SPINAL EPIDURAL ABSCESSES are suppurative infections

Transcription

1 512 ORIGINAL ARTICLE Spinal Epidural Abscess: A 5-Year Case-Controlled Review of Neurologic Outcomes After Rehabilitation David W. Koo, MD, Andrea F. Townson, MD, Marcel F. Dvorak, MD, Charles G. Fisher, MD ABSTRACT. Koo DW, Townson AF, Dvorak MF, Fisher CG. Spinal epidural abscess: a 5-year case-controlled review of neurologic outcomes after rehabilitation. Arch Phys Med Rehabil 2009;90: Objective: To describe the neurologic outcomes of 29 spinal epidural abscess (SEA) patients after rehabilitation compared with a case-controlled traumatic spinal cord injury (TSCI) cohort. Design: Five-year retrospective chart review. Setting: University-affiliated surgical spine unit and inpatient rehabilitation program. Participants: Patients (n 29; 19 men, 10 women) requiring inpatient rehabilitation after SEA and TSCI case controls (n 29) matched by level of injury, American Spinal Injury Association (ASIA) Impairment Scale (), ASIA motor score (AMS), sex, and age. Main Outcome Measure: The primary outcome was a change in AMS from acute admission to discharge from inpatient rehabilitation. Results: Despite having lower admission motor scores, there was a significant trend toward greater neurologic recovery in the SEA group (P.047). In contrast to what is known regarding recovery from complete TSCI, this study shows potential for dramatic recovery in SEA subjects presenting with grade A deficit with a 73% conversion rate to incomplete status; out of 11 SEA subjects with initial grade A, 2 improved to grade B, 1 to grade C, and 5 to grade D. Conclusions: This study shows the potential for a significant improvement in neurologic deficits related to SEA. Based on the results of our study, it is clearly inappropriate to generalize recovery patterns seen in the TSCI patient population to SEAassociated myelopathy because the latter appears to have more favorable outcomes. Key Words: Epidural abscess; Prognosis; Rehabilitation; Spinal cord injuries by the American Congress of Rehabilitation Medicine From the Department of Medicine, Division of Physical Medicine and Rehabilitation, Spinal Cord Injury Program, G.F. Strong Rehabiliation Centre (Koo, Townson); Department of Orthopaedics, Division of Spine, Combined Neurosurgical and Orthopaedic Spine Program, Vancouver General Hospital (Dvorak, Fisher); and International Collaboration On Repair Discoveries, (Koo, Townson, Dvorak, Fisher), University of British Columbia, Vancouver, BC, Canada. Supported by the Rick Hansen Man in Motion Fund (grant no ). The study was reviewed and approved by the ethics and research advisory committees of the University of British Columbia, Vancouver General Hospital and GF Strong Rehabilitation Centre. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to David W. Koo, MD, G.F. Strong Rehab Centre, 4255 Laurel St, Vancouver, BC, Canada V5Z 2G9; david.koo@vch.ca /09/ $36.00/0 doi: /j.apmr SPINAL EPIDURAL ABSCESSES are suppurative infections localized adjacent to the dura mater. About half of all infections result from contiguous spread, and the remainder are either hematogenous or of unknown origin. Estimates of SEA incidence vary between 0.2 and 2 cases per 10,000 hospital admissions. The incidence of SEA as a cause of spinal cord impairment has increased dramatically over the past 2 decades, likely because of the increasing prevalence of IVDU, invasive medical procedures, and an aging population. 1,2 The classic presentation of early stages of SEA is back or neck pain in the presence of abnormal inflammatory parameters, including fever, leukocytosis, and elevated erythrocyte sedimentation rate. With disease progression, radiculopathy or myelopathy leading to complete paralysis may develop. Khanna et al 3 found that back or neck pain occurred most commonly (87.8%), followed by fever (61.0%), paresis (53.7%), bladder/bowel dysfunction (36.6%), sepsis (17.1%), radiculopathy (12.2%), and plegia (14.6%). Although the number of reported series and cases are small, it is estimated that SEA results in complete or incomplete paraplegia in approximately 31% of cases and tetraplegia in 3%. 4 It is not surprising that the most commonly cultured organisms are gram-positive cocci and, in particular, Staphylococcus aureus because conditions such as IVDU lead to regular invasion of skin flora into the bloodstream. 4 The recommended treatment for all but early and very late stages of this disease includes surgical decompression and stabilization with prolonged antibiotic therapy. Medical management without surgery is reserved for those at significant surgical risk, those with extensive SEA without neurologic sequelae, and those with complete paralysis lasting more than 48 to 72 hours. 5,6 Savage et al 7 found that medical treatment alone produced excellent or good results for SEA patients 83% of the time when there was no evidence of systemic sepsis and the neurologic examination was normal or stable. Although there is little literature to clearly delineate the issues of timing of surgical intervention, it is the general consensus that early surgical decompression will prevent the progression of neurologic impairment. 3,4 Although there has been extensive reporting on risk factors such as IV drug abuse for developing SEA, there has been limited literature regarding factors influencing neurologic impairment or predicting neurologic and functional recovery. 5,8 The pathogenesis of neurologic deficit in patients with spinal epidural abscess has been the subject of much study and debate ASIA AMS HIV IV IVDU SEA TSCI List of Abbreviations American Spinal Injury Association ASIA Impairment Scale ASIA motor score human immunodeficiency virus intravenous intravenous drug use spinal epidural abscess traumatic spinal cord injury

2 SPINAL EPIDURAL ABSCESS OUTCOMES, Koo 513 at the basic science level. Proposed mechanisms include septic thrombophlebitis, which impairs blood flow, 1 and gradually progressive mechanical compression with direct injury to neural tissue. 9,10 The proposed mechanisms for neurologic injury after SEA are different than that after TSCI, and, therefore, in theory, neurologic outcomes may also be different. Most studies evaluating SEA have focused on acute management and treatment considerations. Few studies have addressed the rehabilitation of the patient with SEA paralysis, particularly exploration of the predictors of recovery and eventual outcomes. 5,11,12 As noted by McKinley et al, 11 there is a need for studies with more controlled matching for levels and completeness of injury. Therefore, the purpose of this study was to describe the neurologic recovery of patients presenting with SEA-related myelopathy after inpatient rehabilitation and to compare them with that of a matched cohort of individuals with TSCI. Secondary outcomes included assessment of the association of baseline variables with neurologic recovery. METHODS Study Design Our retrospective cohort study was composed of patients with SEA admitted to a tertiary care spine center from 1999 to The study was reviewed and approved by the ethics and research advisory committees of the University of British Columbia, Vancouver General Hospital, and GF Strong Rehabilitation Centre. Patient Populations SEA group. The SEA group consisted of a consecutive series of patients with a primary diagnosis of SEA admitted to a tertiary care spinal unit from 1999 to All patients were age 18 and over and had a major neurologic deficit requiring inpatient rehabilitation. The G.F. Strong Rehabilitation Centre is the sole provider of tertiary rehabilitation services for patients with spinal cord injury for the province of British Columbia. Exclusion criteria were patients with comorbidities such as primary neurologic disease, traumatic brain injury, and other conditions that affected neurologic assessment or outcome. Patients with Pott disease (spinal tuberculosis) and fungal infections were excluded. TSCI group. The TSCI control group consisted of casematched individuals who underwent inpatient rehabilitation at G.F. Strong Rehabilitation Centre during the same 5-year period as the SEA group. The TSCI cases were identified through a database review and were selected to be the best-available match based on the following variables in descending order of priority: last normal neurologic level,, AMS at acute admission, age, and sex. Table 1 presents comparative data for the demographic and neurologic variables of the SEA and TSCI groups. The database and charts were reviewed to obtain baseline variables including age, sex, neurologic level, and completeness of injury. Neurologic status (ASIA last normal neurologic level and AMS) was prospectively documented within 72 hours of the neurologic event or presentation to our facility on the patient chart and in a research database by the Combined Neurosurgical and Orthopedic Spine Service. Neurologic classification at discharge from rehabilitation was performed by physiatrists trained and knowledgeable in the neurologic classification of impairment as described by the ASIA. 13 SEA risk factors including IVDU, alcoholism, diabetes mellitus, immune-compromised status (HIV or immunosuppressive treatment), skin infection, urinary tract infection, spinal Table 1: Demographic and Neurologic Characteristics of the SEA Group TSCI at Acute Admission Characteristics SEA Group (n 29) TSCI Group (n 29) Age (y) (mean SD) Sex Male 19 (65.5) 19 (65.5) NS Female 10 (34.5) 10 (34.5) NS Neurologic level (n) Cervical 16 (55.2) 15 (51.7) NS Thoracic 13 (44.8) 14 (48.3) NS Completeness of injury grade A 11 (37.9) 11 (37.9) NS grade B 10 (34.5) 11 (37.9) NS grade C 8 (27.6) 5 (17.2) NS grade D 0 (0) 2 (6.9) NS Admission motor score (mean SD) Average length of stay (d) NS NOTE. Values in parenthesis are percentages. Abbreviation: NS, not significant. surgery, trauma, and malignancy were documented. The causative infective agent was identified through the review of tissue or blood cultures. The primary outcome measure was a change in AMS from acute hospital presentation to discharge from inpatient rehabilitation. Data Analysis The study design used matched pairs (29 cases). The primary analysis was a matched-pairs t test to assess pair-wise differences with respect to age, AMS at acute admission and discharge from rehabilitation, and rate of change of AMS. Multiple regression analyses were performed on the unpaired data (58 cases) by using the AMS as the dependent outcome variable; a binary variable indicating SEA or TSCI group; and age, sex, and as independent predictors. was looked at 3 ways: as an interval measure (1 4 corresponding to A D), as a binary variable of complete versus incomplete (A vs B, C, and D), and as a binary variable of motor complete versus motor incomplete (A and B vs C and D). RESULTS A total of 29 patients with SEA were identified as receiving inpatient rehabilitation at G.F. Strong Rehabilitation Centre between 1999 and The demographics are displayed in table 1. The case-matched TSCI group did not differ significantly from the SEA group in sex, neurologic level, or percentage of motor completeness at presentation. The mean age of the SEA group of 47.6 years was significantly older than the TSCI group of 38.7 years (P.000), and the average initial motor score of 30.8 in the SEA group was significantly lower than the 36.8 in the TSCI group (P.047). Nineteen of the 29 SEA patients were treated with surgical decompression and stabilization, whereas all received IV antibiotic management of a minimum of 6 weeks duration based on culture sensitivities. All abscesses were in the cervical and thoracic region. Complete spinal cord injury, grade A, was the most common neurologic presentation in the SEA group. The pattern of neurology based on did not differ between the SEA and TSCI groups. Identified risk factors for SEA in descending order were IVDU (68%), alcoholism (28%), HIV (18%), diabetes (18%), P

3 514 SPINAL EPIDURAL ABSCESS OUTCOMES, Koo Table 2: SEA Versus TSCI AMS on Acute Admission, Discharge From Rehabilitation, and Total Change in Motor Score Groups Acute Admission AMS Discharge AMS Change in AMS SEA * TSCI * NOTE. Values are mean SD. *P.000. P.047. age older than 65 years (14%), and cutaneous infection (3.5%). There was 1 case in which no identifiable factors were established (3.5%). The rate of viral hepatitis (hepatitis B and/or C) was 53%. All HIV cases were coinfected with viral hepatitis. The causative organism was S aureus in 23 cases (82%), Streptococcus viridans in 1 case (3.5%), and not identifiable in 4 (14%). Table 2 shows that the change in AMS from acute presentation to discharge from rehabilitation was significantly greater in the SEA group. Table 3 correlates SEA cases by from acute presentation to the time of discharge from rehabilitation. For comparison, table 4 shows the changes noted in the TSCI subjects for the same interval. DISCUSSION This study shows the potential for significant neurologic improvement in neurologic deficits related to SEA. The average change in motor score was 26.4 in the SEA group and 16.2 in the TSCI group, both a clinically and statistically significant difference. This result is further strengthened by the SEA group starting with a lower motor score and having an older average age, 2 factors that would tend to bias toward the null hypothesis. With respect to grading, there was an overall conversion rate of 73% from complete ( grade A) to incomplete ( grade B, C, or D) status compared with 9% in the TSCI cohort. Additionally, conversion from a motor complete ( grade A or B) to motor incomplete ( grade C or D) occurred in 76% of SEA cases compared with 32% in the TSCI group. Based on the results of our study, it is clearly inappropriate to generalize recovery patterns seen in the TSCI patient population to SEA-related myelopathy because the latter appears to have more favorable outcomes. Other studies looking at recovery patterns after paralysis from SEA support the potential for significant improvement, but these studies have had several limitations. All are retrospective case series and therefore lack standardized, prospective neurologic evaluation. Most have lacked standardized classification of deficit, often grouping varying degrees of incomplete and complete neurologic deficit under the outcome of paralysis. Furthermore, a majority of these case series lack a control group, preventing comparison to the more Table 3: The Initial SEA Grade at Acute Hospital Admission Correlated to Grade at Discharge From Rehabilitation SEA Grade at Admission SEA Grade at Discharge From Rehabilitation Grade A Grade B Grade C Grade D Grade E grade A (n 11) 3 (27) 2 (18) 1 (9) 5 (46) 0 grade B (n 10) (50) 5 (50) 0 grade C (n 8) (38) 5 (62) 0 grade D (n 0) grade E (n 0) Table 4: The Initial TSCI Grade at Acute Hospital Admission Correlated to Grade at Discharge From Rehabilitation TSCI Grade at Admission TSCI Grade at Discharge From Rehabilitation Grade A Grade B Grade C Grade D Grade E grade A (n 11) 10 (91) (9) 0 grade B (n 11) 0 5 (45) 3 (27) 3 (27) 0 grade C (n 5) (20) 4 (80) 0 grade D (n 2) (100) 0 grade E (n 0) extensively studied outcomes in the TSCI population. 4,6,12,14 SEA outcome studies that did include TSCI patients for comparison reported baseline differences in subject demographics and initial FIM motor scores as limiting factors to their validity. 5,11 With the limitations of these studies in mind, previously reported outcomes after SEA myelopathy have been varied. Rigamonti et al 14 were generally supportive of our conclusions, finding that 45% of 29 cases with severe motor deficit and/or incontinence recovered independent ambulation when surgically decompressed within 48 hours of presentation. A smaller series found that 5 of 7 paralyzed SEA patients died and that despite early surgical intervention ranging from 6 to 12 hours from onset of paralysis to surgery, the remaining 2 did not recover function, although with such high mortality, meaningful statistical analysis of morbidity was not possible. 2 In a retrospective case-controlled study, the SEA group averaged a 15-point increase during rehabilitation and were more likely to suffer pressure ulcer complications compared with the TSCI group, which averaged 30 points. However, in that series, the TSCI group began with a higher FIM score at admission and had a statistically longer length of inpatient rehabilitation stay. 5 Similarly, another series found reduced FIM motor score changes and FIM motor efficiency scores in subjects with infection-related spinal cord disease compared with their TSCI counterparts during rehabilitation; however, the groups were not controlled for age or level and completeness of injury. 11 Previous studies have reported serious neurologic deficit as sequelae of SEA with paralysis rates as high as 21% to 39% in some series. Because our series focused on patients requiring inpatient rehabilitation, all of our patients had a major neurologic deficit at presentation with a high percentage of motor complete (72.4%) patients. Patients with less severe spinal cord involvement ( grade D) were not captured in this series; a probable explanation is that those with less deficit may have received rehabilitation on an outpatient basis or were transferred back to their originating hospitals for general rehabilitation and/or discharge planning. The 29 SEA subjects represent 4.1% of the total SCI admissions to our rehabilitation program during the 5-year period. To our knowledge, this article presents the largest outcome assessment of motor complete SEA myelopathy to date. The 68% rate of IVDU in our case series is higher than that seen in an overall meta-analysis of 854 patients with identifiable comorbid factors, which found 57 (9%) were IV drug users. 4 In the literature, case series in U.S. sites have reported percentages between 17.5% and 59.4%. 2,3,7,8,14 The particularly high incidence of SEA seen in our cohort may stem from the high prevalence of IVDU and HIV in the Greater Vancouver Region. We also theorize that those with ongoing addiction issues may have delayed presentation to medical care, resulting

4 SPINAL EPIDURAL ABSCESS OUTCOMES, Koo 515 in a higher-than-expected level of neurologic impairment requiring inpatient rehabilitation. The number of SEA subjects with motor complete ( grade A and B) impairment in this study was higher than previous studies. 11,12 The rate of conversion from complete ( grade A) to incomplete ( grade B, C, or D) in the SEA group was 73% in this series, which was significantly higher than 9% in the TSCI comparator group. In the SEA literature, Boström et al 12 reported a similar conversion rate from Frankel A to Frankel B, C, D, or E of 75%. This is significantly higher than rates of conversion of grade A cases in the TSCI literature with reported ranges of 4.1% to 25.5%. 18,19 The potential for significant distal motor recovery in the SEA group is also noteworthy in this case series. Of the 11 SEA cases initially classified as motor and sensory complete, 1 improved to grade C and 5 improved to grade D, with an overall 54% rate of distal motor recovery. This is higher than previous reports of distal motor recovery in complete TSCI in which studies have estimated this to occur in 2.5% to 15% of cases. 20,21 In our cohort, overall, 79% of SEA subjects improved by at least 1 classification, in keeping with the 71% improvement in Frankel grades reported by Bostrom et al. 12 In contrast, McKinley et al 11 reported that only 32% of their infectionrelated spinal cord disease group improved in classification; however, their cohort was more etiologically heterogeneous and may have had a ceiling effect because 88% of their subjects were already grade C or D at presentation. Regression analysis revealed no significant prognostic factors with respect to age, sex, and admission to the degree of neurologic recovery. The high rate of neurologic improvement in the SEA group as well as the small sample size in the subgroups may have been a limiting factor in determining statistical significance. Other series have suggested that completeness of neurologic deficit and motor deficit exceeding 36 hours are associated with a poorer neurologic outcome, although the outcome measures were not described. 17 Rigamonti et al 14 found that patients treated within 24 hours of diagnosis had improved outcomes, as did those with incomplete cord lesions with sensory sparing. Methicillin-resistant S aureus and thoracic distribution of SEA were associated with a poorer prognosis. Magnetic resonance imaging findings of narrowing of the central spinal canal of 50% or more, peripheral contrast enhancement, and abnormal spinal cord signal intensity have also been associated with poorer outcomes after SEA. 22 It is important to consider the potential limitations of the current investigation. Despite efforts to control discrepancies between baseline characteristics in the SEA group and the TSCI-matched cases, controlling for level and completeness of injury resulted in a statistically older mean age and lower initial motor score in the SEA group. However, older age and greater severity of paralysis are usually associated with poorer neurologic recovery after SCI and, in theory, should bias against the SEA group when compared with TSCI. A possible confounding factor may have been the reliability of initial assessment in the acute period, given the high rate of IVDU in the SEA group. The exclusion of spinal shock, through the documentation of bulbocavernous reflex return, was not consistently recorded in the acute records, although this would arguably have equally affected both study groups. Fisher et al 23 found that by eliminating potentially misclassified grade A cases by documenting resolution of spinal shock and eliminating subjects with confounding assessment factors including head injury, intoxicating substances, or severe multitrauma, none of the complete TSCI cases showed distal motor recovery by the 2-year follow-up. In contrast to other case series, there were no lumbosacral epidural abscesses or initial grade D impairments in our study. Other series have reported lumbosacral distribution in 23% to 46% of SEA subjects 2,4,6,11,12,17 and have noted grade D severity in up to 45%. 5 Because our series focused on subjects requiring inpatient rehabilitation at a tertiary rehabilitation center, there was a high percentage of motor complete (72.4%) patients. Patients with less severe functional impairments (ie, lumbosacral distribution or grade D) were not captured in this series; a possible explanation is that those with less of a deficit may have received outpatient rehabilitation or were transferred back to a regional hospital for general mobilization and/or discharge planning. Another contributing factor may have been the high rate of IVDU in our series because our experience has been that many patients with active IVDU addiction may not choose to remain in a hospital setting if their mobility status permits otherwise and may therefore avoid transfer to an inpatient rehabilitation unit. Future investigations could explore longer-term outcomes after SEA-associated paralysis, particularly because TSCI outcome studies suggest that improvements can occur for up to 2 years postinjury. However, the feasibility of such follow-up would be challenging in our study group because of the transient living arrangements and noncompliance issues often facing the IVDU population after inpatient discharge. CONCLUSIONS This study shows the potential for significant improvement in neurologic deficits related to SEA, including those initially presenting with complete motor and sensory impairment. Based on the results of our study, it is clearly inappropriate to generalize recovery patterns seen in the TSCI patient population to SEA-associated myelopathy because the latter appears to have more favorable outcomes. Acknowledgments: We thank Caroline Abramson for assisting with data collection and analysis. References 1. Baker AS, Ojemann RG, Swartz MN, Richardson EP Jr. Spinal epidural abscess. N Engl J Med 1975;293: Hlavin ML, Kaminski HJ, Ross JS, Ganz E. Spinal epidural abscess: a ten-year perspective. Neurosurgery 1990;27: Khanna RK, Malik GM, Rock JR, Rosenblum ML. Spinal epidural abscess: evaluation of factors influencing outcome. Neurosurgery 1996;39: Reihsaus E, Waldbaur H, Seeling W. Spinal epidural abscess: a meta-analysis of 915 patients. Neurosurg Rev 2000;23: ; discussion Zafonte RD, Ricker JH, Hanks RA, Wood DL, Amin A, Lombard L. Spinal epidural abscess: study of early outcome. J Spinal Cord Med 2003;26: Curry WT, Hoh BL, Amin-Hanjari S, Eskandar E. Spinal epidural abscess: clinical presentation, management, and outcome. Surg Neurol 2005;63:364-71; discussion Savage K, Holtom PD, Zalavras CG. Spinal epidural abscess: early clinical outcome in patients treated medically. Clin Orthop Relat Res 2005;Oct(439): Ruiz A, Post MJ, Sklar EM, Holz A. MR imaging of infections of the cervical spine. Magn Reson Imaging Clin N Am 2000; 8: Kobrine AI, Evans DE, Rizzoli H. Correlation of spinal cord blood flow and function in experimental compression. Surg Neurol 1978;10: Feldenzer JA, McKeever PE, Schaberg DR, Campbell JA, Hoff JT. Experimental spinal epidural abscess: a pathophysiological model in the rabbit. Neurosurgery 1987;20:

5 516 SPINAL EPIDURAL ABSCESS OUTCOMES, Koo 11. McKinley W, Merrell C, Meade M, Brooke K, DiNicola A. Rehabilitation outcomes after infection-related spinal cord disease: a retrospective analysis. Am J Phys Med Rehabil 2008;87: Boström A, Oertel M, Ryang Y, et al. Treatment strategies and outcome in patients with non-tuberculous spinal epidural abscess a review of 46 cases. Minim Invasive Neurosurg 2008;51: Marino RJ, Barros T, Biering-Sorensen F, et al. International standards for neurological classification of spinal cord injury. J Spinal Cord Med 2003;26(Suppl 1):S 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 Darouiche RO, Hamill RJ, Greenberg SB, Weathers SW, Musher DM. Bacterial spinal epidural abscess. Review of 43 cases and literature survey. Medicine (Baltimore) 1992;71: Maslen DR, Jones SR, Crislip MA, Bracis R, Dworkin RJ, Flemming JE. Spinal epidural abscess. Optimizing patient care. Arch Intern Med 1993;153: Sampath P, Rigamonti D. Spinal epidural abscess: a review of epidemiology, diagnosis and treatment. J Spinal Disord 1999;12: Waters RL, Adkins RH, Yakura JS, Sie I. Motor and sensory recovery following complete tetraplegia. Arch Phys Med Rehabil 1993;74: Geisler FH, Coleman WP, Grieco G, Poonian D; Sygen Study Group. Measurements and recovery patterns in a multicenter study of acute spinal cord injury. Spine 2001;26(24 Suppl): S Katoh S, el Masry WS. Neurological recovery after conservative treatment of cervical cord injuries. J Bone Joint Surg Br 1994;76: Gerhart KA. Spinal cord injury outcomes in a population-based sample. J Trauma 1991;31: Tung GA, Yim JW, Mermel LA, Philip L, Rogg JM. Spinal epidural abscess: correlation between MRI findings and outcome. Neuroradiology 1999;41: Fisher CG, Noonan VK, Smith DE, Wing PC, Dvorak MF, Kwon B. Motor recovery, functional status, and health-related quality of life in patients with complete spinal cord injuries [published erratum appears in Spine 2006;31:1289]. Spine 2005;30: