a urinary source unlikely. Two studies have provided epidemiological evidence that colonization with E. raffinosus can be associated with nosocomial acquisition (1). The infection described in this case was community acquired, and the source of the isolate is not known. It is not known if the patient kept any pets.

Identification of the enterococcal species was complicated by misidentification by the API 20 Strep system, not because the biochemical reactions were inconsistent with this organism but because E. raffinosus is not included in the database for this product. The system identified the isolate as E. avium, and although this error would not have affected patient treatment, it would have affected any epidemiological study that subsequently included this case. The absence of E. raffinosus from the API 20 Strep kit may also have contributed to the paucity of reported infections caused by this species. Empirical treatment for VO is directed towards the most common causes: Staphylococcus aureus and streptococci. Such regimens are frequently ineffective against enterococci, as occurred in this case, highlighting the importance of early invasive sampling to obtain a microbiological diagnosis, particularly in elderly patients. This is the first report of VO caused by E. raffinosus and confirms the pathogenic potential of this bacterium to cause human disease.