Spinal Epidural Abscess

The Time-Sensitive Conundrum of Diagnosing Spinal Epidural Abscess in the ED

Stephen Alerhand, MD
Resident Physician
Mount Sinai Emergency Medicine


  • Spinal epidural abscess is an extremely rare diagnosis but a potentially devastating one.
  • It has proven to be a difficult ER diagnosis.
  • Unfortunately, once the classic symptoms appear and a definitive diagnosis is made, the very symptoms we look for have often become irreversible.
  • How can we change the way we evaluate for SEA in order to minimize diagnostic delay and thereby decrease morbidity/mortality?


  • Previously reported to occur in 0.2-1.2 patients per 10,000 hospital admissions in 1975 by Baker et al.1 Has risen to 2.5-3 patients per 10,000 admissions due to an increase in predisposing conditions and spinal instrumentation.2
  • Collection of pus or inflammatory granulation between the dura and vertebral column, usually in the wider thoracic/lumbosacral regions containing more infection-prone fat tissue. Extends an average of 3-5 spinal cord segments.3


  • Entry into epidural space: via contiguous infected tissue (vertebral body, psoas muscle), hematogenous spread (skin, soft-tissue, urinary, and respiratory tract infections), or iatrogenic inoculation (epidural analgesia, paraspinal steroid injection, lumbar puncture, surgery, nerve block). No source identified in 30-40% of cases. 3, 4
  • Damage to spinal cord caused by: direct compression, thrombosis, and thrombophlebitis of nearby veins, interruption of arterial blood supply, bacterial toxins, inflammatory mediators.

Risk factors

  • Spinal instrumentation, contiguous bony or soft tissue infection, bacteremia secondary to distant infection, diabetes mellitus, HIV, trauma, IV drug abuse, immunosuppressive therapy, cancer, renal failure, alcoholism, tattooing. 5, 6

Classic Triad

  • Spinal pain, fever, neurologic deficits (see discussion below).


  • Imaging: Definitive diagnosis made by MRI, with sensitivity and specificity above 90%. 7, 8 Worse outcomes with: central stenosis >50% and abscess length >3 cm. 9


A: arrows delineate abscess margins; B: asterisk – abscess, arrow – thecal sac

  • Tests: Positive cultures from abscess in 90% and blood cultures 62% 10, and CSF 19% while almost always negative Gram-stain.3 Staph aureus by far most frequent pathogen (60%).11 Lumbar puncture not indicated because as the needle passes through the abscess, pathogens may be iatrogenically pushed into the meninges or subdural space.
  • Labs: ESR >20 mm/h is found in up to 95%3, 12 and in all but one of 63 SEA patients in a study by Davis et al.13 Neither ESR, CRP, nor WBC is specific for SEA. Low platelets (<100) and extremely high ESR (>110) predicted a poor outcome in one study of 46 SEA patients. 14


  • Goals: reduce abscess size, eliminate abscess, eradicate causative organism.
  • Surgery: Most cases call for surgical decompression by laminectomy, hemilaminectomy, or interlaminar fenestration. Neurological improvement unlikely if paresis exceeds 24-36 hours. 15
  • CT-guided needle aspiration combined with antibiotics: for patients with posterior SEA, no neurological deficit, high surgical risk, and who do not respond to antibiotics alone.
  • Antibiotics alone: if high surgical risk or when neurological deficits are present >3 days and unlikely to improve.
    • Antibiotics against Staph, Strep, and Gram-negative bacilli. Example: Vancomycin plus Ceftriaxone/Cefotaxime/Ceftazidime plus Metronidazole.
    • Against Pseudomonas if IV drug use, Coag-negative Staph if implanted device such as epidural catheters.
    • Duration between 4-16 weeks depending on: co-morbidities, isolated microorganism, bactericidal effect of agent.16 Frequent follow-up of neurological status and MRI studies (after 2-4 weeks) required. Efficacy of non-surgical therapy usually apparent within first 48-72 hours.17

How well do ER physicians diagnose SEA?

  • In a study of 63 SEA patients, Davis et al. found that diagnostic delays (multiple ED visits before diagnosis, admission without SEA diagnosis, >24 hrs to a definitive study) were present in 75% of SEA patients.13

Why do we miss it?

  • SEA is an extremely rare diagnosis.
  • Low back pain is a very common reason for physician visits. It is all-too-easy to attribute to degenerative joint etiology. Symptoms may even be masked by NSAID or steroid use.
  • Symptoms to look for are not necessarily present.
    • Classic triad present in only 10-15% of cases at first physician contact in Davis et al.’s study. 13
      • Fever in only 32% at first ED visit, neurological exam documented as normal in 68%.
    • In a review of 75 cases, Rigamonti et al. found the classic triad present in 37%, while 22% had no neurologic deficit with or without back pain.16
    • Over a range of studies, back pain and severe local tenderness were the most frequent early findings at 75% and 58%, respectively.13
  • The earliest symptoms are non-specific.
    • Stage I: back pain (focal, severe), fever, tenderness.
    • Stage II: radicular pain (“shooting”, “electric shocks”), nuchal rigidity/neck stiffness, reflex changes.
  • Not until the abscess has lingered or progressed do the obvious symptoms present.
    • Stage III: sensory abnormalities, motor weakness, bowel and bladder dysfunction.
    • Stage IV: paralysis.
  • Time varies from symptom onset to the patient seeking medical attention. Patients thus present at variable stages.
    • For Davis et al., mean duration was 5 and 9 days, respectively, with median number of ED visits before admission being 2. 13
  • Neurological deficits can be wide-ranging. Motor: focal motor signs along dermatome, complete paraplegia. Sensory: paresthesias, hyperesthesias, pressure sensation.
  • Laboratory markers have limited specificity.
  • MRI is not always readily available to emergency departments.
  • Bed-bound patient sick for other reasons may have “hidden” neurological symptoms.

What are the consequences of missing this diagnosis?

  • Increased risk of residual weakness with diagnostic delay vs no delay (45% vs 13%).13
  • Symptoms often irreversible by the time they present in Stages III, IV. Almost 50% of survivors with residual neurologic deficits, including 15% with paresis or complete paralysis.16, 18
  • Final outcome (after up to 1 year) strongly correlates with severity (Stage III, IV) and duration (>24-36 hours) of neurological deficits before surgery.
  • Mortality rates (2-20%) usually due to severe sepsis and in patients with multiple co-morbidities. Worse outcomes associated with leukocytosis (>14), thrombocytopenia (<100), MRSA, prior spinal surgery, corticosteroid treatment, and HIV infection.13


  • Early diagnosis is the major prognostic factor for favorable outcome.
  • Put SEA on the differential for spinal pain and with recent spinal instrumentation.
  • Use risk factor assessment rather than “classic triad” exam screening.
  • Perform full neurological exam including: reflexes, sensory and motor function, anal sphincter tone, ability to completely void—especially if patient is bed-bound or with other comorbidities. Focus on specific extremities/dermatomes corresponding to affected spine level.
  • Consider ESR as screening tool in patients with spinal pain and a risk factor but without neurological deficits.
  • Consider urgent MRI for high-risk patients with: neurological deficit and focal back pain, deficit and unexplained fever, deficit and elevated ESR, severe focal back pain and fever, severe focal back pain with markedly elevated ESR, unexplained extremely severe focal back pain. 16
  • Advocate for your patient—early surgical intervention unless otherwise indicated.

References / Further Reading

  1. Baker AS, Ojemann RG, Swartz MN, Richardson EP Jr. Spinal epidural abscess. N Engl J Med 1975; 293:463-8.
  2. Sampath P, Rigamonti D. Spinal epidural abscess: a review of epidemiology, diagnosis, and treatment. J Spinal Disord 1999; 12:89-93.
  3. Darouiche RO, Hamill RJ, Greenberg SB, et al. Bacterial spinal epidural abscess. Review of 43 cases and literature survey. Medicine (Baltimore) 1992; 71:369-85.
  4. Danner RL, Hartman BJ. Update on spinal epidural abscess: 35 cases and review of the literature. Rev Infect Dise 1987; 9:265-74.
  5. Reynolds F. Neurological infections after neuraxial anesthesia. Anesthesiol Clin 2008; 26:23.
  6. Reihsaus E, Waldbaur H, Seeling W. Spinal epidural abscess: a meta-analysis of 915 patients. Neurosurg Rev 2000; 23:175-204; discussion 5.
  7. Angtuaco EJ, McConnell JR, Chadduck WM, Flanigan S. MR imaging of spinal epidural sepsis. Am J Roentgenoli 1987; 149:1249-53.
  8. Wong D, Raymond NJ. Spinal epidural abscess. N Z Med J 1998; 111:345-7.
  9. Tung GA, Yim JW, Mermel LA, Philip L, Rogg JM. Spinal epidural abscess: correlation between MRI findings and outcome. Neuroradiology 1999; 41:904-9.
  10. Gellen BG, Weingarten K, Gamache FW Jr, et al. Epidural Abscess. In: Infections of the Central Nervous System, 2nd Ed, Scheld WM, Whitley RJ, Durack DT (Eds), Lippincott-Raven Publishers, Philadelphia 1997. P 507.
  11. Curry WT Jr, Hoh BL, Amin-Hanjani S., Eskandar EN. Spinal epidural abscess: clinical presentation, management, and outcome. Surg Neurol 2005; 53:364-71; discussion 71.
  12. Soehle M, Wallenfang T. Spinal epidural abscesses: clinical manifestations, prognostic factors, and outcomes. Neurosurgery 2002; 51:79-85; discussion 6-7.
  13. Davis DP, Wold RM, Patel RJ, et al. The clinical presentation and impact of diagnostic delays on emergency department patients with spinal epidural abscess. J Emerg Med 2004; 26:285-91.
  14. Tang HJ, Lin HJ, Liu YC, Li CM. Spinal epidural abscess—Experience with 46 patients and evaluation of prognostic factors. Journal of Infection 2002; 45:76-81.
  15. Bluman EM, Palumbo MA, Lucas PR. Spinal epidural abscess in adults. J Am Acad Orthop Surg 2004; 12:155-163.
  16. Rigamonti D, Liem L, Sampath P, et al. Spinal epidural abscess: contemporary trends in etiology, evaluation, and management. Surg Neurol 1999; 52:189-96; discussion 97.
  17. Hlavin ML, Kaminski HJ, Ross JS, Ganz E. Spinal epidural abscess: a ten-year perspective. Neurosurgery 1990; 27:177-84.
  18. Reihsaus E, Waldbaur H, Seeling W. Spinal epidural abscess: a meta-analysis of 915 patients. Neurosurg Rev 2000;23: 175-204.
  19. http://www.ncbi.nlm.nih.gov/pubmed/21417700
  20. http://www.ncbi.nlm.nih.gov/pubmed/21308559
Edited by Alex Koyfman

The post Spinal Epidural Abscess appeared first on emDocs.

38 Year Old Male – Chest Pain and Leg Paralysis.

You are called for severe chest pain.

The patient is a 38 year old male who describes the abrupt onset of a severe pain in his chest about 30 minutes before his wife called EMS. While sweat streams off his face, he tells you that he has never felt pain this intense. He isn’t sure if it’s pleuritic, and he endorses some shortness of breath. The pain radiates to his shoulders, back, and epigastrium. Despite the severity of the pain, he is actually far more worried that his left lower extremity is numb, and that he can’t move it – he repeatedly tells you in a loud voice that “Something’s wrong with my leg! What’s wrong with my leg?”

With the assistance of his wife, you find that he takes HCTZ and lisinopril for HTN, but he doesn’t smoke or use recreational drugs. In fact, he’s a coach for a high school cross-country running team, and looks like he’s in pretty good shape.

Vitals signs are

  • HR: 50
  • RR: 30
  • BP: 230/140
  • SaO2: 99%

Besides profound diaphoresis, the exam is unrevealing. An ECG is obtained:

screenshot696You give him aspirin 325 mg, and 3 sprays of nitroglycerin, with neither a change in his symptoms nor in his vitals. In fact, 10 mg of morphine IV (max per your protocol) doesn’t improve the discomfort, and he is still yelling about both the chest pain and his leg. Your partner mutters to you “I’m starting to think this is mostly anxiety…”

It is almost 2300 hours. You have three choices of destination hospital:

  1. A “stand-alone” ED that is capable of delivering tPA for STEMI within 30 minutes. It’s just around the corner.
  2. A small community hospital ED that just started performing primary PCI, but they’ll have to call a team in from home. Despite that delay, they will activate based on a prehospital report, and their door-to-balloon times have been excellent. They are 20 minutes away.
  3. A level 1 academic hospital that is 35 minutes away. They don’t activate the cath lab based on EMS interpretation, since they usually “want to see it for ourselves.”


Updated IDSA Soft Tissue Infection Guidelines

Our infectious disease experts have bestowed upon us an update in the management of skin and soft tissue infection.  This is particularly relevant in our new Age of MRSA, where over-reaction and antibiotic overuse has become the norm – and almost certain to usher in a new, more dire, era of resistant pathogens.

But, this update provides a lovely pathway describing the treatment options for SSTIs that, happily, includes many narrow spectrum antibiotics.  Non-abscess SSTIs may be managed with simple penicillins or first-generation beta-lactams.  Purulent pathology is managed simply by incision and drainage – and antibiotics added only in the setting of moderate or severe disease.  They also, importantly, note most impetigo should be treated solely with topical antibiotic ointment.

There are a handful of odd statements, however.  Cultures are still recommended, but the authors acknowledge treatment is reasonable without.  Considering avoidance of cultures in uncomplicated SSTI is part of ACEP’s Choosing Wisely Recommendations, I’d prefer to see revised phrasing from the authors in the associated passages.  Recommendations for antibiotic prophylaxis after dog and cat bites appropriately require specific risk-factors for superinfection, but may yet still be overly broad.  At least, however, the level of evidence supporting their recommendation is appropriately cited as “low”.

And, of course, it’s always nice to be reminded of the appropriate treatment of glanders.

Finally, the recommendations do not overtly appear to reflect the influence of financial conflicts, despite many authors declaring substantial COI.

“Practice Guidelines for the Diagnosis and Management of Skin and Soft Tissue Infections: 2014 Update by the Infectious Diseases Society of America”

Two EKG patterns of pulmonary embolism which mimic MI

Introduction with a case
A 45-year old man presented to the hospital with chest pain and dyspnea.   His troponin was positive, and EKG showed T-wave inversions in the inferior leads and V1-V4.   He was pale, diaphoretic, tachycardic, and borderline hypotensive with a systolic blood pressure ranging from 85-110mm.   He was taken urgently for cardiac catheterization, which demonstrated no significant coronary artery disease.   When he was taken off bedrest, he got up to use the commode and had a PEA arrest.   Autopsy revealed pulmonary embolism.  

This post describes two EKG patterns of PE which mimic MI.   Patients presenting with chest pain, these EKG patterns, and troponin elevation are often misdiagnosed with MI.   In one multi-center study, 3% of all PE patients were admitted with an incorrect diagnosis of MI (Kukla 2011).   These EKG patterns are associated with submassive or massive PE, so immediate recognition and appropriate therapy is essential.

Pattern #1: RV strain pattern

­           PE causing RV strain (Panduranga 2013).

RV strain pattern
  • T-wave inversion in V1 and V2
  • At least one of the following:
    • T-wave inversion in lead III
    • The precordial lead with deepest T-wave inversion is V1 or V2

The most common MI-mimic is caused by RV strain with T-wave inversion (TWI) in leads V1-V4.   TWI in these leads typically raises concern regarding the possibility of ischemia involving the left anterior descending coronary artery (i.e., Wellen's pattern).   However, this is also reported in up to 40% of patients with PE (Kukula 2014)(1).

Differentiation from MI

Kosuge 2014 compared 107 PE patients and 248 MI patients who presented with TWI in at least two leads between V1-V4.   The distribution of TWI is different among these two groups: PE patients are more likely to have TWI in the inferior leads and V1-V2, whereas MI patients are more likely to have TWI in V5-V6 (figure below).   Among these patients, the combination of TWI in V1 and lead III was 87% sensitive and 96% specific for PE.   A similar finding was reported by Witting 2012, who found the combination of TWI in V1 and lead III to be 11% sensitive and 95% specific among all patients presenting with PE.

Kosuge 2014 also noticed that the precordial lead with the maximal amount of TWI differed between patients with PE versus MI (figure below).   By combining the criteria of (maximal TWI in V1-V2) and/or (TWI in V1 and III), this increased the sensitivity for PE to 98% while maintaining a specificity of 92% compared to MI.  

For additional detail and examples, please see several great videos by Amal Mattu discussing this EKG pattern here, here, and here.

Differentiation from chronic RV strain

RV strain pattern may be seen in chronic pulmonary hypertension (for example due to severe COPD or obesity hypoventilation syndrome).   Clinical history as well as prior EKG, echocardiographic, or CT angiography data can help sort out whether this is acute or chronic.

Pattern #2: RV injury pattern

Less commonly, PE presents with widespread ST elevations and depressions.   For example, the EKG below is from a PE patient with RV dilation, repeated syncope, and elevated troponin at Genius General Hospital.

This pattern has been discussed occasionally in the literature, and was recently re-introduced in a case series by Zhong-qun 2013.   These authors proposed that ST elevation in aVR with ST depression in leads I and V4-V6 is associated with massive PE.  

RV injury pattern (2)
  • ST elevation in aVR and ST depression in lead I (3)
  • ST elevation in V1-V3 and/or ST depression in V4-V6 
Comparing EKGs of patients with PE who presented with widespread ST changes reveals the RV injury pattern.   In the limb leads, the current of injury consistently projects towards the right leading to ST elevation in aVR and ST depression in lead I (see below), as described by Zhong-qun.   In the precordial plane, some patients have a current of injury which projects anteriorly causing ST elevation in V1-V3 (red arrow below), whereas other patients have a current of injury which points more laterally causing isoelectric ST segments in V1-V3 (orange arrow below).   Depending on the direction of these currents, there may or may not be ST depression in V4-V6.   All patients have some ST deviation present in the precordial leads (ST elevation in V1-V3 and/or ST depression in V4-V6).

Direction of the current of injury varies between the red arrow and the orange arrow (4).   

STE in aVR is common in PE but generally ignored.   Among a series of 500 patients with PE, Kukula 2014 found STE in aVR in 36% of patients, and in 65% of patients with cardiogenic shock from PE.   STE in aVR was the most sensitive EKG finding among patients with cardiogenic shock due to PE.   Janata 2012 found STE in aVR among 34% of patients with PE, and similarly found this more often among patients with severe PE.   

Differentiation from MI

RV injury pattern is easily mistaken for MI.   Some PE patients have a combination of Q-waves and STE in the anterioseptal and/or inferior leads, closely simulating ST-elevation MI.   It is increasingly recognized that ST elevation in aVR may be seen in MI as a marker of severe coronary artery disease, so this feature may also lead clinicians towards a diagnosis of MI.

One feature which may be helpful to distinguish PE from MI is rightward deviation of the terminal component of the QRS vector.   In patients with RV injury pattern, RV conduction is delayed and the terminal portion of the QRS complex reflects the right ventricle contracting.   Thus the final portion of the QRS complex usually has an axis between +90 degrees and -150 degrees, which produces a terminal S-wave in lead I and a terminal R-wave in lead III.   This is not usually seen in myocardial infarction unless there is underlying complete right bundle branch block.   It's easy to remember this pattern because the terminal forces in leads I and III point towards each other:

Another feature which may help distinguish PE from MI is TWI if present.   PE patients may simultaneously have both RV injury and RV strain patterns.   As discussed above, the distribution of TWI may help differentiate PE from MI.  

Clinical Correlation Required

Relationship to PE severity

Both RV strain and RV injury patterns tend to occur in more severe PE.   Echocardiographic RV dysfunction occurs in 75-80% of patients with either pattern (Kosuge 2014, Janata 2012).   There are many case reports of PE patients who are stable at admission, but then suddenly crash with new-onset hypotension and a repeat EKG revealing an RV injury pattern.  

Synergy of EKG with echocardiography

These EKG patterns are usually associated with echocardiographic features of pulmonary embolism including right ventricular dilation.   MI is also often associated with abnormalities visible on echocardiography.   Therefore for patients who present with these EKG patterns, bedside echocardiography is often helpful in differentiating between PE and MI. 

The Bottom Line

RV strain pattern
  • T-wave inversion in V1 and V2
  • At least one of the following:
    • T-wave inversion in lead III
    • The precordial lead with deepest T-wave inversion is V1 or V2

RV injury pattern
  • ST elevation in aVR and ST depression in lead I
  • ST elevation in V1-V3 and/or ST depression in V4-V6 

  • Acute PE often mimics MI by causing two patterns, which may occur together or separately.   Both patterns are usually associated with submassive or massive PE, requiring immediate recognition and treatment.   
  • RV injury pattern must be included in the differential diagnosis of ST elevation in aVR.
  • RV injury pattern is not specific for PE.   In the presence of this pattern, a terminal S-wave in lead I and a terminal R-wave in lead III may point towards PE.
  • When there is uncertainty regarding whether the patient has PE or MI, there should be a low threshold to obtain immediate bedside echocardiography.   


(1) The frequency of EKG findings among case series depend on factors including the intensity of CT scanning to detect small pulmonary emboli and the study location (i.e., Kukla 2014 was performed a tertiary care centers, which tend to receive patients with submassive or massive PE in transfer).   Therefore, the sensitivity may not translate from one center to another.

(2) These findings have been reported in the literature previously.   This combination of features is based on my review of the above cardiograms, and has not been previously reported to my knowledge.   This should be considered a preliminary finding at best.   These findings are clearly not specific for PE, but rather are intended to suggest the possibilityof PE.  

(3) Since aVR and lead I point in nearly opposite directions, ST elevation in aVR and ST depression in lead I are really measuring the same electrical current.   If there is equivocal ST elevation in aVR then lead I may be used as a tie-breaker: significant ST depression in lead I would suggest that there is indeed some real ST elevation in aVR.  

(4) Appreciate permission of Mike Cadogan to use this image, with modifications (original image at LITFL site).  


Ciliberti 2012   Massive pulmonary embolism with acute coronary syndrome-like electrocardiogram mimicking acute left main coronary artery obstruction.   Journal of Emergency Medicine.

George 2010   aVR - the forgotten lead.   Experimental and Clinical Cardiology.

Janata 2012   The role of ST-segment elevation in lead aVR in the risk assessment of patients with acute pulmonary embolism.   Clinical Research in Cardiology.

Lee 2013   Critical ostial left main and right coronary artery stenosis secondary to takayasu arteritis in a young female simulating pulmonary embolism at presentation.   Journal of Invasive Cardiology.  

Kosuge 2014   Differences in negative T waves between acute pulmonary embolism and acute coronary syndrome.   Circulation Journal.

Kukla 2014   Electrocardiographic abnormalities in patients with acute pulmonary embolism complicated by cardiogenic shock.   American Journal of Emergency Medicine.

Levis 2011 EKG Diagnosis: Pulmonary Embolism.   The Permanente Journal.

Livaditis 2004   Massive pulmonary embolism with ST elevation in leads V1-V3 and successful thrombolysis with tenecteplase.   Heart.

Mohsen 2013   Variable ECG findings associated with pulmonary embolism.   BMJ Case Reports.

Petrov 2013   Submassive pulmonary embolism - a watch-and-wait strategy with anticoagulation alone or advanced therapy with thrombolysis.   American Journal of Medicine Studies.  

Spodick 1972   Electrocardiographic responses to pulmonary embolism: Mechanisms and sources of variability.   American Journal of Cardiology.

Toprak 2014   Pulmonary embolism with ST-segment elevation in V1-3 and AVR treated successfully by catheter directed high-dose bolus thrombolytic therapy under cardiopulmonary resuscitation.   American Journal of Emergency Medicine.

Ullman 2001   Electrocardiographic manifestations of pulmonary embolism.   American Journal of Emergency Medicine.

Zhong-qun 2013   A new electrocardiogram finding for massive pulmonary embolism: ST elevation in lead aVR with ST depression in leads I and V4-V6.   American Journal of Emergency Medicine.