- Epidemiology & risk factors
- Clinical findings
- Laboratory studies
- Imaging studies
- Antibiotic selection & timing
- Interventional management
- Questions & discussion
route of spread
- Improved imaging techniques have demonstrated that the majority of epidural abscesses occur secondary to tracking from a nearby infected musculoskeletal structure (e.g., septic facet joint, discitis, vertebral osteomyelitis).(31021957) Infection may also track from more distal structures (e.g., psoas abscess, retropharyngeal abscess, perinephric abscess) or be iatrogenically introduced during procedures.
- Hematogenous spread of bacteremia (e.g., from endocarditis) can certainly lead to spinal epidural abscess – perhaps in about half of cases. However, it appears that bacteria first affect the musculoskeletal components of the spinal cord (e.g., discitis or osteomyelitis), and only then spread to cause a spinal epidural abscess.
- As with brain abscess, in roughly a third of cases the source of infection remains unknown (cryptogenic).(30273242)
anatomy of SEA
- Abscess occurs in the epidural space, a space between the vertebrae and the dura mater (image above). Within the skull, the epidural space is a potential space (since the dura is normally adherent to the skull). However, in the spinal cord the epidural space is normally a real space that contains fat and blood vessels (this explains why epidural abscesses are ten times more common in the spinal cord than within the skull.)
- Abscesses span 3-4 vertebral levels on average, but they can extend for longer lengths along the spinal cord.(32694325)
- Abscesses most often occur in the dorsal (posterior) thoracolumbar spine (likely because the epidural space is larger in the thoracic and lumbar segments).(34623097)
- The thoracic spine is most often affected, followed by the lumbar and then the cervical regions.(35063888)
mechanism of neurological damage
- 🗜 Despite some ongoing controversy, the primary mechanism of spinal cord damage seems to be direct compression of the cord.(31021957)
- Abscesses in the cervical or thoracic cord may pose the greatest danger, as there is little extra space around the spinal cord in these areas. Alternatively, abscesses occurring below conus medullaris at L1 may pose less risk of compression.(31021957)
- Prompt surgical decompression may alleviate pressure on the cord, leading to reversal of neurological deficits which are less than ~36 hours old.(31021957)
epidemiology & risk factors
risk factors for SEA:
- Risk factors for bacteremia:
- IV drug abuse (yet this accounts for only ~20% of cases).
- Chronic indwelling venous catheter (e.g., hemodialysis, chemotherapy).(31021957)
- Spinal procedures:
- Spinal steroid injection.
- Lumbar puncture.
- Epidural anesthesia (especially indwelling epidural catheters).
- Spinal surgery.
- Indwelling spinal hardware.
- Immunosuppressive comorbidities, including:
- Diabetes (may be single most important risk factor).(Louis 2021)
- Chronic liver or kidney dysfunction.
- Immunosuppressive therapy (including steroid).
- Overall rate of SEA is increasing over time due to:
- Aging of the population, with accumulation of comorbidities.
- Increased spinal instrumentation, including for anesthesia.
- Increased IV drug abuse.
Typically, back pain precedes radicular pain, with spinal cord dysfunction occurring last:
#1) earliest findings
- Local back pain:
- Occurs in most patients.
- Often the initial complaint (in the absence of other distracting processes).
- Pain may be focal or diffuse.(31021957)
- There may be tenderness to palpation.
- Systemic inflammation:
- Fevers, chills, and/or night sweats may occur.
- Fever is absent in most patients, so absence of fever doesn't exclude SEA.(30273242)
#2) radicular symptoms
- Radicular pain may be the most common symptom.
#3) spinal cord dysfunction
- Motor and/or sensory deficits.
- Bowel, bladder, and/or sexual dysfunction.
- (These features may be silent among patients with altered mental status, chronic debilitation, or intubation.)
Traditional teaching was to diagnose SEA on the basis of a triad of pain, fever, and neurological deficits. However, this triad is present only in ~10% of cases.(35063888) Neurologic dysfunction is a late finding, so ideally patients should be diagnosed prior to the development of any neurologic deficit.(28779448)
- ~60% of patients may be bacteremic.(32694325)
- Leukocytosis is present in only 70% of cases, so this is not diagnostically helpful.(34623097)
- Erythrocyte Sedimentation Rate (ESR) and C-Reactive Protein (CRP) might be helpful in sorting out musculoskeletal pain from more serious processes (e.g., malignancy, inflammation, infection). Completely normal inflammatory markers render the diagnosis of SEA unlikely. However, elevated inflammatory markers are extremely nonspecific among critically ill patients.
- CRP level should be obtained as a baseline value, which may be useful to track disease progression over time (e.g., to help determine the duration of antibiotic therapy).(31021957)
- ⚠️ Lumbar puncture is contraindicated in patients with probable/definite spinal epidural abscess, due to the risk of puncturing the abscess and seeding bacteria into the subarachnoid space (thereby causing meningitis).
- MRI with contrast is the test of choice, with sensitivity and specificity >90%.(32694325) Noncontrast MRI has reduced sensitivity for smaller abscesses.
- Abscess is hyperintense on T2, and iso/hypointense on T1 (compared to the spinal cord). Similar to brain abscess, epidural abscesses may display restricted diffusion (i.e., hyperintensity on diffusion-weighted imaging (DWI) with a reduced apparent diffusion coefficient (ADC))(figure below).(33324773)
- Contrast enhancement is often present, which may help determine the maturity of the abscess:
- Enhancement throughout the lesion suggests an early, phlegmonous stage.(33324773) This may not be drainable percutaneously, as it lacks a liquid core.
- A rim-enhancing pattern of contrast enhancement suggests a true abscess.(Wijdicks 2019)
- Adjacent vertebral osteomyelitis (T1-hypointensity) and/or discitis (T2-hyperintensity) are common.
- 💡 Consider MRI of the entire spine if there is a concern for skip lesions, which occur in ~15% of cases.(35063888) Risk factors for the presence of skip lesions include:(32694325)
- Delayed presentation (e.g., symptoms for >7 days).
- Extraspinal infection.
- ESR >95 mm/hr.(35063888)
CT with IV contrast
- May be performed if MRI cannot be obtained.
- CT scan has lower sensitivity for epidural abscess, but it may show reactive bony changes.(34623097)
general principles of SEA microbiology
- As discussed above, most spinal epidural abscesses result from hematogenous seeding of the spinal cord. Therefore:
- (1) Anaerobes are generally not involved (unlike brain abscesses).
- (2) Infection is usually monomicrobial.(35063888)
commonly involved pathogens (25081964)
- Staphylococcal species:
- MSSA (40%)
- MRSA (20%)
- Coagulase-negative staphylococcus (7%)
- Streptococcus spp. (7%)
- Gram-negative bacteria (8%)
- Polymicrobial (5%)
- Mycobacterium tuberculosis.
timing of antibiotic therapy
- Ideally, antibiotic therapy should be held until an organism has been obtained (e.g., from positive blood cultures, surgical cultures, or radiology-guided aspiration). Antibiotics may be delayed in patients who are hemodynamically and neurologically stable.(31021957)
- For patients who are hemodynamically unstable or neurologically compromised, antibiotics should be started as soon as possible (prior to surgery or radiological aspiration, but after obtaining blood cultures).
empiric antibiotic selection
- Empiric therapy can generally focus on gram-positive organisms (e.g., vancomycin or linezolid).(31021957) However, the addition of gram-negative coverage may also be reasonable and this is often recommended (e.g., ceftriaxone or cefepime/piperacillin-tazobactam, among patients with risk factors for pseudomonas such as nosocomial SEA following a procedure). Addition of a beta-lactam to vancomycin also provides added coverage for streptococci or methicillin-sensitive Staphylococcus aureus in case the vancomycin level is subtherapeutic.
- The abscess is located outside of the dura, so antibiotics don't need to penetrate into the cerebrospinal fluid.(31021957) However, it could be reasonable to use meningeal dosing of antibiotic to enhance penetration of abscesses which have not been surgically drained (e.g., ceftriaxone 2 grams IV q12hr).
- Other active infections may require adjustment of the antibiotic regimen, particularly if it is suspected that the epidural abscess is a metastatic complication of another primary infection.
duration of antibiotic therapy
- Most patients receive at least 4-8 weeks of treatment.(30273242)
- Treatment duration is generally determined after ICU departure, in consultation with infectious disease specialists.
When possible, early surgery may often improve neurological outcomes. Similar to malignant spinal cord compression, neurological outcomes relate to the neurological condition prior to decompressive surgery. CT-guided aspiration is occasionally a therapeutic option, but realistically the decision is most often between surgical therapy versus medical therapy (i.e., antibiotics and clinical observation).
- Therapeutic aspiration is technically difficult and not commonly performed. Therapeutic drainage is possible only if all of the following criteria are met:(31021957)
- (1) Posterior abscess location.
- (2) MRI demonstrates a true abscess with a liquid core (e.g., dark signal on T1; only the rim of the abscess enhances with contrast).
- (3) Lack of bony destruction.
- CT-guided aspiration of surrounding structures may be useful to obtain a culture of the infecting organism (e.g., sampling of paraspinal collections, facet joints, or intradiscal spaces).(31021957)
evidentiary basis regarding whether to perform surgery
- No high-level evidence exists (e.g., prospective RCTs). Given the rarity of SEA, such evidence may never exist.
- Available evidence consists of retrospective studies describing cohorts of patients who initially received nonoperative management. Risk factors for failure of medical management were determined (table below).
- This data is substantially limited by selection bias, because it is based on a carefully selected group of patients who were chosen to undergo medical management. Thus, these data may not apply to a broader range of SEA patients.
summary of factors to consider regarding operative vs. nonoperative management
- Neurologic status:
- Intact neurologic status may favor nonoperative management.
- Neurologic deficits lasting <48 hours, or incomplete deficits, are perhaps the strongest indications for intervention (to prevent ongoing or progressive neurological impairment).
- Complete neurological deficit lasting >48-72 hours is controversial. Some authors feel that this supports nonoperative management (given reduced likelihood of neurological recovery). However, other authors state that it's impossible to know at which point the deficit becomes irreversible.(32694325, 31021957)
- Technical issues:
- Panspinal involvement may favor nonoperative management, as this makes surgery more difficult (decompressive laminectomy of the entire spine is a huge surgery).
- Patients with compression fracture or destruction of the spine may require surgery to stabilize the spine.
- This is tricky. Although comorbidity is often used as a rationale to avoid surgery, it's unclear whether this argument is valid.
- Comorbidity increases the risk of surgical complications, but it also increases the risk of failing nonoperative management. For example, patients >65 years old and patients with diabetes are more likely to fail nonoperative management.(32694325)
rough guide to the decision to operate
- Below is one decision aid regarding whether to perform surgery.
- Overall, the vast majority of patients are managed surgically. Surgery should generally be viewed as the “default” therapy, with nonoperative treatment reserved for carefully selected patients.(32684428)
- ⚠️ Please note that the decision regarding whether to undergo surgery should be made by an orthopedic surgeon or neurosurgeon (who should be consulted immediately). Decision rubrics such as the one below are inevitably a simplification of reality.
if nonoperative management is initially pursued:
- Failure of nonoperative management occurs in 30% of patients.(32684428)
- Indications of treatment failure may include:(32694325)
- Neurological deterioration.
- Persistent or worsening symptoms.
- Progression on serial imaging.
- Persistently positive blood cultures.
- Patients should be observed carefully for early indications of treatment failure. These may be an indication for delayed surgical intervention.
Follow us on iTunes
questions & discussion
To keep this page small and fast, questions & discussion about this post can be found on another page here.
- Diagnosis of spinal epidural abscess is incredibly difficult, since most patients lack the classic triad of fever, back pain, and neurologic deficits. There should be a low threshold for MRI with contrast.
- If a spinal epidural abscess is diagnosed, it's probably wise to consult immediately with an infectious disease specialist and a surgeon (either orthopedics or neurosurgery, depending on your hospital). When to start antibiotics and which interventions to perform (surgery or interventional radiology drainage/sampling) may be challenging decisions that require collaboration.
Guide to emoji hyperlinks
- = Link to online calculator.
- = Link to Medscape monograph about a drug.
- = Link to IBCC section about a drug.
- = Link to IBCC section covering that topic.
- = Link to FOAMed site with related information.
- 📄 = Link to open-access journal article.
- = Link to supplemental media.
- 28779448 Alerhand S, Wood S, Long B, Koyfman A. The time-sensitive challenge of diagnosing spinal epidural abscess in the emergency department. Intern Emerg Med. 2017 Dec;12(8):1179-1183. doi: 10.1007/s11739-017-1718-5 [PubMed]
- 30273242 Chow F. Brain and Spinal Epidural Abscess. Continuum (Minneap Minn). 2018 Oct;24(5, Neuroinfectious Disease):1327-1348. doi: 10.1212/CON.0000000000000649 [PubMed]
- 31021957 Babic M, Simpfendorfer CS, Berbari EF. Update on spinal epidural abscess. Curr Opin Infect Dis. 2019 Jun;32(3):265-271. doi: 10.1097/QCO.0000000000000544 [PubMed]
- 32684428 Gardner WT, Rehman H, Frost A. Spinal epidural abscesses – The role for non-operative management: A systematic review. Surgeon. 2021 Aug;19(4):226-237. doi: 10.1016/j.surge.2020.06.011 [PubMed]
- 32694325 Schwab JH, Shah AA. Spinal Epidural Abscess: Diagnosis, Management, and Outcomes. J Am Acad Orthop Surg. 2020 Nov 1;28(21):e929-e938. doi: 10.5435/JAAOS-D-19-00685 [PubMed]
- 33324773 Tetsuka S, Suzuki T, Ogawa T, Hashimoto R, Kato H. Spinal Epidural Abscess: A Review Highlighting Early Diagnosis and Management. JMA J. 2020 Jan 15;3(1):29-40. doi: 10.31662/jmaj.2019-0038 [PubMed]
- 34623097 Bhattacharyya S, Bradshaw MJ. Infections of the Spine and Spinal Cord. Continuum (Minneap Minn). 2021 Aug 1;27(4):887-920. doi: 10.1212/CON.0000000000001031 [PubMed]
- Louis ED, Mayer SA, Noble JM. (2021). Merritt’s Neurology (Fourteenth). LWW.
- 35063888 Long B, Carlson J, Montrief T, Koyfman A. High risk and low prevalence diseases: Spinal epidural abscess. Am J Emerg Med. 2022 Mar;53:168-172. doi: 10.1016/j.ajem.2022.01.008 [PubMed]