Treatment of a Staphylococcus lugdunensis cervical epidural abscess

Abstract

First reported in 1988, Staphylococcus lugdunensis is a virulent coagulase-negative Staphylococcus species often associated with endocarditis. This is the first case report describing this infection in the cervical spine. A 58-year-old, left-handed Caucasian woman with no significant medical history presented with neck and left arm pain. Neurological examination revealed mild left deltoid 4/5 weakness and myelopathy. She was found to have an epidural cervical spinal cord mass that was causing severe cord compression and underwent partial anterior cervical decompression of the mass. On entering what was believed to be the mass, yellow-tinged fluid was encountered. An abscess was immediately suspected, so the dura was not entered. The specimen sent for culture was identified a day later as S. lugdunensis. A two-dimensional echocardiogram was negative for vegetations. A peripherally inserted central catheter line was placed and the patient discharged on 6 weeks of physical therapy and 6–8 weeks of intravenous cefazolin.

Background

Staphylococcus lugdunensis is a unique coagulase-negative staphylococcus that resembles S. aureus in its pathogenicity.¹ It is a part of the normal human skin flora but is also associated with numerous infections including skin and soft tissue infection, prosthetic joint infection and endocarditis. It was first reported to cause infective endocarditis in 1988.² Due to its identification as a coagulase-negative Staphylococcus, the virulence of S. lugdunensis is often underestimated. Adherence factors and biofilm formation are thought to contribute to the destructive nature of S. lugdunensis and it has emerged as an important pathogen associated with high morbidity and mortality.³

S. lugdunensis has also been reported as a cause of brain abscess, peritonitis, joint and bone infections, vertebral osteomyelitis and discitis.⁴ It is also associated with infections in postsurgical, immunocompromised and immunocompetent patients. The most common risk factors for S. lugdunensis infections include use of catheter and synthetic devices, such as prosthetic joints, or after arthroscopic procedures and an immunocompromised state.³ Most commonly, infections with S. lugdunensis occur in a nosocomial setting. Outcome for infection with S. lugdunensis is highly sensitive to antibiotics but invasive infections have a high rate of associated morbidity and mortality. To date, there have been no published reports of an epidural spinal infection caused by S. lugdunensis. This is the first case report describing this infection in the cervical spine and its implications.

Case presentation

The patient is a 58-year-old woman with no significant medical history who presented to the emergency department in May complaining of intractable neck and left arm pain for 1 month. She had several emergency room visits during the month where she was prescribed pain medications and discharged home. The pain was constant, dull and did not fit any dermatomal distribution. She did not have any loss of bowel or bladder function.

The patient had no prior history, clinical or systemic symptoms consistent with, or any risk factors for tuberculosis. No foreign travel, no night sweats or TB contacts. She also denied any intravenous drug use.

Her neurological examination revealed mild left deltoid 4/5 weakness with 3+ reflexes bilaterally in all muscle groups with bilateral Hoffman and Babinski signs. She did not have any muscle wasting. There were no skin findings to suggest intravenous drug use.

Investigations

Her primary care physician eventually ordered MRI of the cervical spine, which showed a cystic structure within the anterior epidural space at C5/C6 resulting in moderate spinal canal stenosis (figure 1).

Figure 1.

Sagittal T2 MRI showing a ventral C5–C6 cervical epidural mass causing severe cord compression. There is a septation seen on axial imaging giving it a somewhat loculated appearance.

She was promptly admitted to the hospital from the clinic. She had blood cultures, which were negative. Two-dimensional echocardiogram revealed no vegetations. Erythrocyte sedimentation rate at the time of diagnosis was 57 mm/hour (reference 0–20 mm/hour) and C reactive protein was 1.5 g/dL (reference range <0.5 mg/dL). Both inflammatory markers had normalised when rechecked at the completion of intravenous antibiotics. A repeat CT of the cervical spine after completion of intravenous antibiotics demonstrated stable anterior cervical discectomy and fusion C5—without evidence of osseous fusion. There were no hardware complications.

Treatment

Intraoperatively, a C5 corpectomy was performed and the mass was identified (figure 2). It was dural-based, abutting the bone and suspicious for infection. As the mass was decompressed, it expressed yellow-clear fluid that resembled a burnt-out abscess. A frozen section sent to our pathology laboratory showed inflammatory changes with many neutrophils and no cancer cells. Although the dura was thickened, the decision was made to leave the dura intact as infection was suspected.

Figure 2.

Sagittal CT scan of the cervical spine showing a partial C5 corpectomy with C5–C6 anterior cervical discectomy and fusion.

Postoperatively, the patient was immediately started on intravenous vancomycin. The cultures were identified as S. lugdunensis and infectious disease was consulted. The patient was switched to cefazolin on postoperative day 3, with rifampin added on postoperative day 5. Sensitivities were confirmed. Infectious diseases recommended a total of 8 weeks of intravenous cefazolin (2 gm every 8 hours) and oral rifampin (600 mg every night), followed by cephalexin 500 mg by mouth twice daily for an additional 6 months. A two-dimensional echocardiogram was completed, which was negative for vegetations.

Outcome and follow-up

The patient was discharged home on postoperative day 5. She returned for numerous follow-up visits over the following 8 weeks with 100% medication compliance and resolution of both her symptoms and prior elevated inflammatory markers.

Discussion

This case report presents an unusual clinical presentation of S. lugdunensis infection. Unlike typical epidural infections, no associated risk factors, clinical features, imaging characteristics or external sources of infection were identified. Furthermore, S. lugdunensis has not been previously reported to infect the epidural space.

Urgent neurosurgical intervention is generally indicated when a compressive epidural abscess results in a progressive neurological deficit. In a neurologically intact patient without spinal instability, an epidural infection may be treated with antibiotics and close neurological monitoring without surgical intervention if cultures and sensitivity are available.⁵ In this case presentation, the absence of associated risk factors, clinical characteristics or radiological features specific to infection made surgical intervention important for both successful diagnosis and treatment.

Established risk factors for epidural infections include diabetes, immunosuppressive therapy, HIV, malignancy, renal insufficiency and spinal trauma.⁶ Urological infections are known to reflux into the epidural space via Batson’s plexus.⁷ Endocarditis can result in haematogenous seeding resulting in either primary epidural infection or spread from associated psoas abscess, diskitis and osteomyelitis. Systemic markers of infection may be elevated, such as white blood cell count, erythrocyte sedimentation rate, C reactive protein and body temperature.⁶ On MRI, epidural abscesses usually enhance and are commonly associated with adjacent diskitis, osteomyelitis and paraspinal muscle abscess.⁸ When adjacent foci of infection are present and blood cultures are negative, image-guided biopsy is often pursued to provide cultures. In this case, image-guided biopsy was not possible due to the location of enhancing tissue and need for surgical decompression.

S. lugdunensis is a unique coagulase-negative staphylococcus that exhibits virulence far more severe than S. epidermidis. Instead, its pathogenicity is comparable to S. aureus. ¹ While the full extent of pathogenic mechanisms by which S. lugdunensis causes severe infections is not yet completely understood, the organism produces numerous virulence factors, adhesions and biofilm.⁹ To the best of our knowledge, central nervous system infection with this organism is limited to patients with previous neurosurgical procedures. The presence of hardware can increase the risk for biofilm formation, such as in the case of shunt infection.¹⁰ Apart from this, there is one case report of a septic embolus from S. lugdunensis endocarditis resulting in acute stroke.¹¹

Most S. lugdunensis strains are beta-lactam susceptible. In a series of 70 isolates, 80% were beta-lactamase resistant and 10% were oxacillin resistant.¹² The incidence of S. lugdunensis infection may be under-reported as some laboratories do not routinely identify coagulase-negative staphylococcus.¹³ Unlike S. epidermidis, S. lugdunensis should be presumed to be a true pathogen in most cases. Among 229 S. lugdunensis clinical isolates in one series, less than 15% were considered contaminants or colonising organisms.¹⁴

Although the pathophysiology of the infection in this patient was unclear, colonisation of the skin was the most likely source of a suspected transient bacteraemia that resulted in haematogenous seeding of the spine. The patient did not have any history of skin trauma or wounds, no lines or recent hardware placement prior to presentation. Rifampin was added in this case as all Staphylococcus species are known to form biofilm, and this patient required new hardware placement. Although lifelong antibiotic suppression was considered for the newly placed indwelling hardware, ultimately it was felt unnecessary because the hardware was not placed until after extensive and careful debridement of the infection.

There is no definitive infectious disease consensus regarding the optimal treatment duration of vertebral osteomyelitis with hardware placement. In native vertebral osteomyelitis, current Infectious Disease Society of America (IDSA) guidelines recommend at least 6 weeks of intravenous antibiotics, with some committee members and other publications recommending longer courses of intravenous antibiotics (6–12 weeks) followed by at least 3 months of oral suppression antibiotics when S. aureus is involved.¹⁵ When hardware is involved, cure is even more difficult to achieve in staphylococcus infections due to biofilm formation and antibiotic duration is generally lengthened. Current IDSA guidelines involving hardware joint infections recommend at least 6 weeks of intravenous antibiotics combined with rifampin, followed by 6 months of oral suppressive antibiotics. In this case, infectious disease consult recommended a total of 8 weeks of intravenous cefazolin (2 g every 8 hours) combined with oral rifampin (600 mg before sleep) followed by oral cephalexin (500 mg twice daily) for an additional 6 months. These recommendations were within the suggested length of treatment per IDSA guidelines as this patient required placement of new hardware for a cervical corpectomy into the infected bone at the time of diagnosis.

Learning points.

• When considering epidural abscesses as pathology, it is important to obtain intraoperative cultures and avoid intentional or unintentional durotomy.

• Because of the destructive nature of this particular species, we suggest all coagulase-negative Staphylococcus spinal infections be fully speciated in the laboratory as identification of Staphylococcus lugdunensis would alter both prognosis and treatment.

• Due to the particular virulence of S. lugdunensis, we recommend an aggressive approach with close follow-up and repeat imaging to ensure eradication of infection.

• We suggest S. lugdunensis spinal osteomyelitis with hardware involvement be treated similarly to S. aureus, with 6–8 weeks of intravenous antibiotics combined with oral rifampin, followed by de-escalation to oral antibiotic suppression for 3–6 months.