Surgical Neurology 66 (2006) 141 – 147 www.surgicalneurology-online.com
Infection
Spinal epidural abscess due to Brucella ¨ mer Tural, MDa, AYkVn Gfrgqlq, MDa,4, Baki S. Albayrak, MDa, Evlin Gfrgqlq, MDc, O Tamer Karaaslan, MDa, Orhan Oyar, MDb, Mesut YVlmazd Departments of aNeurosurgery and bRadiology, University of Suleyman Demirel, School of Medicine, 32260, Isparta, Turkey c Department of Infectious Disease, Isparta State Hospital, 32100, Isparta, Turkey d Department of Infectious Disease, University of Istanbul, School of Medicine, C¸apa, Turkey Received 17 August 2005; accepted 29 October 2005
Abstract
Background: Brucellar spinal epidural abscess (SEA) is a rarely encountered clinical entity during the course of the systemic Brucella infection. Methods: We reported 9 patients diagnosed with Brucellar SEA with a mean follow-up of 20 months. Spinal epidural abscess was detected by magnetic resonance imaging in all cases. Brucella diagnosis was established by specific blood tests. Patients were administered antibiotics for a duration of 6 to 12 weeks. Results: Spinal epidural abscess was localized in lumbar region in 6 patients, dorsal in 2 patients, and cervical in 1 patient. Abscess mimicked disk herniation clinically in 3 patients. Although neurologic examination was normal in 6 patients, we detected motor deficit in 3 patients. Symptoms regressed in all patients but 1 after the institution of antibiotic regimens, and all recovered fully without any sequel. Surgical drainage of abscess was performed in 1 patient. Conclusions: Proper antibiotic regimens in required doses and duration should be the primary treatment in Brucellar SEA. The criteria for terminating antibiotic therapy are clinical recovery and dissolution of abscess images radiologically. Lastly, should any neurologic deterioration be detected during the course of medical treatment, surgical decompression is to be considered. D 2006 Elsevier Inc. All rights reserved.
Keywords:
Antibiotic therapy; Brucella; Spinal epidural abscess
1. Introduction Brucellosis is a systemic infectious disease, caused by a nonencapsulated, nonmotile, gram-negative bacilli belonging to the genus Brucella. This disease is a zoonosis transmitted to humans from infected animals such as goats, sheep, cows and dogs [3,8,11]. Although brucellosis is a controlled disease in developed countries, its prevalence is not exactly known in developing ones and still remains as a serious health problem. The prevalence is higher in Mediterranean area, the Arabian Peninsula, Mexico, and
Abbreviations: CRP, C-reactive protein; CSF, cerebral spinal fluid; MRI, magnetic resonance imaging; SEA, spinal epidural abscess; WBC, white blood cell. 4 Corresponding author. Sqleyman Demirel Bulvary´ 101 Cadde No 73 Sener Apt A Blok, Kat 3 Daire 5 32100 Isparta, Turkey. Fax: +90 246 2327350. E-mail address:
[email protected] (A. Gfrgqlq). 0090-3019/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.surneu.2005.10.019
Central and South America [6,25]. Brucellosis is endemic in Turkey. During last decade (between 1991 and 2000), approximately 9000 cases yearly were reported to Turkish Ministry of Health (incidence, 14/100 000). A screening of general population for seropositivity (Wright test, positive at N 1/160; n = 1054/58 707, 1.8%) reveals a much higher rate in Turkey [6,24]. Up to 50% of cases with brucellosis may have spinal involvement that is mostly seen as spondylitis [2,11], whereas abscess formation is very rare. In the present study, 9 cases of Brucellar SEA are discussed from the point of clinical and laboratory findings, diagnostic evaluations, and treatment methods particularly. 2. Patients and methods We evaluated 9 patients (3 men and 6 women with the mean age of 62.5 years) who were treated with the diagnosis
142
Table 1 Clinical characteristics of the patients 1
2
3
4
5
6
7
8
9
Age/gender Duration of symptoms Presentation
72/F 2 mo
53/M 45 d
53/F 15 d
47/M 1 mo
76/F 4 mo
69/F 45 d
53/F 1 mo
67/M 2 mo
73/F 3 mo
Low back pain and pain radiating down from hip Wright test: 1/640 Normal
Low back pain and pain in left leg Wright test: 1/320 + Normal
Pain in the dorsal Pain in the region and, difficulty dorsal region in walking Wright test: 1/1280 Wright test: 1/160
T8-T10
T7-T8
Low back pain and pain in the right knee and thigh Wright test: 1/320 + Strength in extensors, Normal flexors of forearms and wrists bilaterally 3/5, 4/5, 4/5, 4/5, respectively C6 -T1 L3-L4
S+R+T
R+T+C
R + T + (T S)
R+T
R+T
S+T
R+T+C
R+T+C
S+T
3 + 6 wk + 6 wk
3 mo
3 mo
2 mo
6 wk
3 + 6 wk
6 wk
3 mo
3 mo + operation
Serologic tests Lasegue Initial neurological (4/5 motor strength) examination
Location of abscesses Antibiotherapy regimen Duration of antibiotherapy
Normal
Neck pain, weakness in both arms, fever and chills Wright test: 1/1280
S, streptomycin; T, tetracycline; R, rifampin; (T-S), trimethoprim-sulfomethaxazole; C, ciprofloxacine.
Low back pain and Low back Low back pain in left leg pain, fever pain and diffuse and chills arthralgia Wright test: 1/160 Wright test: 1/320 Wright test: 1/320 + Strength in dorsal Normal Normal flexors in left leg 4/5
L4 -L5
L5-S1
L3-L4
L5
L5-S1
A. Go¨rgu¨lu¨ et al. / Surgical Neurology 66 (2006) 141 – 147
Patient
A. Go¨rgu¨lu¨ et al. / Surgical Neurology 66 (2006) 141 – 147
143
Fig. 1. T1-weighted MR images showing cervical epidural abscess (black arrow) extending from C5-9. A: Pretreatment. B: Posttreatment.
of Brucellar SEA retrospectively. Diagnosis of SEA was established by MRI and serologic tests specific for Brucella species (Wright test N 1:160). Antibiotic therapy was started after the confirmation of diagnosis by serologic test. All of these patients had been on antibiotic regimens for a period of 6 to 12 weeks. Etiologic microorganism was cultured from abscess material in 1 patient who subsequently underwent surgical drainage. Patients were followed up by regular neurologic examinations, MRI, and serologic tests. Mean follow-up period was 20 months (6-30 months). 3. Results Patient characteristics, presenting symptoms, diagnostic blood tests, initial neurologic examination, location of abscesses, duration of antibiotic therapy, and antibiotic regimen of each patient are shown in Table 1. The duration of symptoms before hospitalization ranged from 15 days to 4 months (median, 2 months). All patients suffered from
pain localized over the involved vertebrae. Mild motor deficits were detected in 3 patients, whereas 6 patients had no deficit. Initial diagnosis was erroneous in 3 patients with low back pain who had been treated medically for lumbar disk herniation before they were admitted to our clinic. Fever and chills were encountered in 2 patients. The presence of abscesses was confirmed with MRI in all patients (Figs. 1-4). The abscesses were located in the lumbar spine (6 patients), followed by the thoracal spine (2 patients), and the cervical spine (1 patient). Radiological findings of spondylodiscitis were found in all patients but the one with cervical involvement. Results of the standard tube agglutination test for Brucella (Wright test) were more than 1:160 in all patients. Peripheral WBC counts were determined to be mildly elevated (between 12 000 and 14 000 cells per microliter) only in 6 patients. C-reactive protein was found high in 7 patients (mean CRP level was 13.0 F 8 mg/dL). Sedimentation rate was above normal in 3 patients. Blood
Fig. 2. Axial MR images showing epidural abscess located (black arrow) in dorsal region. A and B: Pretreatment.
144
A. Go¨rgu¨lu¨ et al. / Surgical Neurology 66 (2006) 141 – 147
images disappeared completely in MRI with Gadolinium (2 -10 weeks; mean, 4.2 weeks), and they recovered completely without neurologic deficits (Figs. 1- 4). Pretreatment hypointense and hyperintense appearance of vertebral bodies in T1- and T2-weighted MRI scans, respectively, became isointense in both sequences of posttreatment MRI. Peripheral WBC count and CRP returned to normal ranges within 3 to 7 days, although sedimentation rate normalized later (around 3 weeks). Titers of specific antibodies to Brucella have shown progressive decrease, and they returned to normal values for more than a year in 3 patients (mean duration, 6 months). 4. Discussion
Fig. 3. T1-weighted sagittal MRI scans. Abscess formation (black arrow) extends along the length of 3 thoracic vertebrae. A: Pretreatment. B: Posttreatment.
cultures were negative in all patients but 1 that yielded Brucella abortus (patient 3). Considerable decrease in the size of the abscesses (mean decrease, 30%) were observed 2 to 5 weeks (mean, 3 weeks) after antibiotic therapy. Complaints regressed completely in all patients but 1 (patient no. 9) who suffered from progressive sciatalgia and developed weakness in plantar flexion. Abscess was drained through hemilaminectomy in this patient. Severe lumbar pain was dissolved quickly after surgical drainage, and he was neurologically normal in third month. The rest of the patients was treated with antibiotics for an average of 8.5 F 0.5 weeks. Abscess
Brucella is transmitted from animals to humans via 3 main routes: the digestive tract (eating unpasteurized milk or contaminated meat), by direct contact with infected tissues (blood, lymph, conjunctiva, or broken skin), or by the respiratory system (inhalation by laboratory personnel of the live bacteria or clinical isolates) [21,23]. Regarding risk factors of our patients, 1 was a veterinarian, 3 were involved with stock-breeding, and 3 were infected by ingestion of unpasteurized milk products. The primary site of localized infection is mostly seen in the musculoskeletal system [22,23]. The spinal involvement of brucellosis is generally seen in the subacute chronic phases. The age of patients with vertebral involvement was more than 50 years in the literature, as in our series [3,8]. Brucella spondylodiscitis generally arises from upper end plate where blood supply is rich, and it may spread to whole vertebra, disk space, and adjacent vertebra, depending on the virulence of the microorganism, size of the inoculum, and immune status of the host. Therefore, Brucellar SEA mostly results from
Fig. 4. T2-weighted axial MR images showing abscess formation (black arrow) in a lumbar segment. A: Pretreatment. B: Posttreatment.
A. Go¨rgu¨lu¨ et al. / Surgical Neurology 66 (2006) 141 – 147
direct invasion followed by spondylitis [7]. However, SEA may develop by direct hematogenous route very rarely without spondylitis. One of our patient (patient no. 3) was shown to have SEA without any accompanying pathological change in the bony structure of the corpus with MRI; however, clinical progress and response to antibiotics were not different from the other patients. The lomber region is the most common site for abscess formation [3,8], whereas cervical region is a rare location as in one of our patients [3,8,17]. Spinal epidural abscess is primarily a bacterial infection. Staphylococcus aureus is the bacterium most frequently isolated, followed by various aerobic and anaerobic microorganisms, Mycobacterium tuberculosis, fungi, and parasites [2,19,25]. Brucella sp. are among the very rarely encountered causes of SEA. The classic triad of fever, pain, and neurologic deficits is seen in patients with SEA [2,10,18,19]. However, because of the subacute chronic course of Brucellar SEA, findings of a systemic infection (malaise, chills, and periodic nocturnal fever [undulant fever]) is not commonly encountered (2 of our patients suffered from systemic complaints). Rarely, it can mimic a herniated disk with acute radicular symptoms [14,15]. Pain with palpation and local tenderness are the most common but nonspecific findings. However, these nonspecific but localizing findings may be more valuable and guiding for those who are living in endemic regions or for those who traveled to these regions. Magnetic resonance imaging is considered to be the gold standard for the diagnosis of SEA. Meanwhile, this modality is also useful in revealing the spondylodiscitis, medulla, root compression, and surrounding anatomical structures. Epidural abscess is seen as a mass lesion within the spinal canal and outside to the spinal cord, hyperintense on T2-weighted, and slightly hypointense on T1-weighted magnetic resonance (MR) images. Abscess is better delineated with contrast enhancement with Gd-DTPA [7,22]. There is no particular finding in MRI for Brucellar spinal infections. Two forms of spinal brucellosis are known, focal, and diffuse. Focal Brucellar osteomyelitis appears as focal areas of abnormal signal intensity and usually localized in the anterior aspect of an end plate of a vertebra at the discovertebral junction. Diffuse form of Brucellar osteomyelitis is defined as diffuse abnormal signal intensity in the adjacent vertebra and the intervening disk. Computed tomography and radionuclide bone scanning may be applicable and useful for the detection of bony changes in the spine [21,23]. The polymerase chain reaction offers another sensitive and specific way of detecting Brucellar DNA in clinical samples. The diagnostic application of polymerase chain reaction in brucellosis may be a good alternative compared with conventional methods by providing results in a short period; however, its disadvantages such as low detection limit of bacterial DNA in samples increased risk of carry-over contamination, and higher cost must be further
145
evaluated. Although the progress to date is promising, there is still room for improvement especially with neurobrucellosis. The definite diagnosis of brucellosis is made when Brucella is isolated from blood, bone marrow, or other tissues. Nevertheless, isolation of the pathogen may be difficult because of the intracellular localization of the bacterium and chronic nature of the disease [21,22]. Although Brucella grows slowly, automated culture systems (eg, BACTEC, BACT/ALERT) have shortened the isolation time from clinical sample including CSF to days. WBC count is generally normal, and routine laboratory tests are often not helpful. Rose-Bengal shows rapid agglutination, and it can be used as a screening test [17,22]. When brucellosis is suspected, detection of high or rising titers of specific antibodies to Brucella ( z 1:160, standard tube agglutination test or z 1:320 [Coombs’ test]) in the serum helps a presumptive diagnosis [21,22]. The sensitivity and specificity of the these tests are very high (around 98% for both). Diagnosis was done by agglutination tests in all of our patients, and blood cultures were positive in only 1 patient. It is reported in the literature that successful treatment modality of the patients with SEA is aggressive surgical debridement [13,20]. There is a consensus that nonsurgical treatment can be reserved for limited number of patients who carry considerable surgical risk, and for those who have no neural tissue damage or who have a complete paralysis lasting for more than 3 days [12,18]. Because Brucellar SEA mostly develops in the subacute-chronic phase of the disease process and Brucella has a good antibiotic response, recently, many authors report in the favor of antibiotic therapy, which should be administered beforehand in required doses and last until full recovery [3,4,9]. Therapeutic efficacy of antibiotic therapy in patients with Brucellar SEAs was reported by various authors as case reports in the literature [4,9,19]. Bingol et al [4] reported successful medical treatment of spastic paraparetic patients with intramedullary granuloma due to Brucella species. To the best of our knowledge, the present series appears to be the largest patient group reported regarding Brucellar SEA in the literature so far, and we achieved complete cure both clinically and radiologically by medical treatment alone in all cases except 1. An emergent decompression may be mandatory during the course of antibiotic therapy if neurologic deterioration occurs, as was the case with one of our patients. Currently, there is still no consensus in choosing the appropriate antibiotics and the treatment duration in patients with Brucellar spondylitis and SEA [16,17,22]. However, there is a general agreement that monotherapy in short-term usage is associated with a high relapse rate (30%) [4]. The most recent reports have recommended either cotrimoxazole and/or streptomycin and/or rifampin and doxycycline in treatment [1,8]. Streptomycine is not to be given for more than 3 weeks. Quinolones may also be added to the drug regimens [17,22]. These drugs may be
146
A. Go¨rgu¨lu¨ et al. / Surgical Neurology 66 (2006) 141 – 147
synergistic, and they have good intracellular and central nervous system penetration [5]. It is also recommended that antibiotic therapy should be prolonged up to 6 weeks to 1 year in patients with spondylitis and SEA [2,3,16,17]. However, relapses generally occur within the 6-month period after the completion of therapy [4,21]. We administered similar treatment protocols to our patients for a duration of 6 to 12 weeks, and we did not experience any relapse in long follow-up periods (mean, 20 months). We determined the criteria for terminating antiobiotherapy as clinical recovery and considerable decrease in abscess images radiologically. Although all of our patients recovered fully, it took more than a year for some to turn their serologic tests to normal. Therefore, opposite to the idea of some authors, we think that conversion of serologic tests to normal should not be assumed as a recovery criterion in patients with Brucellar SEA. Finally, it seems logical to consider the probability of Brucella as an etiologic agent in patients with SEA who are the inhabitants of endemic regions and who traveled to these regions. Antibiotic therapy could be primary treatment and is effective in many patients. However, should anticipated recovery not be achieved or neurologic deterioration be observed despite proper antibiotic regimens in required doses and duration, surgical decompression is to be performed. Acknowledgments The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. References [1] Al-Eissa YA. Clinical and the rapeutic features of childhood neurobrucellosis. Scand J Infect Dis 1995;27:339 - 43. [2] Akalan N, Ozgen T. Infection as a cause of spinal cord compression: a review of 36 spinal epidural abscess cases. Acta Neurochir 2000; 142:17 - 23. [3] Ariza J, Gudiol F, Valverde J, et al. Brucellar spondylitis: a detailed analysis based on current findings. Rev Infect Dis 1985;7:656 - 64. [4] Bingol A, Yucemen N, Meco O. Medically treated intraspinal bBrucellaQ granuloma. Surg Neurol 1999;52:570 - 6. [5] Bucher A, Gaustad P, Pape E. Chronic neurobrucellosis due to Brucella melitensis. Scand J Infect Dis 1990;22:223 - 6. [6] Cetin ET, Coral B, Bilgic¸ A, et al. Tqrkiye’de insanda bruselloz insidansinin saptanmasi. Doga Tr J Med Sci 1990;14:324 - 34. [7] Ceviker N, Baykaner K, Goksel M, et al. Spinal cord compression due to Brucella granuloma. Infection 1989;17:304 - 5. [8] Colmenero JD, Reguera JM, Martos F, et al. Complications associated with Brucella melitensis infection: a study of 530 cases. Medicine 1996;75:195 - 211. [9] Geijo Martinez P, Perez Gil MA, Ruiz Ribo D, et al. Brucella spondylitis with epidural abscess healed by drug treatment. Apropos of a case diagnosed by nuclear magnetic resonance. An Med Interna 1996;13:203 - 4.
[10] Khan SH, Hussain MS, Griebel RW, et al. Title comparison of primary and secondary spinal epidural abscesses: a retrospective analysis of 29 cases. Surg Neurol 2003;59:28 - 33. [11] Lifeso RM, Harder E, McCorkell SJ. Spinal brucellosis. J Bone Joint Surg Br 1985;67:345 - 51. [12] Naderi S, Yuceer N, Mertol T, et al. Course and prognosis of spinal epidural abscess. Report of five cases. Eur J Orthop Surg Traumatol 2000;10:199 - 202. [13] Nussbaum ES, Rigamonti D, Standiford H, et al. Spinal epidural abscess: a report of 40 cases and review. Surg Neurol 1992;38: 225 - 31. [14] Ozates M, Ozkan U, Bukte Y, et al. Lumbar epidural brucellar abscess causing nerve root compression. Spinal Cord 1999;37:448 - 9. [15] Ozerbil OM, Ural O, Topatan HI, et al. Lumbar spinal root compression caused by Brucella granuloma. Spine 1998;23:491 - 3. [16] Pina MA, Ara JR, Modrego PJ, et al. Brucellar spinal epidural abscess. Eur J Neurol 1999;6:87 - 9. [17] Pina MA, Modrego PJ, Uroz JJ, et al. Brucellar spinal epidural abscess of cervical location: report of four cases. Eur Neurol 2001;45:249 - 53. [18] Reihsaus E, Waldbaur H, Seeling W. Spinal epidural abscess: a metaanalysis of 915 patients. Neurosurg Rev 2000;23:175 - 204. [19] Rigamonti D, Liem L, Sampath P, et al. Spinal epidural abscess: contemporary trends in etiology, evaluation, and management. Surg Neurol 1999;52:189 - 96. [20] Soehle M, Wallenfang T. Spinal epidural abscesses: clinical manifestations, rognostic factors, and outcomes. Neurosurgery 2002;51: 79 - 85. [21] Solera J, Lozano E, Martinez-Alfaro E, et al. Brucellar spondylitis: review of 35 cases and literature survey. Clin Infect Dis 1999;29: 1440 - 9. [22] Tekkok IH, Berker M, Ozcan OE, et al. Brucellosis of the spine. Neurosurgery 1993;33:838 - 44. [23] Trujillo IZ, Zavala AN, Caceres JG, et al. Brucellosis. Infect Dis Clin North Am 1994;8:225 - 41. [24] Yilmaz M, Ozaras R, Ozturk R, et al. Epileptic seizure: an atypical presentation in an adolescent boy with neurobrucellosis. Scand J Infect Dis 2002;34:623 - 5. [25] Young EJ. Brucella species. In: Mandell GL, Bennett JE, Dolin R, editors. Principles and practice of infectious diseases. Philadelphia7 Churchill Livingstone; 2000. p. 2386 - 93.
Commentary Brucellosis is a worldwide zoonotic disease infecting humans, domestic, and feral animals and marine mammals. It is commonly present in the Mediterranean basin, Middle East, and Latin America. Between 2% and 53% of patients with brucellosis develop spondylodiscitis, osteomyelitis, and arthritis. Since about 500 000 cases of brucellosis occur annually worldwide, spondylodiscitis and secondary epidural abscess must occur more often than recognized. The study by Gfrgqlq et al describes the vague symptoms that precede and accompany the patient’s presentation as well as the difficulty in obtaining positive cultures from sterile sites but, at the same, time emphasizes the value of MRI and agglutinating antibodies to Brucella in establishing a diagnosis. Serum agglutinating antibody test or the enzyme-linked immunosorbent assay are 2 serologic techniques used to confirm a diagnosis of brucellosis, although most recently, the use of a polymerase chain reaction appears to be more sensitive than culture in
