Spinal Epidural AbscessÐExperience with 46 Patients and Evaluation of Prognostic Factors

Transcription

1 Journal of Infection (2002) 45: 76±81 doi: /jinf , available online at on Spinal Epidural AbscessÐExperience with 46 Patients and Evaluation of Prognostic Factors H.-J. Tang 1, H.-J. Lin 2, Y.-C. Liu 3 and C.-M. Li* 1 1 Section of Infectious Diseases, 2 Section of Neurology, Department of Internal Medicine, Chi-Mei Medical Center, Tainan and 3 Section of Infection Diseases, Department of Internal Medicine, Kaohsiung Veteran General Hospital, Kaohsiung, Taiwan Objective: Spinal epidural abscess is a rare infectious disorder, often with a delayed diagnosis, and associated with significant morbidity and mortality rates. We conducted a retrospective study to define its clinical characteristics and to evaluate its prognostic factors. Methods: The medical charts of 46 patients (36 men and 10 women) with spinal epidural abscess over a 10-year period (from July 1991 to May 2000) were reviewed and analyzed. Results: A high proportion of patients had underlying diseases of diabetes (46%), frequent venous puncture (35%), spinal trauma (24%), and history of spinal surgery (22%). The initial accurate diagnostic rate was 11%. Localized spinal pain (89%), paralysis (80%), fever/chills (67%), and radicular pain (57%) were the common manifestations. The erythrocyte sedimentation rate (ESR) was elevated uniformly (mean, 86.6 mm/h) when measured. Staphylococcus aureus was the most common cause of spinal epidural abscess isolated from blood (39%) and pus (50%). Low platelet counts (< /L) (odds ratio (OR): 8.8, 95% confidence interval (CI): 1.0±77.8), extremely high ESR (110 mm/h) (OR: 4.8, 95% CI: 0.9±25.4), and cervical spine epidural abscess (OR: 5.2, 95% CI: 0.8±32.0) predicted a poor outcome. Conclusions: Localized back pain in a febrile patient with significant risk for epidural abscess warrants an immediate magnetic resonance imaging examination. The presence of thrombocytopenia, extremely elevated ESR, or evidence of spinal cord compression should prompt aggressive treatment. # 2002 The British Infection Society Introduction In 1948, Heusner studied 20 patients with spinal epidural abscess (SEA) and emphasized the importance of early diagnosis [1]. The medical literature for SEA has two common threads: reports of poor prognoses and appeals for rapid treatment. Nonetheless, despite more than 100 publications during the past quarter century, rates of death and paralysis reported recently are similar to those recorded by Baker and Colleagues in 1975 [2,3]. The diagnosis of SEA is a challenge largely because of its rarity among patients with back pain [4]. Among all patients admitted to hospitals, the incidence is approximately one to two cases per 10,000 [5]. Many * Please address all corresponding to: C.-M. Li, Section of Infectious Diseases, Department of Internal Medicine, Chi-Mei Medical Center, 901 Chung Hwa Road, Yung Kang City, Tainan, Taiwan, ROC. Tel.: , ext 3170; Fax: ; address: 8409d1@mail.chimei.org.tw (Chien-Ming Li). physicians would have had no chance to see a case during their practice. Furthermore, the initial symptoms may mimic several medical conditions. The patient with SEA may present with abdominal pain or flank pain and without high fever, neck stiffness, or elevated white blood cell count. He may be treated for urolithiasis or gastrointestinal disturbance. The nature of SEA as a great imitator makes early diagnosis rather difficult, and the consequences of delayed intervention are potentially devastating. Since the time of the review by Baker et al, new imaging modalities have become available. In addition, effective antimicrobial agents have been developed and various risk factors have emerged as significant. We reviewed our experience with 46 patients with SEA over a 10-year period and describe the clinical characteristics and explore the potential prognostic factors for outcome. As only increased awareness and swift management will improve the outcome of SEA, our goal is to enhance its early recognition and treatment. 0163±4453/02/$35.00 # 2002 The British Infection Society

2 Spinal Epidural Abscess 77 Methods and Materials We searched the computer records of Chi-Mei Medical Center, a 1200-bed teaching hospital in southern Taiwan for patients with SEA (ICD-9:324.1) from July 1991 to May All the medical records, laboratory data, radiological imaging studies, and operative notes were reviewed. The inclusion criteria required the identification of an abscess at surgery, or radiological evidence of SEA with a positive blood or aspirated pus culture. Infective spondylitis, paraspinal abscess without epidural involvement, and tuberculous SEA were not included in this study. The radiological diagnosis of SEA was based upon computerized tomography±myelography or magnetic resonance imaging (MRI) results. The outcome was assessed at the last follow-up clinic visit for survivors. The outcome was classified as good if there was significant improvement in neurological deficits including muscle power and paresthesia, as well as pain relief. A poor outcome was defined as no improvement in neurological impairments, death, or relapse of disease. Statistical analysis Comparison was made between patients with good outcomes and those with poor outcomes. Continuous data were compared using Student's t test or the Mann± Whitney test, as appropriate. Categorical data were compared using the chi-square test or Fisher's exact test, as appropriate. Logistic regression modeling was used to analyze the impact of potential prognostic factors on outcome. Variables with a P value less than 0.2 in the univariate analyses were entered to a multivariate model, and then a backward, stepwise method was used to select the final model. The two-tailed significance level was set at P ˆ Results Clinical features A total of 46 patients with SEA were included in the review. There were 36 men and 10 women, with a mean age of 60 years (range: 31±89 years). Their clinical presentations are listed in Table I. Back pain was the most common symptom, and radicular pain was a complaint of more than half of the patients. Four-fifths (37/46) of the patients had paralysis; more than onethird (17/46) had sphincter dysfunction at the time of admission. Fever and chills, the common symptom Table I. abscess. Clinical presentation of 46 patients with spinal epidural Symptoms/signs Number of patients % Fever/chills Spinal pain Confusion 3 7 Radicular pain Bowel/bladder dysfunction Local tenderness Neck stiffness 8 17 Paresthesia Paralysis (paraparesis, paraplegia) Table II. abscess. Underlying diseases of 46 patients with spinal epidural Underlying diseases Number of patients % Diabetes mellitus Liver disease 5 11 Chronic intravenous injection Spinal surgery Trauma of spine and sign indicating infection, were found in two-thirds (31/46) of the patients. The underlying diseases are summarized in Table II. Almost half (21/46) of the patients were diabetics. Frequent intravenous injection of steroids, nonsteroidal anti-inflammatory drugs (NSAIDs), or illicit drugs (heroin) was found in 35% (16/46) of the patients. Prior spinal trauma (11 patients) and spinal surgery (10 patients) were also common predisposing factors. The duration of symptoms before hospitalization ranged from 1 to 180 days (median: 7 days). The length of hospital stay varied from 7 to 77 days (median: 32 days). The initial diagnosis was in error in 34 patients (74%) and included sepsis of unknown origin (9 patients), spondylitis (9 patients), renal stone or abscess (4 patients), acute pyelonephritis (3 patients), degenerative joint disease of spine (4 patients), spinal metastasis (2 patients), meningitis (2 patients), and deep neck infection (1 patient). The epidural abscess was located in the cervical spine in 9 patients, thoracic spine in 14 patients, and lumbar spine in 23 patients. Extension of the abscess along the spine ranged from 1 to 13 vertebral bodies with a median of 2 vertebral bodies. The comorbidity included infective spondylitis (osteomyelitis or discitis) in 35 patients (76%), and paraspinal abscess (including psoas muscle abscess) in 25 patients (54%).

3 78 H.-J. Tang et al. Table III. Bacteriological findings from blood and aspirated pus cultures in 46 patients with spinal epidural abscess. Microorganism Blood Pus Number of patients (n ˆ 46) % Number of patients (n ˆ 28) Staphylococcus aureus (OSSA)* Staphylococcus aureus (ORSA)** Viridans streptococci Streptococcus agalactiae Klebsiella pneumoniae Escherichia coli Salmomnella enteritidis Coagulase (±) Staphylococci Serratia marcescens Enterococcus fecalis Morganella morgani No growth *OSSA: oxacillin susceptible Staphylococcus aureus. **ORSA: oxacillin resistant Staphylococcus aureus. Bacteriologic findings Blood cultures were obtained from all 46 patients and aspirated pus, from 28 patients (Table III). Staphylococcus aureus was the predominant pathogen, accounting for 39% (18/46) of the blood pathogens and 50% (14/28) of the pus culture pathogens. The overall blood culture positive rate was 70% (32/46). Four pus cultures (14%) yielded no organisms. The organisms isolated from the epidural pus were always the same as those identified in blood cultures. Outcome and prognostic factors Twenty-five patients underwent posterior decompression and drainage of the abscess, while 21 patients underwent medical treatment alone. The outcome was assessed at the last clinic visit for survivors, with a median follow-up period of 20 weeks (range: 4±48 weeks). Among the 46 patients, 33 (72%) were classified as having a good outcome, and 13 (28%, including 5 deaths and 7 relapses) as having a poor outcome. Table IV lists the univariate analysis of the potential prognostic factors for the outcomes. Only a low platelet count of less than /L emerged as significantly associated with poor outcome. Additionally, cervical spine involvement and high erythrocyte sedimentation % Table IV. Univariate analysis of potential prognostic factors for outcomes of 46 patients with spinal epidural abscess. Variables No. of patients (%) P Good outcome Poor outcome Age (year) * Gender Men 26 (79) 10 (77) Women 7 (21) 3 (23) Underlying disease Diabetes mellitus 13 (39) 8 (62) Liver disease 3 (9) 2 (15) Frequent injections 12 (36) 4 (31) Prior spine surgery 8 (24) 2 (15) Trauma 8 (24) 3 (23) Duration of symptoms, 7 (1±180) 7 (1±45) 0.208** median (range), d Comorbidity Osteomyelitis 25 (76) 8 (62) Disciitis 27 (82) 8 (62) Paraspinal abscess 18 (55) 7 (54) Treatment Surgical 17 (52) 8 (62) Medical 16 (48) 5 (38) Length of stay, 30 (7±77) 37 (7±76) 0.583** median (range), d Involved spine level Cervical 4 (12) 5 (38) Noncervical 29 (88) 8 (62) Involved range vertebral level vertebral levels vertebral levels vertebral levels vertebral levels vertebral levels vertebral levels vertebral levels 1 0 Lab data White blood cells 13 (39) 6 (946) /L Hemoglobin < 100 g/l 7 (21) 1 (9) Platelets < /L 2 (6) 4 (41) ESR 110 mm/h 6 (18) 6 (46) P values not specified are for c 2 test or Fisher's exact test, as appropriate. ESR: erythrocyte sedimentation rate. *Student's t test. **Mann±Whitney test. rate (ESR) greater than 110 mm/h appeared to have borderline significance. In multivariate logistic analysis (Table V), low platelet count remained a significant independent prognostic factor, with an odds ratio of 8.8 for a poor outcome (95% confidence interval 1.0±77.8). According to the data analysis, the infectious bacteria and clinical outcomes were not directly related.

4 Spinal Epidural Abscess 79 Table V. Multivariate analysis comparing poor outcome to good outcome. Variable OR 95% CI P Platelets < /L ± ESR 110 mm/h ± Cervical spine ± OR: odds ratio; CI: confidence interval; ESR: erythrocyte sedimentation rate. Discussion SEA is a collection of pus or inflammatory granulation tissue between the dura mater and the surrounding fatty tissue. Some authors associate pyogenic abscess with acute disease, and granulation tissue formation with a chronic course [1], while others feel that this distinction cannot be made [6]. SEA may be divided into three phases: acute, subacute, and chronic [2,7]. Acute SEA accounts for at least two-thirds of the reported cases [8,9]. Symptoms of acute SEA can be measured in hours to days. Patients with chronic SEA often present with similar signs and symptoms to those with acute SEA, but with a more prolonged course, measured in weeks to months. These are usually associated with a less virulent organism. Evidence of acute infection including fever, leukocytosis, and spinal tenderness is frequently absent in the chronic subtype. Symptoms and signs of subacute SEA often demonstrate both acute and chronic features. Among our 46 patients, 31 (67%) presented with acute type and the remainder had subacute or chronic type SEA. SEA is a rare entity. If it is not diagnosed and treated promptly, it results in high morbidity and mortality. A previous review by Nassbaum [5] suggested that the incidence of SEA has increased due to a greater number of spinal procedures including lumbar puncture, spinal surgery, epidural catheterization for pain control, the increased use of illicit intravenous drugs and a larger population of patients with immunodeficiency syndrome [10]. In our study, the use of illicit intravenous drugs was rare. However, frequent venous puncture for steroid or NSAID injection and acupuncture for treating joint pain or lower back pain were noted in more than onethird of patients (35%). The source of infection in SEA is, in most cases, hematogenous dissemination of bacteria from a localized focus elsewhere in the body, especially the skin. The bacteremia is often silent, and in some cases it is not possible to identify the focus. When a local infection is adjacent to the spine, spondylitis or paravertebral abscess for example can also propagate directly to the epidural space [11]. Another obvious route of infection to the epidural space is direct contamination from a penetrating wound or from a medical procedure, such as diagnostic or therapeutic punctures and surgical complications. As the abscess increases in size, infection follows the path of least resistance and extends laterally along the dural sheath. It can stretch along the entire length of the spinal cord [3]. The sites of predilection are often the lower thoracic and lumbar regions posterior to the spinal cord and the reasons for this are largely anatomical. The epidural space is wider at the thoracolumbar level and there is also more extensive extradural venous plexus (Batson's epidural plexus of veins). In addition, the dura is tightly, ventrally adherent to the posteriorlongitudinal ligament. A ``locus minoris resistantiae'' is one of the reasons why hematogenous spread of infection affects a particular site [12]. The mechanism of spinal cord necrosis is still unclear. The following three causes are most accepted: a decrease in arterial blood flow, venous thrombosis, or direct compression of the spinal cord. In our patients, the most common presenting symptoms and signs included pain (spinal pain in 89%, radicular pain in 57%), paralysis (80%), and fever/ chills (67%). The symptoms of SEA are easy to recognize in typical cases with fever, back pain, tenderness of the spine and radiating root pain followed by weakness, bowel or bladder dysfunction and paralysis of the limbs. Sometimes, septicemia may be the dominant presentation and the neurological symptoms may go unnoticed. This may also be the case among patients who are confined to bed for other reasons. In the presence of a chronic, localized infection, such as spondylitis, severe pain may be the initial symptom, without obvious signs of systemic infection. Weeks or months may elapse before extension of the infection into the epidural space produces neurological deficits. Pain is the most consistent symptom of SEA, occurring virtually in all patients at some time during their illness. It was accompanied by localized tenderness at the affected level in 90% of the cases. In chronic infections of the spine, malaise, body weight loss, and systemic upset are among the common presentations. In our study, the initial, accurate diagnosis rate was only 26% (12/46). The low accuracy of the diagnostic rate might have been due to: (1) incomplete neurological examination; (2) no detailed history taken from patients with sepsis syndrome; (3) poorly performed physical examinations on patients with septic syndrome; (4) lack of obvious typical symptoms and signs in some patients; (5) mistaken degenerative joint disease of the spine in an afebrile patient; (6) abuse of intravenous NSAID or steroids masking the symptoms and leading to transient bacteremia.

5 80 H.-J. Tang et al. Laboratory data suggestive of SEA are usually nonspecific. Patients with acute abscesses frequently have leukocytosis. The ESR is nearly always markedly elevated [11], as was seen in our study. The single infectious agent most commonly isolated is Staphylococcus aureus, responsible for more than 50% of all reported cases in the literature. The proportion in our study was lower (39%), and among these 18 patients, oxacillin-resistant isolates were found in 7, all who had had prior spinal surgery. Danner and Hartman found that in 5 studies from 1930 to 1982 the mean isolation frequency of S. aureus was 62%, for other Gram-positive cocci it was 10%, for aerobic gram negative bacilli the mean isolation frequency was 18%, and for anaerobes it was 2% [12]. Cases of cryptococcosis, actinomycosis, nocardiosis, aspergillosis, Fusobacterium nucleatum, Haemophilus aphrophilus, Torulopsis grabrata, and Eikenella corrodens have also been reported [13±15]. No discrepancy between blood and pus pathogens was found in this study. Although some reports showed good results with antibiotics alone, most suggest that emergency or urgent surgical drainage was warranted [16±18]. In view of clinical outcomes, there was no difference between surgical therapy and medical therapy alone in our study. Nonetheless, for some patients, medical treatment may be better than surgical treatment. Reasons for medical treatment include (1) being a poor surgical candidate, such as a patient with a bleeding tendency, (2) length of the abscess prohibits surgery, (3) lack of obvious neurological symptoms and signs, and (4) complete paralysis for more than 3 days. On the other hand, patients with neural compression from mass effect, spinal instability with deformity or failure to obtain a satisfactory culture of the infecting organism were candidates for surgical intervention [2,18]. In 1996, Rohit and Khanna found the following factors influenced outcome [19]: (1) age: better outcomes in younger patients. This was probably related to the higher prevalence of poor medical condition in older patients as well as possible better spinal cord plasticity in younger patients. (2) The degree of thecal sac compression: severe thecal sac compression (>50%) had a significantly poorer outcome. (3) Abscess location: better outcomes were noted when the SEA was in the lumbosacral area. This was due to less severe nerve root or cauda equina compression than direct cord compression in the cervical or thoracic spine. As patients with anteriorly located SEA had an increased risk of osteomyelitis, they had a poorer prognosis. (4) Surgical findings: patients with pustular abscesses had significantly better outcomes than did patients with granulation tissue. This likely reflected the acute or chronic presentation. (5) Septic presentation: patients with initial septic presentation had a poorer prognosis, (6). Duration of symptoms (such as bowel/bladder dysfunction or paralysis): longer duration lead to a poorer prognosis. In our analyses, we found that only low platelet count was a significant independent risk factor for poor outcome, while extremely high ESR (110 mm/h) and cervical spine involvement had borderline significance contributing to poor outcome. For the latter, the unstable estimates of odds ratios might have been due to small sample size. A low platelet count implied clinical sepsis and impending disseminated intravascular coagulation and, thus, accounted for poor outcome. This finding is consistent with the report of Rohit and Khanna that septic presentation was a prognostic factor for a poor outcome [19]. ESR is a highly sensitive but less specific parameter for inflammation. Elevated ESR might have indicated long disease duration and serious severity. In cervical epidural abscess, direct compression of the spinal cord might have accounted for the grave outcome. In conclusion, in view of the facts that the highly diverse clinical presentations that could mimic many other diseases, and the still excessive morbidity and mortality that could be due to delays in diagnosis, early diagnosis of SEA remains the most essential factor to prevent devastating outcomes. All physicians treating patients for back pain must have a high index of suspicion, particularly in those patients who are diabetics or chronic intravenous drug users with a history of fever and spinal tenderness. Appropriate laboratory data such as ESR and counts of blood cells could indicate an inflammatory process or impending sepsis, and a spine MRI is warranted for early accurate detection of the lesions. References 1 Heusner AP. Nontuberculous spinal epidural infections. N Engl J Med 1948; 239: 845± Baker AS, Ojemann RG, Swartz MN, Richardson EP Jr. Spinal epidural abscess. N Engl J Med 1975; 293: 463± Hlavin ML, Kaminski HJ, Ross JS, Ganz E. Spinal epidural abscess: a ten-year perspective. Neurosurgery 1990; 27: 177± Deyo R. Early diagnostic evaluation of low back pain. J Gen Intern Med 1986; 1: 328± Nussbaum ES, Rigamonti D, Standiford H, Numaguchi Y, Wolf AL, Robinson WL. Spinal epidural abscess: a report of 40 cases and review. Surg Neurol 1992; 38: 225± Del Curling O Jr, Gower W, McWhorten JM. Changing concepts in spinal epidural abscess: a report of 29 cases. Neurosurgery 1990; 27: 185± Browder J, Meyers R. Infections of the spinal epidural space: an aspect of vertebral osteomyelitis. Am J Sur 1987; 37: 4±26. 8 Reihsaus E, Waldbaur H, Seeling W. Spinal epidural abscess: a metaanalysis of 915 patients. Neurosurg Rev 2000; 23: 175± Altrocchi P. Acute spinal epidural abscess vs. acute transverse myelopathy. Arch Neurol 1963; 9: 17±25.

6 Spinal Epidural Abscess Doroviche RO, Hamill RJ, Greenberg SB, Weather SW, Musher DM. Bacterial spinal epidural abscess: review of 43 cases and literature survey. Medicine 1992; 71: 369± Verner EF, Musher DM. Spinal epidural abscess, symposium on infections of central nervous systems. Med Clinic North Am 1985; 69: 375± Danner RL, Hartman BJ. Update of spinal epidural abscess: 35 cases and review of the literature. Rev Infect Dis 1987; 9: 265± Kaufman DM, Kaplan JG, Litman N. Infectious agents in spinal epidural abscess. Neurology 1980; 30: 844± Kannangara DW, Tanaka T, Thadepalli H. Spinal epidural abscess due to Actinomyces israelii. Neurology 1981; 31: 202± Allison L, Harvey JM, Lisa O. A case of Nocardia epidural abscess. J Emerg Med 1998; 16(4): 579± Cooper K. Spinal epidural abscess. A case report. Del Med J 1984; 56: 343± Ravicovitch MA, Spallone A. Spinal epidural abscess: surgical and parasurgical management. Eur Neurol 1982; 21: 347± Korbin W. Spinal epidural abscess: review of literature and report of case. Bull LA Neurol Soc 1958; 23: 21± Khanna PK, Malik GM, Rock JP, Rosenblum ML. Spinal epidural abscess: evaluation of factors influencing outcome. Neurosurgery 1996; 39: 958±964.