Journal Logo


Spinal Epidural Abscess in Adults

Bluman, Eric M. MD, PhD; Palumbo, Mark A. MD; Lucas, Phillip R. MD

Author Information
Journal of the American Academy of Orthopaedic Surgeons: May 2004 - Volume 12 - Issue 3 - p 155-163
  • Free


Spinal epidural abscess is an infection characterized by the accumulation of purulent material in the space between the dura mater and the osseoligamentous confines of the spinal canal. Although referred to by Morgagni in 1761,1 spinal epidural abscess was not clearly defined as a clinical entity until 1820 by Bergamaschi.2 The first successful surgical procedure for this condition was performed by Barth in 1901.3 Reports from the first half of the twentieth century suggested that most cases were fatal.4 Even though subsequent refinement in diagnostic and management methods improved prognosis, the morbidity and mortality associated with spinal epidural abscess remain significant even in more recent series.5-9

Although the definitive management of spinal epidural abscess is generally provided by a spine surgeon, the decisive factor in the outcome of most cases is the rapidity with which the diagnosis is established by the initial contact physician. The clinical features, as well as principles of diagnosis and management, are similar for the adult patient with either a pyogenic or fungal infection involving the extradural space. Infections of the epidural space and epidural abscesses may be associated with spontaneous occurrence from vertebral osteomyelitis or, with increasing incidence, from administration of extradural medications. Thus, it is critical for all physicians to recognize the features of spinal epidural abscess, be able to initiate an appropriate laboratory and radiologic work-up, and have an understanding of basic management principles.


Historically, patients diagnosed with spinal epidural abscess have accounted for 0.2 to 1.2 cases per 10,000 hospital admissions.10 Although still considered rare, spinal epidural abscess has been increasing in incidence over the past decade. Rigamonti et al11 documented approximately 12.5 cases per 10,000 admissions at a large referral center. Reasons proposed for the rising incidence include aging of the general population, a growing number of patients with the predisposing factors of intravenous drug abuse and human immunodeficiency virus (HIV) infection, an increasing incidence of traumatic injuries, more frequent utilization of invasive diagnostic and therapeutic spinal procedures, improved sensitivity of neuror adiologic imaging techniques, and changing hospital referral patterns that concentrate these patients at tertiary care centers.11,12

Spinal epidural abscess has a peak age incidence in the sixth and seventh decades of life.5,13,14 When all large series are taken in aggregate, male predominance is 2:1.15 Predisposing systemic conditions include diabetes mellitus, intravenous drug abuse, end-stage renal disease, alcoholism, HIV infection, malignancy, morbid obesity, long-term corticosteroid use, and septicemia.7,11,15-20 Local conditions that predispose an individual to epidural space infection include recent spine trauma, spinal surgery, and extrathecal injection or catheter placement. The predisposition to spinal epidural abscess imparted by these factors likely relates to impairment of physiologic or structural contributors to host immunocompetence.

Of the microorganisms shown to be causative agents of spinal epidural abscess, Staphylococcus aureus is the most prevalent, accounting for approximately 70% of cases. Streptococcus species are the next most common isolateand are responsible for the infection in approximately 7% of cases. Although less common in general, gram-negative bacilli are frequently isolated from intravenous drug abusers. Mycobacterium tuberculosis, fungal species, and parasitic organisms are rare causes of spinal epidural abscess, especially without associated vertebral osteomyelitis.15,21 In some patients (up to 40% in some series), cultures are sterile, and the infecting organism cannot be identified.22


The three basic mechanisms by which an abscess can form in the epidural space are hematogenous spread, direct extension from an infected contiguous structure, and iatrogenic inoculation. In cases of hematogenous spread, the skin and soft tissues are the most frequent site of origin of the bacteremia and may be responsible in up to 25% of patients.23,24 Other common sources of pathogens include the respiratory tract, the genitourinary system, the abdominal viscera, the oral cavity, and, in endocarditis, the heart valves.13,15 In direct extension (or lymphatic spread) from an infected contiguous structure, the usual source of infection is an adjacent focus of vertebral osteomyelitis. Hlavin et al13 documented the presence of anterior osteomyelitis in two thirds of the patients with spinal epidural abscess. Microorganisms also can originate from the local soft tissues, as in the case of a psoas or retropharyngeal abscess. Iatrogenic inoculation of the epidural space can occur with manipulation of the spine or vascular system during invasive diagnostic or therapeutic techniques. Depending on the series, between 14% and 22% of spinal epidural abscesses occur in patients who have undergone surgery or percutaneous spinal procedures.5,13,23

Most epidural infections caused by hematogenous seeding was previously thought to have been localized to the posterior epidural space8,23 (Fig. 1). However, Soehle and Wallenfang9 recently suggested that posterior abscesses are not the most common. A spinal epidural abscess caused by direct extension from vertebral body osteomyelitis is generally confined to the anterior portion of the spinal canal23 (Fig. 2). Circumferential extension of the abscess around the perimeter of the thecal sac is uncommon. Once established, the craniocaudal spread of a spinal epidural abscess typically is limited to 3 or 4 vertebral segments.5,7,13 Although panspinal epidural abscesses have been described,11,25 such cases are exceedingly rare.

Figure 1
Figure 1:
Posterior spinal epidural abscess. A 41-year-old man presented with back pain, fever, and urinary retention. His erythrocyte sedimentation rate was elevated at 29 mm/hr. A, Sagittal section of a gadolinium-enhanced T1-weighted MRI scan demonstrates a posterior epidural abscess spanning the L1-L3 levels. The cranial and caudal extent of the abscess is delineated by arrows. Note the peripheral enhancement of the lesion. Also evident is osteomyelitis of the T12 vertebral body (asterisk) with erosion of the superior end plate and extension of infection into the T11-12 intervertebral disk. B, Axial image of a gadoliniumenhanced T1-weighted MRI scan at the L3 level shows marked compression of the thecal sac (arrow) by the abscess (asterisk). The patient underwent emergent surgical treatment in the form of L1-L3 decompressive laminectomies along with débridement of the purulent material. Intraoperative cultures grew S aureus. The infection resolved after a 6-week course of intravenous nafcillin. (Courtesy of Glen Tung, MD, Providence, RI, and Henry Bohlman, MD, Cleveland, OH.)
Figure 2
Figure 2:
Sagittal T2-weighted MRI scan of an anterior spinal epidural abscess associated with lumbar vertebral osteomyelitis. (Courtesy of Glen Tung, MD, Providence, RI, and Henry Bohlman, MD, Cleveland, OH.)

The usual limitations on the longitudinal and circumferential spread of infection can be explained, at least in part, by spinal canal anatomy. Specifically, uniform space does not encircle the thecal sac. Using cryomicrotome sectioning of in situ lumbar-spine specimens, Hogan26 showed that the anterior epidural space is potential rather than actual, with the dura mater directly applied to the posterior longitudinal ligament. Posteriorly, the true epidural space is characterized by a repeating metameric arrangement in which the epidural contents are distributed in segmented areas over the dorsal and lateral surface of the dural sac (Fig. 3). Additional longitudinal septations along the sides of the canal divide the epidural space in the axial plane.

Figure 3
Figure 3:
Segmental arrangement of the epidural space. The epidural contents (shaded) are discontinuous circumferentially and longitudinally. (Adapted with permission from Hogan QH: Lumbar epidural anatomy: A new look by cryomicrotome section. Anesthesiology 1991;75:767-775.)

The pathologic basis for the neural impairment associated with spinal epidural abscess is controversial. Several authors have proposed an ischemic mechanism resulting from occlusion of spinal cord arterial inflow, thrombosis of its venous drainage system, or abscess-induced vasculitis.2,27,28 Although little experimental evidence supports a vascular mechanism, it may explain the rapid and irreversible loss of neurologic function observed in some patients. Other investigators advocate direct compression of the spinal cord or cauda equina as the primary pathophysiologic mechanism.10,23 In this scenario, pressure on the neural elements can be caused by the abscess either alone or in combination with retropulsed bone and disk material associated with fracture of an infected vertebral body. Using a rabbit model of spinal epidural abscess, Feldenzer et al1 provided experimental support for the mechanical compression theory. It is likely that the primary etiology of neural damage is mechanical compression, with vascular compromise playing a secondary role.

Clinical Features

Clinical manifestations of spinal epidural abscess are variable. Spinal pain is the most common presenting complaint. Seventy percent to 90% of patients report this symptom; most describe it as relentless and as the worst pain they have experienced. Fever is noted in 60% to 70% of individuals.8,9,15,22,29,30 Localized spinal tenderness, although often found on physical examination, is not universally present. Paraspinal muscle spasm, limitation of spinal motion, and difficulty with ambulation also may be present. Up to one third of patients present with a significant neurologic deficit.15

Neurologic symptoms and signs are frequently encountered in patients with spinal epidural abscess. In a 1999 study from a large referral center, a neural deficit was identified in 53 of 75 patients (71%).11 The severity of neurologic compromise varies, ranging from monoradiculopathy to a complete cord lesion. In 1948, Heusner27 described the typical clinical picture of spinal epidural abscess as progressing through four phases: spinal pain; nerve root pain and spinal tenderness; motor weakness, loss of sensation, and impaired sphincter control; and complete paralysis. In fact, the rate of progression of neurologic compromise is variable. Plegia, for instance, may take days to develop or may become manifest within hours after the onset of neurologic symptoms.31

Increased vigilance is necessary with patients who have undergone catheterization of the epidural space. Signs and symptoms of spinal epidural abscess may appear up to 60 days after the date of catheterization.32,33 Royakkers et al34 reported on two patients with catheter-related epidural abscesses who presented with systemic symptoms and signs of infection in the absence of a neurologic deficit. For this reason, the clinician should be highly suspicious of the individual with a history of recent epidural catheterization or injection who presents with systemic or localized signs of infection, even in the absence of neurologic findings.

Failure to establish the diagnosis of spinal epidural abscess in a timely fashion is common. In a febrile patient with spinal pain, delayed recognition of epidural infection results from failure to consider the diagnosis. Other conditions that can cause back pain with fever include osteomyelitis, subdural abscess, intra-abdominal or retroperitoneal abscess, endocarditis, leukemia, and, in the presence of nuchal rigidity or mental status change, meningitis. When neurologic compromise is present, the differential diagnosis must be expanded to encompass disk herniation, spinal neoplasm, epidural hematoma, transverse myelitis, syringomyelia, and cord infarction.

Laboratory Data

Serologic tests are useful in the diagnosis and management of spinal epidural abscess. Essentially all patients have an elevated erythrocyte sedimentation rate (ESR) on admission;11,13,22,23 the aggregate data of Reihsaus et al15 demonstrated only 7 of 117 patients (6%) with ESR values within the normal range. Rigamonti et al11 reported a mean ESR of 51 mm/hr in their cohort of patients. C-reactive protein (CRP) levels also should be tested in suspected or confirmed cases because levels of this acute phase reactant rise more quickly at the onset of inflammation and return to normal levels more rapidly than the ESR does.35 Although CRP levels have not been extensively studied in relation to spinal epidural abscess, Soehle and Wall enfang9 demonstrated that CRP values followed over the first 2 weeks after surgery can be a prognosticator of eventual outcome. An increased peripheral white blood cell count with a left shift is common and has been reported in 60% to 77% of patients.11,13,22

Blood cultures always should be drawn; they provide early isolation of the causative microbe in approximately 60% of patients.5 An intracutaneous tuberculin skin test performed with a purified protein derivative also should be done to identify previous mycobacterial infection. The search for potential causes of bacteremia includes screening for pathologic conditions of the genitourinary tract by analysis and culture of the urine. Cardiac ultrasound is warranted to rule out endocarditis, especially in intravenous drug abusers. In the immunocompromised patient, a nutritional panel may provide useful information.

In the past, lumbar puncture to obtain cerebrospinal fluid (CSF) for analysis, and epidural aspiration to obtain culture material, were part of the diagnostic work-up for spinal epidural abscess.13,22 However, both procedures carry the risk of seeding the intrathecal compartment with resultant meningitis. Additionally, dural puncture caudal to the level of a block to CSF flow can cause acute deterioration in neurologic status. Therefore, diagnostic procedures that penetrate the spinal canal generally should not be done until after spinal epidural abscess has been ruled out.

Radiologic Findings

Radiographs of the spine are not a good screening tool for the presence of spinal epidural abscess; however, they should be obtained as the initial study in all cases because abnormalities consistent with vertebral osteomyelitis will be identified in a subset of this patient population. Anteroposterior and lateral radiographs also are helpful in assessing stability of the spine when bony involvement is present. Foreign bodies (eg, the remnant of a catheter) also may be identified. In addition, the identification of preexisting conditions such as osteoporosis, spondylosis, vertebral metastases, and spinal deformity may be useful when making subsequent treatment decisions.

Recent advances in spinal imaging techniques have substantially changed the radiologic assessment of epidural infection. Gadolinium-enhanced magnetic resonance imaging (Gd-MRI) is currently the modality of choice for examination of the spine in patients with a suspected spinal epidural abscess. Gd-MRI should be obtained on an emergent basis whenever the diagnosis of spinal epidural abscess is suspected. With this imaging modality, the abscess is accurately localized and defined because of its pattern of contrast-material uptake. Unlike unenhanced MRI scans, the hyperintense signal of the epidural mass on Gd-MRI allows for demarcation of the abscess from the nonenhancing thecal sac and neural elements36 (Fig. 1). Gd-MRI has the added advantages of facilitating definition of associated vertebral osteomyelitis and identifying many of the entities in the differential diagnosis. Although highly sensitive and specific, Gd-MRI has the potential to overestimate the presence and extent of purulent material within the spinal canal.

When MRI is contraindicated (eg, for an elderly patient with a pacemaker), computed tomography (CT) with myelography provides a sensitive means of defining the location and extent of an extradural compressive lesion. Typically, this study will show an area of partial or complete blockage to CSF flow with associated neuralelement compression (Fig. 4). However, the appearance of an epidural abscess on the CT-myelogram is not specific and cannot be distinguished from other lesions that compress the thecal sac. To avoid the risk of iatrogenic meningitis, a high cervical puncture at C1-C2 should be used to access the subarachnoid space.37

Figure 4
Figure 4:
Anteroposterior (A) and lateral (B) thoracic myelograms of a spinal epidural abscess. This 46-year-old diabetic patient presented with back pain, fever, and paraparesis. Myelography showed a complete block to contrast-material flow at the T8 level. After emergent surgical decompression and débridement, an 8-week course of intravenous antibiotics was administered. The patient made a complete neurologic recovery. (Courtesy of Glen Tung, MD, Providence, RI, and Henry Bohlman, MD, Cleveland, OH.)

Radionuclide scans play little, if any, role in the diagnosis of a spinal epidural abscess. Although epidural infection can be identified, falsenegative rates are high with the use of technetium-, gallium-, and indium isotope-labeled compounds. These nuclear medicine scans lack anatomic detail and, more importantly, will result in a delay of diagnosis and institution of treatment.


With few exceptions, antimicrobial therapy should be administered as soon as the diagnosis of spinal epidural abscess has been established and cultures of the blood, urine, and sputum have been obtained. The initial intravenous antibiotic regimen should provide broad-spectrum coverage of the most common causative organisms (ie, Staphylococcus and Streptococcus species). An agent active against gramnegative bacteria should be added for the patient who is immunocompromised, has a history of intravenous drug abuse, or has had recent infection or manipulation of the genitourinary tract. The regimen is tailored to the specific pathogen once final cultures and antibiotic sensitivities are available.

Because of compromise of the host immune response and the potential progression of infection, corticosteroids should not be administered in most cases of spinal epidural abscess. The risk-benefit ratio seems to be especially unfavorable for patients who have no evidence of myelopathy. However, intravenous steroids might be considered for neuroprotection when a patient with a progressive neurologic deficit is being prepared for surgery.

Surgical Management

The fundamental treatment of a spinal epidural abscess is emergent surgical decompression of the neural elements and drainage of the abscess, followed by long-term antibiotic therapy. Although many factors influence the surgical decision-making process, the choice of a specific approach and technique depends primarily on the location of the abscess within the spinal canal and the assessment of spinal stability.

Because most posterior abscesses are not associated with vertebral osteomyelitis, a dorsal approach to the epidural space by laminectomy is effective. In general, spinal stability is maintained after decompression, making arthrodesis unnecessary. Supplemental instrumentation and fusion should be considered if the structural integrity of the posterior elements is compromised by infection, deformity, or the decompressive procedure. With adequate débridement and thorough antibiotic irrigation of the wound, primary closure over drains is recommended.

Surgical management of an anterior epidural abscess is more complex because of the frequent association of spinal osteomyelitis. Structural compromise of the vertebral body, intervertebral disk, and longitudinal ligaments by the infection can lead to spinal instability and deformity. An anterior surgical approach generally is required to adequately decompress the neural elements and drain the epidural purulence. The decompression must include débridement of all infected osseous tissue, followed by structural bone grafting. Autogenous iliac crest or fibula is the optimal source of bone graft for the arthrodesis, but allogeneic material may be a reasonable graft alternative.38,39 When the posterior elements are intact, this type of anterior construct combined with external bracing usually maintains spinal stability during the healing process.

In certain patients, the anterior arthrodesis may need to be combined with spinal instrumentation. Supplemental fixation should be strongly considered when more than one vertebral body has been resected or when the structural integrity of the anterior reconstruction is compromised by osteopenia or deficiency of the posterior stabilizing elements. Anterior instrumentation enhances stability of the reconstructed spine segments, but its use in the presence of vertebral osteomyelitis remains controversial. In many cases, adding posterior instrumentation in a delayed fashion to avoid implant placement in the area of active infection may be more appropriate.40 With respect to the timing of posterior fixation, proceeding immediately after the anterior procedure may be acceptable, but few data support this approach. If the posterior operation is necessarily delayed, stability of the thoracolumbar spine usually can be maintained with a thor acolumbosacral orthosis. In the cervical region, skeletal traction or a halo vest can be used to stabilize the spinal column before definitive posterior fixation.

Management of epidural abscess by percutaneous drainage of the spinal canal has been reported.29,41 Although perioperative morbidity is likely reduced compared with standard open surgery, percutaneous measures provide limited access to the pathologic anatomy and a less thorough means of abscess evacuation and débridement of infected tissue. Thus, this technique cannot currently be advocated as a primary treatment modality. However, percutaneous abscess drainage may be appropriate as an adjunct to antimicrobial therapy. The method may be particularly applicable to a suboptimal surgical candidate in whom the purulent material is easily accessible in the posterior epidural space. Recently, an endoscopic technique of direct abscess visualization and débridement has been described.42 Further study is necessary to document the safety and efficacy of spinal endoscopy as a treatment for epidural abscess.

Antimicrobial therapy is mandatory after surgery. Assuming adequate débridement of the infected tissue, intravenous antibiotics should be continued for a minimum of 4 weeks.5,7,11 A more extended regimen of parenteral antibiotics is required with osteomyelitis, a highly virulent organism, suboptimal débridement, or an immunocompromised host. Although the use of oral agents after the intravenous regimen is controversial, we advocate continuation of oral antibiotics until there is good clinical and laboratory evidence that the infectious process has resolved. Infectious disease consultation should be obtained to optimize antibiotic coverage and to aid in determining possible sources of bacteremia, such as infection of the skin, urinary tract, and cardiac valves.

Nonsurgical Management

Although urgent surgical drainage remains the primary treatment of spinal epidural abscess, antimicrobial therapy without surgery recently has been advocated. However, a true index of the efficacy of nonsurgical management is difficult to derive from the published case reports14,43-45 and retrospective studies.22,25,46 The results of surgery have not been compared with antibiotics alone in a prospective randomized trial. In addition, several reports have documented a not able risk of neurologic deterioration during medical management.5,13,45 The available data do not support the generalized use of nonsurgical therapy for spinal epidural abscess.

Nevertheless, nonsurgical management may be appropriate in certain cases. These include clinical situations in which the patient is a poor surgical candidate because of severe concurrent medical illness, an extensive or panspinal abscess precludes adequate surgical drainage, or complete plegia has been present for more than 48 to 72 hours. In these individuals, every reasonable attempt should be made to identify the pathogen. In most patients, the organism can be isolated from the blood. When blood cultures are sterile and the spinal epidural abscess is associated with vertebral osteomyelitis or psoas abscess, CT-guided biopsy often will identify the microbe. In rare cases, a limited open (or an endoscopic) biopsy of the epidural space may be required.

A more controversial, although not unreasonable, indication for medical management involves the patient with a spinal epidural abscess but no significant neurologic deficit. In this scenario, there is a spectrum of canal involvement. For instance, in cases of vertebral osteomyelitis with associated disk-space infection, the MRI scan may show some degree of longitudinal extension of infected material behind the posterior cortex of the vertebral bodies without significant compression of the neural elements. These MRI findings likely represent an early stage of the epidural infectious process that could develop into a focal abscess with a substantial mass effect on the contents of the dural sac. When considering a nonsurgical approach for the neurologically intact patient, the abscess ideally should be in the lumbosacral region (such that the spinal cord is not at risk). Also, the mass effect on the neural elements should be minimal. Additionally, the spinal column should not be significantly compromised structurally. The success of nonsurgical management in this situation depends on the immune status of the patient, along with early isolation of the causative organism and administration of a culture-specific antibiotic regimen.

Because of the potential for rapid and irreversible neurologic deterioration during nonsurgical treatment, vigilant surveillance of the patient is necessary. Frequent clinical examination, interval determination of the ESR and CRP level, and serial Gd-MRI help to monitor the response to medical therapy. A prolonged period of hospitalization may be needed in some cases. After the patient is discharged, the recurrence of symptoms consistent with spinal epidural abscess should prompt a rapid reassessment.

The required duration of medical management is variable. In the absence of vertebral osteomyelitis, intravenous antibiotics generally are administered for 4 to 12 weeks.7,23,25,44 A longer course of parenteral antimicrobial therapy may be necessary in the presence of concurrent spinal column infection, a compromised host immune system, recurrent bacteremia, continued elevation of the ESR, or persistent enhancement of the abscess on serial MRI scans. Many patients also will need to be maintained on oral antibiotics for an extended period. In general, pharmacologic treatment can be discontinued when the reversible clinical manifestations of epidural infection (ie, fever and spinal pain) have subsided, the ESR and CRP level have normalized, and the Gd-MRI scan shows abscess resolution.


There has been a substantial reduction over the past 50 years in mortality resulting from spinal epidural abscess.15 Reihsaus et al15 performed an extensive review and analysis of the literature pertaining to spinal epidural abscess. This revealed that in the last decade of the 20th century, approximately 15% of those diagnosed with this condition succumbed to it. Two subsequent series29,30 published in 2002 suggest a trend of reduction, with reported mortality rates ≤10%. Despite this apparent reduction in mortality associated with spinal epidural abscess, permanent neurologic sequelae remain common.

Prognosis is affected by the patient's medical condition at presentation. The combination of an epidural abscess and overt sepsis is usually lethal.8,24 Host immunocompetence has an important influence on outcome. A suboptimal response to treatment has been described in patients taking oral corticosteroids and in those with HIV infection.5,12 Poor results have been correlated with the following risk factors: multiple medical comorbidities, infection with methicillinresistant staphylococci, and previous spinal surgery.11

The neurologic outcome in a patient with epidural infection depends in part on the location of the abscess. In one recent study, 36% of patients with a thoracic abscess had a poor outcome compared with 12.5% with cervical and 7.7% with lumbar spinal epidural abscesses.11 These findings are similar to those reported by Khanna et al.8 The results of the study by Tang et al,30 although not statistically significant, demonstrated a better outcome for noncervical epidural abscesses.

Additionally, neurologic outcome has been shown to be inversely related to the severity of the neural deficit at presentation. Rigamonti et al11 showed that patients presenting with significant neurologic compromise were more likely to have a poor outcome. Of 29 patients with severe neural dysfunction, 6 (21%) died, and 10 (34%) did not regain the ability to ambulate. In 24 patients with a moderate deficit, 2 (8%) died, and 7 (29%) were unable to walk. All 22 patients with a normal presenting neurologic examination had resolution of the infection and normal ambulatory function. Lu et al29 concluded that the only statistically significant (P < 0.001) prognostic factor was the presence of a pretreatment motor deficit.

Any substantial delay in diagnosis or management can have a detrimental effect on the overall clinical outcome. Whereas Heusner27 in 1948 only warned of the negative impact of a surgical delay on neurologic recovery, Danner and Hartman5 in 1987 demonstrated that partial neurologic improvement occurred in patients who had surgery with deficits of ≤36 hours' duration, whereas no recovery was observed in 9 of 11 patients with an established deficit of >36 hours. Maslen et al22 documented no neurologic recovery in patients who were plegic for >12 hours before surgery. Khanna et al8 concluded that recovery is more likely when decompression is done ≤72 hours of the onset of a neural deficit. Rigamonti et al11 subsequently reported that patients operated on >36 hours after the onset of a severe motor deficit rarely recovered useful muscle function.

The outcome of nonsurgical management is difficult to ascertain from the limited available data. In a 1992 literature review, Wheeler et al14 analyzed the outcome after medical management of 37 patients with spinal epidural abscess. This group was culled from three individual case reports and 18 case series published between 1970 and 1990. Twenty-three (62%) of the patients recovered completely, 2 died, 1 worsened, and the remaining 11 either showed no improvement or recovered partially. Patients with localized spinal or radicular pain and those with an incomplete neurologic deficit were more likely to show recovery than were those with complete cord lesions or in septic shock.


Spinal epidural abscess is a rare but potentially devastating disorder. In most cases, the major determinants of outcome are the rapidity with which the diagnosis is made and treatment initiated. The responsibility for recognizing the infection often falls on the physician who first evaluates the patient. A high index of suspicion is warranted in the patient who presents with spinal pain or a neurologic deficit in conjunction with fever or an elevated ESR, in the immunocompromised patient with back pain, and in the patient who has undergone recent epidural catheter placement and presents with systemic signs of infection, localized spinal pain, or neurologic compromise. Once the diagnosis of spinal epidural abscess is considered, blood cultures and plain radiographs of the spine should be added to the screening serologic tests.

In the presence of neurologic dysfunction, Gd-MRI of the entire spinal column should be obtained immediately. If a spinal epidural abscess is identified, emergent surgical treatment is recommended. Intraoperative cultures serve to guide the mandatory long-term antibiotic regimen. If a substantial delay before surgery cannot be avoided, broad-spectrum antibiotics should be started before the operation. Indications for nonsurgical therapy are few and may include a medical profile that prohibits surgical intervention, an extensive or panspinal abscess not amenable to surgery, and the presence of complete plegia of >36 to 48 hours' duration.

Emergent screening Gd-MRI of the spine also should be done in the absence of neurologic dysfunction. When an epidural abscess is confirmed in the neurologically intact patient, surgical decompression and débridement (with or without spinal reconstruction) is the treatment of choice in the presence of significant neural element compression. Antimicrobial agents are not administered until after intraoperative cultures are obtained. Nonsurgical therapy of spinal epidural abscess is warranted when surgery is contraindicated because of comorbid disease. Nonsurgical management also may be considered in cases in which the risk of neurologic demise is judged to be low based on the location and morphology of the abscess, the immune status of the patient, and the virulence of the organism.

Acknowledgment: We would like to thank Drs. Glen Tung and Henry Bohlman for providing the radiologic images used in this article.


1. Feldenzer JA, McKeever PE, Schaberg DR, Campbell JA, Hoff JT: Experimental spinal epidural abscess: A pathophysiological model in the rabbit. Neurosurgery 1987;20:859-867.
2. Browder J, Meyers R: Infection of the spinal epidural space. Am J Surg 1937;37:4-26.
3. Markham JW: Surgery of the spinal cord and vertebral column, in Walker AE: A History of Neurologic Surgery. New York, NY: Hafner, 1967, pp 364-395.
4. Mixter WJ, Smithwick RH: Acute intraspinal epidural abscess. N Engl J Med 1932;207:126-131.
5. Danner RL, Hartman BJ: Update of spinal epidural abscess: 35 cases and review of the literature. Rev Infect Dis 1987;9:265-274.
6. Ericsson M, Algers G, Schliamser SE: Spinal epidural abscesses in adults: Review and report of iatrogenic cases. Scand J Infect Dis 1990;22:249-257.
7. Del Curling OJr, Gower DJ, McWhorter JM: Changing concepts in spinal epidural abscess: A report of 29 cases. Neurosurgery 1990;27:185-192.
8. Khanna RK, Malik GM, Rock JP, Rosenblum ML: Spinal epidural abscess: Evaluation of factors influencing outcome. Neurosurgery 1996;39:958-964.
9. Soehle M, Wallenfang T: Spinal epidural abscesses: Clinical manifestations, prognostic factors, and outcomes. Neurosurgery 2002;51:79-87.
10. Baker AS, Ojemann RG, Swartz MN, Richardson EP Jr: Spinal epidural abscess. N Engl J Med 1975;293:463-468.
11. Rigamonti D, Liem L, Sampath P, et al: Spinal epidural abscess: Contemporary trends in etiology, evaluation, and management. Surg Neurol 1999;52: 189-197.
12. Koppel BS, Tuchman AJ, Mangiardi JR, Daras M, Weitzner I: Epidural spinal infection in intravenous drug abusers. Arch Neurol 1988;45:1331-1337.
13. Hlavin ML, Kaminski HJ, Ross JS, Ganz E: Spinal epidural abscess: A ten-year perspective. Neurosurgery 1990;27:177-184.
14. Wheeler D, Keiser P, Rigamonti D, Keay S: Medical management of spinal epidural abscesses: Case report and review. Clin Infect Dis 1992;15:22-27.
15. Reihsaus E, Waldbaur H, Seeling W: Spinal epidural abscess: A metaanalysis of 915 patients. Neurosurg Rev 2000;23:175-205.
16. Boszczyk BM, Krause P, Bolay H, Hohmann F, Mayer HM: Spinal epidural abscess following blunt pelvic trauma. Eur Spine J 2000;9:80-84.
17. Sillevis Smitt P, Tsafka A, van den Bent M, et al: Spinal epidural abscess complicating chronic epidural analgesia in 11 cancer patients: Clinical findings and magnetic resonance imaging. J Neurol 1999;246:815-820.
18. Still JM, Abramson R, Law EJ: Development of an epidural abscess following staphylococcal septicemia in an acutely burned patient: Case report. J Trauma 1995;38:958-959.
19. Korovessis P, Sidiropoulos P, Piperos G, Karagiannis A: Spinal epidural abscess com plicated closed vertebral fracture: A case report and review of literature. Spine 1993;18:671-674.
20. Byers K, Axelrod P, Michael S, Rosen S: Infections complicating tunneled intraspinal catheter systems used to treat chronic pain. Clin Infect Dis 1995;21:403-408.
21. Khwaja MS, Dossetor JF, Lawrie JH: Extradural guinea-worm abscess: Report of two cases. J Neurosurg 1975;43:627-630.
22. Maslen DR, Jones SR, Crislip MA, Bracis R, Dworkin RJ, Flemming JE: Spinal epidural abscess: Optimizing patient care. Arch Intern Med 1993;153:1713-1721.
23. Martin RJ, Yuan HA: Neurosurgical care of spinal epidural, subdural, and intramedullary abscesses and arachnoiditis. Orthop Clin North Am 1996;27:125-136.
24. Redekop GJ, Del Maestro RF: Diagnosis and management of spinal epidural abscess. Can J Neurol Sci 1992;19:180-187.
25. Leys D, Lesoin F, Viaud C, et al: Decreased morbidity from acute bacterial spinal epidural abscesses using computed tomography and nonsurgical treatment in selected patients. Ann Neurol 1985;17:350-355.
26. Hogan QH: Lumbar epidural anatomy: A new look by cryomicrotome section. Anesthesiology 1991;75:767-775.
27. Heusner AP: Nontuberculous spinal epidural infections. N Engl J Med 1948;239:845-854.
28. Lasker BR, Harter DH: Cervical epidural abscess. Neurology 1987;37:1747-1753.
29. Lu CH, Chang WN, Lui CC, Lee PY, Chang HW: Adult spinal epidural abscess: Clinical features and prognostic factors. Clin Neurol Neurosurg 2002;104:306-310.
30. Tang HJ, Lin HJ, Liu YC, Li CM: Spinal epidural abscess: Experience with 46 patients and evaluation of prognostic factors. J Infect 2002;45:76-81.
31. Phillips GE, Jefferson A: Acute spinal epidural abscess: Observations from fourteen cases. Postgrad Med J 1979;55:712-715.
32. Nordberg G, Mark H: Epidural abscess after epidural analgesia treated successfully with antibiotics. Acta Anaesthesiol Scand 1998;42:727-731.
33. Kindler CH, Seeberger MD, Staender SE: Epidural abscess complicating epidural anesthesia and analgesia: An analysis of the literature. Acta Anaesthesiol Scand 1998;42:614-620.
34. Royakkers AA, Willigers H, van der Ven AJ, Wilmink J, Durieux M, van Kleef M: Catheter-related epidural abscesses: Don't wait for neurological deficits. Acta Anaesthesiol Scand 2002;46:611-615.
35. Kushner I: C-reactive protein and the acute-phase response. Hosp Pract (Off Ed) 1990;25:13, 16, 21-28.
36. Post MJ, Sze G, Quencer RM, Eismont FJ, Green BA, Gahbauer H: Gadoliniumenhanced M R in spinal infection. J Comput Assist Tomogr 1990;14:721-729.
37. Smith AS, Blaser SI: MR of infectious and inflammatory diseases of the spine. Crit Rev Diagn Imaging 1991;32:165-189.
38. Eismont FJ, Bohlman HH, Soni PL, Goldberg VM, Freehafer AA: Pyogenic and fungal vertebral osteomyelitis with paralysis. J Bone Joint Surg Am 1983;65:19-29.
39. Emery SE, Chan DP, Woodward HR: Treatment of hematogenous pyogenic vertebral osteomyelitis with anterior debridement and primary bone grafting. Spine 1989;14:284-291.
40. Yilmaz C, Selek HY, Gurkan I, Erdemli B, Korkusuz Z: Anterior instrumentation for the treatment of spinal tuberculosis. J Bone Joint Surg Am 1999;81:1261-1267.
41. Hori K, Kano T, Fukushige T, Sano T: Successful treatment of epidural abscess with a percutaneously introduced 4-French catheter for drainage. Anesth Analg 1997;84:1384-1386.
42. Roselli R, Iacoangeli M, Pompucci A, et al: Anterior cervical epidural abscess treated by endoscopy-assisted minimally invasive microsurgery via posterior approach. Minim Invasive Neurosurg 1998;41:161-165.
43. Kwon NS, Heffer SM, Levy PD: Spinal epidural abscess potentially misdiagnosed as multiple myeloma. Hosp Physician 2001;37:54-59.
44. Messer HD, Lenchner GS, Brust JCM, Resor S: Lumbar spinal abscess managed conservatively: Case report. J Neurosurg 1977;46:825-829.
45. Harrington P, Millner PA, Veale D: Inappropriate medical management of spinal epidural abscess. Ann Rheum Dis 2001;60:218-222.
46. Mampalam TJ, Rosegay H, Andrews BT, Rosenblum ML, Pitts LH: Nonoperative treatment of spinal epidural infections. J Neurosurg 1989;71:208-210.
© 2004 by American Academy of Orthopaedic Surgeons