Infective Spondylitis with Epidural Abscess Formation Caused by Roseomonas mucosa: A Case Report and Literature Review

Roseomonas mucosa (R. mucosa) is a pink-pigmented, aerobic, nonfermentative, slow-growing Gram-negative coccus typically isolated from the natural environment, human skin, and hospital environment. This pathogen, in most circumstances, leads to infections in immunocompromised hosts, but it may sometimes invade immunocompetent individuals. Bacteraemia is the most common form of infection caused by R. mucosa. In contrast, only two case reports have described R. mucosa-related epidural abscess formation and infective spondylitis. In this case report, we shared the history and treatment experience of a 76-year-old female who was diagnosed with infective spondylitis and epidural abscess caused by R. mucosa. She received a local transdermal injection into the lower back to relieve her back pain two months before symptom onset, which was considered to be associated with this infection episode. After admission to the hospital, neurosurgeons performed emergent decompression and debridement. She was treated with intravenous ceftriaxone for four weeks, followed by oral ciprofloxacin for another eight weeks. The patient recovered well without any sequelae and had no relapse of infection at least six months after the end of treatment. In addition to the case report, we reviewed the literature for reported cases caused by R. mucosa. Our experience suggests that clinicians should include R. mucosa as one of the possible healthcare-associated pathogens among individuals who have undergone transdermal procedures. We believe that this article will help clinicians better recognize R. mucosa infection.


Introduction
Te Roseomonas species was frst documented in 1993 [1], and in 2003, R. mucosa was established as a separate species from Roseomonas gilardii because of its distinct genotype and phenotype [2]. Tis Gram-negative coccus has been isolated from the natural environment [3][4][5][6], human skin [7], and hospital environment [8]. It is generally believed that people infected with R. mucosa are primarily immunocompromised [9] or have underlying diseases [10]. However, a few reports have revealed that R. mucosa also infects immunocompetent individuals and might lead to severe infections [11].
Epidural abscess is a severe central nervous system infection [12] and may lead to fever, back pain, neurological defcits, impaired motor function, paralysis, and even death [13]. R. mucosa is an infrequent cause of epidural abscess formation. In this case report, we present the disease course of a female diagnosed with infective spondylitis and epidural abscess formation caused by R. mucosa. In addition, we reviewed the literature and listed all published articles on R. mucosa infection.

Clinical Course.
A 76-year-old female with independent daily activity presented with progressive low back pain for fve days and fever for three days. She had a medical history of type 2 diabetes mellitus under reasonable medical control and left breast cancer, for which she had undergone surgical resection fve years prior.
Two years prior, she was diagnosed with spinal retrolisthesis over L2-L5. In addition, an orthopaedist noted a collapsed disc at L3-L4 with bilateral nerve root compression. She had been experiencing intermittent low back pain and numbness over the left leg since that time and sought rehabilitation, taking painkillers as needed. Orthopaedics did not suggest surgical intervention in consideration of her severe osteoporosis and independent daily activity status. To relieve the low back pain, the patient sought various treatment modalities over the past two years. Two months before admission, she underwent a so-called "nerve block" procedure by receiving a local transdermal injection into her lower back. One month prior, she received radiofrequency coagulation pain therapy at a local hospital.
Her low back pain became more severe and unbearable fve days before this admission. She fell at home three days prior and was admitted to a local hospital. Her body temperature was 39.0°C upon admission, with otherwise typical vital signs. Because of the severe painful disability, she could no longer stand and laid in bed all day. She was also put on a Foley catheter due to difculty in urination. Under the suspicion of infective spondylitis, contrast L-spine magnetic resonance imaging (MRI) was performed and revealed an L3-L4 collapsed disc with nerve compression, which was not signifcantly diferent from previous fndings. Having persistent fever and progressive low back pain, she was transferred to Kaohsiung Veterans General Hospital (KVGH), a tertiary teaching hospital, for further management.
Upon admission to KVGH, her vital signs were as follows: blood pressure 132/63 mmHg, pulse rate 74/min, respiratory rate 17/min, body temperature 38.4°C, and oxygen saturation 97%. On physical examination, her consciousness was clear, and there was a marked tender point over her lower back. Te muscle power over the bilateral lower limbs was symmetrically decreased (grade 2 out of 5), while the muscle power over the bilateral upper limbs was intact. A bilateral straight leg raise test was positive, and the deep tendon refex over the bilateral lower limbs was normal. Laboratory data reported leukocytosis, elevated erythrocyte sedimentation rate and C-reactive protein (15.38 mg/dl). Urine analysis showed pyuria, and urine culture yielded no pathogen. Under a high suspicion of infective spondylitis with/without epidural abscess formation, cefepime and vancomycin were given as empirical antibiotics and an MRI was repeated on hospital Day 3. Te images showed a focal, small, lobulated fuid collection at the posterior epidural space of L4-L5, with adjacent soft tissue swelling at the epidural space and posterior column of the L4-L5 level. Infective arthritis over the L4-L5 facet joint with adjacent epidural abscess formation leading to severe spinal stenosis at the L4-L5 level was concluded by the radiologist (Figure 1).
Emergent partial laminectomy of L3-L4 and internal decompression of L4-L5 was therefore conducted for debridement and decompression of the lesions. Some dirty fuid and granulation tissue collected during the surgery later yielded R. mucosa, which was identifed by matrix-assisted laser desorption/ionization time of fight mass spectrometry (MALDI-TOF MS). Both fungal and mycobacterial cultures of the lesions were negative. An antimicrobial sensitivity test of R. mucosa was conducted by the E-test method and showed susceptibility to imipenem (0.5 µg/ml), ceftriaxone (1.5 µg/ml), amikacin (0.5 µg/ml), and ciprofoxacin (0.125 µg/ml). Te pathological report described a bony tissue with acute and chronic infammatory cell infltration, bone necrosis, and focal granulation tissue, which is compatible with osteomyelitis.
After the spinal surgery, there was a notable improvement in the patient's muscle strength. We utilized the Medical Research Council scale to assess the patient's muscle power, which scores muscle strength from 0 to 5, with 0 indicating no contraction, 1 indicating trace contraction, 2 indicating active movement with gravity eliminated, 3 indicating active movement against gravity, 4 indicating active movement against resistance, and 5 indicating normal power. Te patient's muscle power in both lower extremities increased from grade 2 to 4 within two days following the surgery. However, during a follow-up MRI, an abscess measuring 1.5 cm in diameter was discovered at the L3-L4 level. In addition, an early stage of infective spondylodiscitis was noted at L4-L5 ( Figure 2). Sonography-guided aspiration of the abscess was attempted but failed due to the small size of the lesion. We treated the patient with intravenous ceftriaxone 2 gm q12 h for four weeks during the hospital course. Upon discharge from the hospital, her muscle power was grade 4 over the bilateral lower limbs. She could walk under the assistance of a four-foot walker but could not stand up by herself. Te patient received another two months of oral antibiotics of ciprofoxacin 500 mg PO q12 h at the outpatient department. She recovered three months after the infection. Her muscle power in both lower limbs was graded as 5. She was able to walk with the assistance of a four-foot walker and could stand up independently. Although she still experienced some soreness in her left leg, there was no mention of any low back pain. Tere was no relapse of infection at least six months after the end of treatment.

Laboratory Method.
Te laboratory of KVGH inoculated the epidural abscess specimen onto blood agar, eosin methylene blue (EMB) agar, chocolate agar, Columbia C.N.A. agar, and thioglycolate broth. Te following day, pink colonies were observed on EMB agar. After subsequent subculture and identifcation using MALDI-TOF MS, the colonies were identifed as Roseomonas mucosa. Turbidity was also observed in the thioglycolate broth the following day, and subsequent subculture and identifcation yielded the same result of Roseomonas mucosa.
Te KVGH laboratory routinely uses VITEK MS MAL-DI-TOF (BioMérieux, France) to identify bacterial species. Te process for identifying bacteria using MALDI-TOF MS is as follows. First, a toothpick is used to pick up a colony from the agar plate and spread it evenly onto the center of a sample plate to form a thin bacterial flm. Ten, 1 μl of VITEK MS-CHCA matrix solution is pipetted on the bacterial flm, allowed to dry and then proceeded for identifcation. Te strain of Roseomonas mucosa was identifed using MALDI-TOF MS three times, with consistent results indicating a single bacterium and a high confdence value of 99.9.
Te KVGH laboratory routinely uses the VITEK 2 system (BioMérieux, France) for antibiotic susceptibility testing. However, in the case of this Roseomonas mucosa, VITEK 2 system failed to generate the antibiotic susceptibility testing result. Terefore, the E-test method was used instead, with the following procedure: frst, 1-2 colonies were picked up from the agar using a sterile cotton swab and suspended in tryptic soy broth adjusted to McFarland No. 0.5 turbidity. Ten, the swab was used to streak the Mueller-Hinton agar in three directions and allowed to dry for 2-3 minutes. Te E-test strip was then carefully placed on the agar surface and incubated in a non-CO 2 35°C ± 2°C incubator for 16-20 hours before interpreting. Antimicrobial susceptibility was interpreted according to the latest Clinical and Laboratory Standards Institute Guideline (M100-S 31).

Discussion
In the past, it was generally believed that R. mucosa is not highly pathogenic to immunocompetent populations. A systematic review [9] suggests that R. mucosa infections occur more frequently in immunocompromised patients or patients with severe comorbidities, including dialysis [14,15], cancer [16,17], HIV [18], and autoimmune disease [10]. In addition, the placement of catheter should be considered a risk factor [19,20], and the assessment of whether to remove the relevant catheter after infection is considered an essential part of the treatment [21]. However, there is a growing body of the literature reporting that R. mucosa can potentially cause severe infections in immunocompetent patients [11,22]. Te most common infection caused by R. mucosa is bacteraemia. Other types of infections, although uncommon, have been reported sporadically (Table 1). We searched PubMed using the keyword combinations "Roseomonas," "Roseomonas mucosa," "epidural abscess," and "infective spondylitis." Only one case report described an epidural abscess due to R. mucosa infection [27]; another article reported a male with infective spondylitis caused by R. mucosa infection [11]. To the best of our knowledge, this article is the second reported case of infective spondylitis and epidural abscess caused by R. mucosa in the literature.     According to previous reports, R. mucosa is usually resistant to penicillins and cephalosporins and susceptible to carbapenems, tetracyclines, fuoroquinolones, and aminoglycosides [9,35]. Our drug susceptibility tests of the R. mucosa isolate were similar to the literature. We treated the patient with intravenous ceftriaxone for four weeks, followed by oral ciprofoxacin for another eight weeks. Te outcome was good, and there was no relapse of infection at least six months after the end of treatment.
R. mucosa is an easily misidentifed species. A previous study [36] compared the identifcation ability of three routine laboratory bacterial identifcation methods for R. mucosa, including VITEK 2, MALDI-TOF MS, and 16S rRNA gene sequencing. Te VITEK 2 system could misidentify the isolate as Rhizobium radiobacter, Sphingomonas paucimobilis, or R. gilardii [10,27]. In comparison, MAL-DI-TOF MS and 16S rRNA gene sequencing had a more reliable identifcation ability. Generally speaking, 16S rRNA sequencing is often considered a superior method to MALDI-TOF MS for bacterial identifcation, as it provides greater taxonomic resolution at the species level. However, MALDI-TOF MS can also provide accurate identifcation of bacteria by comparing the mass-to-charge ratio and the mass spectra obtained by measuring the time of the bacteria proteins, which have been laser ionized, fight to the detector with those in a reference database [37]. Te main reason for using MALDI-TOF in this current report is that it is the routine method in the KVGH laboratory and is most relevant to clinical practice. After all, 16S rRNA sequencing is mostly used for research purposes rather than clinical applications. In addition, MALDI-TOF is relatively inexpensive, has a higher accessibility, and can provide results within a few hours.
After compiling much of the literature on the subject, we have summarized the valuable lessons learned from this case and our treatment advice for similar cases. First, according to previous reports, R. mucosa is present in human skin [7] and medical settings [8]. A previous report [27] described that invasive interventions might cause skeletal muscle infections caused by the Roseomonas species. Our patient underwent several medical procedures before the onset of infection, including nerve block with local lidocaine injection and radiofrequency coagulation. We suspected that this infection episode was associated with one of the medical procedures, especially nerve block injection. Terefore, we suggest that among individuals who have healthcareassociated infections, especially those who have undergone transdermal procedures, clinicians should include R. mucosa as one of the possible pathogens [8]. Second, to identify the type of microorganism infected, we recommend using MALDI-TOF MS or 16S rRNA gene sequencing to avoid misidentifying strains. Finally, for the treatment, we recommend that once the location of the infection is identifed, treatment should prioritize the removal of the source of the infection. Besides, since the antibiotics recommended by the Infectious Diseases Society of America for the prevention of postsurgical skin and soft tissue infections are usually not efective against R. mucosa [38], we recommend the use of carbapenems, tetracycline, fuoroquinolone, or aminoglycosides when R. mucosa infection is identifed, as R. mucosa is usually sensitive to these antibiotics.
Tere are two major limitations to this report. Te frst limitation is related to the method of bacterial identifcation utilized. Instead of using the gold standard of 16S rRNA sequencing, we relied on the clinically common MALDI--TOF MS method. While MALDI-TOF MS is commonly used in clinical settings and provides rapid results, it may not provide the same level of taxonomic resolution as 16S rRNA sequencing. Te second limitation is that although we have reviewed most of the literature on the diseases related to R. mucosa, we only encountered one actual case, and our clinical experience with this species is still insufcient.
In conclusion, although rare, R. mucosa can cause epidural abscess formation and infective spondylitis in immunocompetent hosts. Since R. mucosa has fewer antibiotic choices, clinicians should arrange surgical removal or drainage as soon as possible once abscess formation occurs. In addition, R. mucosa should be considered a pathogen associated with medical procedures.

Data Availability
Te health record data used to support the fndings of this case report are restricted to ensure patient privacy. Specifc health record data have already been included within the article in a manner that does not compromise patient confdentiality.

Ethical Approval
Te Institutional Review Board of Kaohsiung Veterans General Hospital approved the study protocol (KSVGH22-CT6-02).

Consent
Te informed consent was waived due to the retrospective nature of this report and minimal risks to the participant.

Conflicts of Interest
Kuan-Sheng Wu has received grants from Kaohsiung Veterans General Hospital to conduct this study.

Authors' Contributions
Zong-Han Lin and Yu-Chang Lu contributed equally to this article.