Original Paper Eur Neurol 1992;32:328-333
Department of Neurology. Showa University School of Medicine, Tokyo, Japan
KeyWords Spinal epidural abscess Spinal leptomeningitis (spinal subarachnoid abscess) Myelography Magnetic resonance imaging Spinal cord infarction Staphylococcus aureus Escherichia coli
Acute Spinal Epidural Abscess and Spinal Leptomeningitis: Report of 2 Cases with Comparative IMeuroradiological and Autopsy Study Abstract A 21-year-old male developed back pain, fever, and rapidly progressive quadriparesis. Lumbar tap yielded frank pus which was confirmed on magnetic resonance imaging (MKI) to be located mainly in the cervical epidural space. Conservative antibiotic remedy was partially effective for restoration of the neurological deficits. A 82-year-old female noticed low-back pain which was rapidly accompanied with clouding of consciousness, paraplegia, and sphinc ter disturbances. Lumbar puncture revealed thick pus which was best depicted on MRI in the thoracolumbar subarachnoid space. At autopsy, spinal sub arachnoid abscess or leptomeningitis was confirmed, and a spinal infarction previously unrecognized on MRI was found. Usefulness and shortcomings of MRI in the diagnosis of paraspinal infections are discussed.
Introduction Acute spinal epidural and subarachnoid abscesses, otherwise termed spinal leptomeningitis, are rarely en countered in neurological practice and often escape a cor rect diagnosis, resulting in permanent neurological defi cits. Both initially present with back pain, high fever, and nerve root pain followed by rapidly progressing para- or tetraparesis and sphincter disturbances in the more ad vanced stage and possible death [1-5]. As is often the case, clinicians face the difficulty in differentiating the two similar but different diseases just by the clinical signs [4, 5]. Although the early utilization of magnetic reso nance imaging (MRI) is advocated for correct diagnosis [6, 7]. there have been few studies comparing MRI with other radiological methods in these two paraspinal infec tious states. The purpose of this report is to present cases
Received: November 14. 1991 Accepted after revision: June 16. 1992
of a spinal epidural and a subarachnoid abscess and to discuss the usefulness and limitations of MRI as a diag nostic tool.
Case Reports Patient I A 2 1-year-old male, with a previous history of cellulitis in the left leg 6 months before, suddenly developed progressively worsening low-back pain and fever on October 4. 1990. and subsequently was admitted to a local hospital. An epidural block was successful in con trolling the back pain, but on the following morning, the patient noticed severe nuchal pain that limited neck movements and numb ness of ankles and feet. As lumbar puncture revealed gray-yellowish pus from which Staphylococcus aureus was islolated. a tentative diag nosis of suppurative myelitis was established. Although intravenous ampicillin sodium was started immediately, the patient continued to have high fever and rapidly developed quadriparesis. numbness
Toshiya Fukui. MD Department of Neurology Showa University School of Medicine 1-5-8 Hatanodai. Shinagawa-ku Tokvo 142 (Jaoan)
© 1992 S. Karger AG. Basel 0014-3022/92/0326-0328 $2.75/0
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
Toshiya Fukui Hiroo Ichikawa Nobuyuki Kawate Taneyoshi Nozawa Koujiro Sugita
below the sensory level C-5 dermatome, and urinary retention. At this point he was transferred to our hospital for further investigation and treatment. Physical examination on admission revealed an acutely ill obscse male with a body temperature of 37.9 °C and severe spontaneous pain at spinous processes of C-6 and 7, with marked tenderness. There was no evidence of concurrent dermal, dental, or rhino-oph thalmologic infections. Neurological examination revealed that the patient was alert and oriented. The cranial nerves were unremark able. He showed flaccid tetraparesis, more prominent in the lower as well as the left extremities, total areflexia with positive bilateral Babinski signs, hypocsthesia of all modalities at C-5 level extending caudally. and bowel and bladder disturbances. Acute inflammatory reactions such as leukocytosis ( 13.400/mm3), increased erythrocyte sedimentation rate (120mm/h). and strongly positive C-reactive protein (25.5 mg/dl) were noted on admission. Lumbar tap at L-4/5 again produced purulent fluid with countless numbers of polymorphonuclear leukocytes, protein 4.058 mg/dl. and glucose 7 mg/dl. with the culture growing S. aureus. The roentgenograms and nonenhanced computerized tomograms (CT) of the cervical spinal column were unremarkable, and there was no sign of osteomyelitis. Myelography or myelo-CT was not per formed because it was considered to be inappropriate when the pres ence of an acute paraspinal infection w'as highly suspected. Sagittal cervical MRI (Toshiba MRT-50A. 0.5 T) revealed a sharply delineated mass of increased signal intensity with gadopcntctate dimegluminc (Gd-DTPA) injection on Tl-weighted image, lo cated around and mainly dorsal to the spinal cord. The lesion was situated in the epidural space between C-l and C-7, without invasion into the intracranium. Thoracic and lumbar MRI records were too poor, due to motion artifacts, for the lower end of the abscess to be evaluated. The epidural fat tissue, which would appear as a highintensity band in epidural space, could not be isolated from the abscess in the low'er cervical region (fig. I A). T2-weighted scans were not available. Based on the clinical and laboratory manifestations, he was diagnosed as having an acute spinal epidural abscess. The patient was soon responding to intensive intravenous antibi otic therapy with piperacillin sodium while waiting for surgical inter vention, which turned out to be unnecessary once quick and promis ing recovery' commenced. The abscess had disappeared on follow-up MRI I month after the onset of the symptoms (fig. IB) w'hen the patient did not have fever, nuchal pain, or paresis of the upper extremities and was capable of ambulating with assistance, albeit he had remaining spastic paraparesis and moderate sensory deficits in the legs.
Fig. 1. Patient l. With Tl-weighted (spin-echo technique, TR 300 ms. TE 14 ms) sagittal MRI with Gd-DTPA injection, the abscess is depicted in the epidural space as a sharply demarcated linear lesion of high signal intensity (arrows in A). The lesion is not seen after successful antibiotic treatment (B). The high-intensity area anterior to the spinal cord suggests the presence of a small amount of pus (A).
Patient 2
Fig. 2. Patient 2. Lumbar myelography shows a partial block of the subarachnoid space at T-9 and L-l due to an extramedullary lesion. It is inconclusive whether the lesion is located intradural or not.
329
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
An 82-ycar-old female, without a previous medical history of dia betes mcllilus or any suppurative diseases, noticed vague low-back pain while cleaning her house on December 31. 1989. The next morni ng, her back pain was also accompanied by weakness of the legs and fever of 38.0 °C. The paraparesis of the lower extremities rapidlyworsened. followed by mild clouding of consciousness, total loss of all sensations in the lower extremities, and loss of bladder and rectal control within days. She was admitted to a hospital with a suspicion of acute meningomvclitis. Myelography performed there showed a partial block of the spinal column between T-9 and L-l by an extra medullary lesion which was either intra- or extradural (fig. 2). On myelo-CT. the spinal cord appeared to be compressed anteriorly by localized masses which projected into the enhanced subarachnoid
A
B
Fig. 4. Patient 2. A Tl-weighted (spin-echo technique. TR 300 ms. TE 14 ms) sagittal MR1 (A) and a transaxial image at T-10 (B) with Gd-DTPA. The lesion (arrows in A) is located dorsal to the spi nal cord and anterior to the epidural fat from T-9 through L-l, but the lesion seems to extend further rostrally and caudally. A mass lesion of low signal intensity which is surrounded by a high-intensity band (black arrow in B) is present behind the spinal cord (asterisk in B). Epidural fat tissue (white arrow in B) can be recognized clearly separate from the mass lesion. The band of low signal intensity between them represents dura mater and arachnoid membrane.
330
Fukui/ichikaw'a/Kawate/Nozaw'a/Sugiia
Neuroradiological Study in Paraspinal Infections
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
Fig. 3. Patient 2. Myelo-CT at L-l shows a few extramedullary masses in the subarachnoid space. An outer band of low density cor responding to the epidural fat tissue is clearly recognized.
space in the same segment as shown on myelography (fig. 3). She was then on ampicillin sodium for a few days which did not bring about any amelioration of the symptoms. When transferred to our hospital on the 9th day. she was febrile (38.6 °C) and slightly disoriented, complaining of nausea and head ache with meningeal irritation signs, but was devoid of spinal percus sion tenderness or any other noticeable skin infection. Neurological examination revealed that the patient had total paraplegia in the legs, all modal sensory disturbances of a partial degree at L-l to L-3 and total sensory loss below L-3, brisker but equal tendon reflexes in the upper extremities, areflexia in the legs with positive bilateral Babinski signs, and sphincter disturbances. However, she did not show any cranial nerve palsy, weakness, or cerebellar ataxia of the upper extremities. Acute inflammatory signs such as leukocytosis (20,400/mm3), increased erythrocyte sedimentation rate (30 mrn/h). and positive Creactive protein (1.3 mg/dl) were remarkable among the laboratory Endings on admission. Urine culture grew Candida species under use of ampicillin sodium. A lumbar tap yielded purulent fluid which con tained more than 170.000 cells/mm1. 97% of which polymorphonu clear leukocytes, 5,000 mg/dl protein. 44 mg/dl glucose, and culture of the fluid grew Escherichia coii. The roentgenograms of the vertebral column showed no sign of osteomyelitis. Tl-weighted sagittal MR1 with Gd-DTPA of the tho racic region disclosed an ill-defined mass of the same or slightly higher signal intensity as the spinal cord, displacing it anteriorly from the level of T-7 at least to L-2. The lesion was recognized to have much more extensive rostral-caudal distribution than was inferred from myelography and myclo-CT (fig. 4A). The transaxial T lweighted MRI scan at T-10 level showed more or less the same find ings as myelo-CT. i.e.. the spinal cord was anteriorly dislocated by a few blocks of mass lesions of the same intensity as the spinal cord, some of which appeared to be encapsulated by a membrane-like structure enhanced by GD-DTPA. apparently inside a circular band of low intensity which would correspond to dura mater and arach noid membrane. Epidural fat was conspicuously depicted outside the dural sac and characterized by its markedly high signal intensity (fig. 4B). Combined with lumbar tap results, these mass lesions were regarded as subarachnoid pus. No abnormalities were detected in the spinal cord parenchyma on Tl-weighted image with Gd-DTPA. and a T2-weighted scan was not available. From these observations, the patient was diagnosed as having an acute spinal subarachnoid abscess or spinal leptomeningitis, caused by E. coli at the lower thoracic and upper lumbar regions. The pres ence of clouding of consciousness suggested concomitant purulent meningitis as a result of extension of the inflammatory process into the intracranium as inferred from the MRI findings, although mycloCT suggested otherwise. Candida in the urinary tract suggested prior infection by suppurative bacteria including E. coli which may have been adventitiously removed by antibiotic therapy. Because of her high age. the severity of paraplegia, and a poor nutritional state, she was rendered to be a poor candidate for surgery' and was immediately placed on piperacillin sodium and latamoxef sodium in place of ampicillin sodium. This regimen was effective for her meningitis and back pain, but no neurological improvement in myelopathy was noted. She was on the same regimen w'ithout any neurological improvement and expired on the 21st day because of concomitant cerebral hemorrhage and hematemesis from disseminated intravas cular coagulopathy.
At autopsy, no purulent exudate was encountered on the surface of the spinal dura mater, but once it was opened, there was sticky purulent material covering the spinal cord from C-2 to cauda equina, the thickest pus being present at T-9 to L-3. There was no gross com pression or deformity of the spinal cord. Microscopically, the puru lent material was confirmed as an aggregation of polymorphonuclear cells confined to the subarachnoid space. The spinal cord per sc did not show any infiltration with acute inflammatory cells in its paren chyma. Abscess masses had no encapsulating membrane against the inference from MRI findings. The most striking pathological change was a spinal cord infarction most prominently seen at T-12 and L-l (fig. 5). The whole gray matter and patchy areas of the white matter at T-12 (fig. 5A) and the anterior two thirds of the spinal cord at L-l (fig. 5B) were completely infarcted and microscopically infiltrated by macrophages, the presence of which suggested that the infarction was in a subacute to early chronic stage. A mild ischemic change was also noted in anterior and lateral columns between T-10 and T-12. The anterior spinal artery was patent, and there was no engorgement of the spinal venous plexus.
Discussion
Fig. 5. Patient 2. Mainly the gray matter and patchy areas of the white matter at T-12 (A) and an extensive area of the ventral two thirds of the spinal cord at L-l (B) are grossly infarcted. These areas may correspond to the territories irrigated by the anterior spinal artery. Note a large amount of abscess underneath the arachnoid membrane (arrowheads) and the absence of any direct compression to the spinal cord. Kliiver-Barrera. X 8.
fat, dura mater, arachnoid membrane, spinal cord, and abscess if any. If an abscess is to be distinguished from the normal structures in the spinal canal, it would necessitate intrathecal injection of contrast medium; therefore, myelo-CT would carrv the same risks as mvelography [ 6], MRI is being recognized as a noninvasive, repeatable, and potentially capable examination tool for delineating the location and the extent of an epidural abscess [2, 3, 6, 7], The MRI presentation of an epidural abscess in patient 1 may be characterized by a sharp margin, inabil ity of the abscess to extend into the intracranium, and dis appearance of normally bright high intensity of epidural fatty tissue. The fact that a cervical abscess has sharp
331
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
As Heusner [8] and others [4, 5. 9] described, the clini cal features of both epidural and subarachnoid abscesses follow similar courses and are characterized by back pain, radicular pain, limb paresis, and rapidly progressing quadriplegia. Therefore, these two entities are essentially indistinguishable from each other on clinical grounds. It is discussed that coexisting cerebral meningitis, ventricu litis, and spinal vascular lesions are more common in subarachnoid abscesses and that spinal percussion tender ness is more common in epidural abscesses [4. 10], but these are not absolute distinctive differences [11], Al though some authors maintain that the distinction be tween the two is of little practical significance, since surgi cal drainage and antibiotic treatment are all-important in both situations [12], it would be of neuroradiological interest as well as an adjunct help in selecting candidacy and timing for surgical treatment to know the exact loca tion and the extent of a lesion. Myelography is in some cases a successful method in distinguishing an epidural from a subarachnoid abscess [ 13.14], but in others it does not suffice [ 12, 15, 16], Some potential risk is also suggested by implanting pathogenic organisms into the subarachnoid space from an epidural abscess via spinal tapping [16]. Since patient 1 had an accelerated exacerbation of neurological symptoms after epidural block to control the back pain, it is possible that an abscess originally localized in the lumbar region might have been disseminated into the cervical area. With noninvasive plain CT. it is impossible to distin guish the soft tissues within the spinal canal, i.e., epidural
332
abscess per se. As the lower thoracic to upper lumbar spi nal cord is vulnerable to ischemia because this region cor responds to a borderzone between two adjacent radicular arteries [20, 21 ]. it is possible that radicular arteries were compressed by abscess or that arterial spasms were in duced by the irritative nature of infection, resulting in spi nal infarction at T-12 and L-l. It is uncertain why the spinal cord infarction escaped detection by the most sensitive diagnostic tool of MRI. but presumably may it be either because spinal infarction at a specific stage of disease progression may appear isosignaled on Tl-weighted image as noninjured normal spi nal parenchyma or because the spinal cord imaging in the thoracolumbar region is apt to be obscured by artifacts from surrounding structures. For more accurate diagnosis of a spinal infarction, spinal imaging not only on Tlweighted but also on T2-weightcd MRI setting would be recommended. Early detection of spinal infarction which may conse quently accompany spinal leptomeningitis would be of great importance in determining a conservative treatment strategy in lieu of surgical intervention, since decompres sion and removal of abscess may not be effective or suit able in reducing neurological impairments of those af flicted with this ailment once accompanied by spinal infarction. According to the previous literature [4, 5, II. 12. 19], the most common causative organism in both spinal epi dural and subarachnoid abscess is S. aureus. Patient 1 with an epidural abscess was not an exception, but patient 2 was one of the rare cases whose spinal leptomeningitis was caused by an organism other than this common bac terium. E. coli is regarded as a rare cause of paraspinal infections and shares only 2% of all causative organisms in a series of patients reviewed by Takenaka et al. [11], In either case, it was impossible to ascertain the exact entry route of infection to the paraspinal regions, but it was speculated that hematogenic dissemination of a causative organism from the respiratory or urinary tract, for in stance. might be more probable than direct extension of infection from nearby structures to the spinal canal.
Acknowledgements The authors would like to express sincere gratitude to Dr. Nobutaka Arai for providing pathological findings of the spinal cord and to Ms. Ramona Severson for syntactic revision of the manuscript.
Fukui/Ichikawa/Kawate/Nozawa/Sugita
Neuroradiological Study in Paraspinal Infections
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
boundaries and that it does not extend into the intracra nium. even though it is closely adjacent to the foramen magnum, would suggest that the abscess was located in the tight and closed epidural space which exists between the spinal dura mater proper and the periosteum of spi nous processes, both being joined to form one sheet of intracranial dura mater. Since the disappearance of epi dural fat on CT is regarded as a pathognomonic feature in epidural abscess [17], and whether or not the abscess located in the epidural space can be identified on myeloCT with ease, transaxial MRI seems to have no additional diagnostic advantage over myelo-CT except for its low ered risk due to the absence of contrast medium. Sagittal MRI. however, would be far more superior to myelo-CT in depicting the longitudinal extent of the disease. Spinal subarachnoid abscess or leptomeningitis, on the other hand, is a much rarer disease and considered to be one of the subtypes of subdural abscess. Although there have been few reports on the MRI findings in subarach noid abscess, Kurokawa ct al. [ 18] recently reported a case of spinal 'subdural' abscess which is depicted 'intrathecally’ as an isointensity on T1 -weighted and as a strongly high-signaled band on T2-weighted scans. MRI findings of the subarachnoid abscess in patient 2 are remarkable in that the demarcation of the abscess is more ill-defined, that it is easy to identify epidural fat apart from an abscess, and that the rostral-caudal extent of the abscess is more precisely correlated with autopsy findings than myelo-CT. As mentioned above, transaxial myelo-CT may be as efficacious as MRI in defining size, configuration, and location of an abscess in relation to dura mater and spinal cord; the superiority of MRI to myelo-CT in the diagnosis of spinal subarachnoid ab scess. therefore, may exist in its availability of diagnosti cally valuable sagittal scans as well as its being noninvasive and easily repeatable. In this case, enhancement by Gd-DTPA was not much help in separating the lesion from the normal structures, and also careful interpreta tion of ‘ringed enhancement' is needed because the au topsy examination did not prove any capsular membrane around abscess masses. Without similar cases, it remains uncertain as to what it may represent. Paraplegia and sensory loss in the lower extremities observed in patient 2 were most likely to be ascribed to spinal infarction at T-12 and L-l rather than to nonsegmental and diffuse compression, if any, of the spinal cord by abscess. This may quite well coincide with the estima tion of Freedman and Alpers [19] that spinal vascular lesion is more responsible for the presentation of neuro logical symptoms than the compression of spinal cord by
References 8 Heusner AP: Nontubcrculous spinal epidural infections. N Engl J Med 1948;239:845-854. 9 Hirson C: Spinal subdural abscess. Lancet 1965;ii: 1205—1217. 10 Dus V: Spinal peripachymeningitis (epidural abscess). J Neurosurg 1960:17:972-983. 11 Takcnaka K, Kobayashi H. Niikawa S. Hattori T. Ookuma S. Nokura H. Sakai N. Yamada H. Sasaoka I: Spinal subdural abscess. Report o f a case and a review of the literature of 43 cases. Brain Nerve (Tokyo) 1989:41:331-336. 12 Schiller E, Shadlc OW: Extrathecal and intra thecal suppuration. Arch Neurol 1962:7:33— 36. 13 Patronus NJ, Marx WJ, Duda EE: Radiographic presentation of spinal abscess in the subdural space. AJR 1979:132:138-139. 14 Danner RL. Hartman B.I: Update of spinal epi dural abscess: 35 cases and review of the litera ture. Rev Infect Dis 1987:9:265-274. 15 Messer HD, LenchnerGS. Brust JCM. RcsorS: Lumbar spinal abscess managed conservative ly. J Neurosurg 1977:46:825-829.
16 Knudsen LL. Voldby B. Stagaard M: Com puted tomographic myelography in spinal sub dural empyema. Neuroradiology 1987:29:99. 17 Leys D. Lesoin F. Viaud C. Pasquier F. Rous seaux M. Jomin M. Petit H: Decreased morbid ity from acute bacterial spinal epidural abscess using computed tomography and nonsurgical treatment in selected patients. Ann Neurol 1985:17:350-355. 18 Kurokawa Y. I lashi K. Fujishigc M. Tsuchida M. Maeda K, Kaneko M: Spinal subdural em pyema diagnosed by MRI and recovered by conservative treatment. Brain Nerve (Tokyo) 1989:41:513-517. 19 Freedman H. Alpers BJ: Spinal subdural ab scess. Arch Neurol Psychiatry 1948:60:49-60. 20 Lazorthcs G. Gouaze A, Zadeh JO. Santini JJ. Lazorthes Y. Burdin P: Arterial vascularization of the spinal cord. Recent studies of the ana tomic substitution pathways. J Neurosurg 1971:35:253-261. 21 Turnbull IM: Microvasculature of the human spinal cord. J Neurosurg 1971:35:141-147.
333
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
1 Baker AS. Ojemann RG. Swartz MN. Richard son EP Jr: Spinal epidural abscess. N Engl J Med 1975:293:463-468. 2 Hlavin ML. Kaminski HJ. Ross JS. Ganz E: Spinal epidural abscess: A ten-year perspective. Neurosurgery 1990:27:177-184. 3 Curling OD. Gower DJ. McWhorter JM: Changing concept in spinal epidural abscess: A report of 29 cases. Neurosurgery 1990:27:185— 192. 4 Fraser RAR. Ratzan K. Wolpert SM. Wein stein L: Spinal subdural empyema. Arch Neu rol 1973:28:235-238. 5 Dacey RG. Winn HR. Jane JA. Butler AB: Spi nal subdural empyema: Report of two cases. Neurosurgery 1978:3:400-403. 6 Erntell M. Hollas S. Norlin K. Dahlquist E. Nilsson-Ehle I: Magnetic resonance imaging in the diagnosis of spinal epidural abscess. Scand J Infect Dis 1988:20:323-327. 7 Hanigan WC. Asner NG. Elwood PW: Mag netic resonance imaging and the nonoperative treatment of spinal epidural abscess. Surg Neu rol 1990:34:408-413.
Department of Neurology. Showa University School of Medicine, Tokyo, Japan
KeyWords Spinal epidural abscess Spinal leptomeningitis (spinal subarachnoid abscess) Myelography Magnetic resonance imaging Spinal cord infarction Staphylococcus aureus Escherichia coli
Acute Spinal Epidural Abscess and Spinal Leptomeningitis: Report of 2 Cases with Comparative IMeuroradiological and Autopsy Study Abstract A 21-year-old male developed back pain, fever, and rapidly progressive quadriparesis. Lumbar tap yielded frank pus which was confirmed on magnetic resonance imaging (MKI) to be located mainly in the cervical epidural space. Conservative antibiotic remedy was partially effective for restoration of the neurological deficits. A 82-year-old female noticed low-back pain which was rapidly accompanied with clouding of consciousness, paraplegia, and sphinc ter disturbances. Lumbar puncture revealed thick pus which was best depicted on MRI in the thoracolumbar subarachnoid space. At autopsy, spinal sub arachnoid abscess or leptomeningitis was confirmed, and a spinal infarction previously unrecognized on MRI was found. Usefulness and shortcomings of MRI in the diagnosis of paraspinal infections are discussed.
Introduction Acute spinal epidural and subarachnoid abscesses, otherwise termed spinal leptomeningitis, are rarely en countered in neurological practice and often escape a cor rect diagnosis, resulting in permanent neurological defi cits. Both initially present with back pain, high fever, and nerve root pain followed by rapidly progressing para- or tetraparesis and sphincter disturbances in the more ad vanced stage and possible death [1-5]. As is often the case, clinicians face the difficulty in differentiating the two similar but different diseases just by the clinical signs [4, 5]. Although the early utilization of magnetic reso nance imaging (MRI) is advocated for correct diagnosis [6, 7]. there have been few studies comparing MRI with other radiological methods in these two paraspinal infec tious states. The purpose of this report is to present cases
Received: November 14. 1991 Accepted after revision: June 16. 1992
of a spinal epidural and a subarachnoid abscess and to discuss the usefulness and limitations of MRI as a diag nostic tool.
Case Reports Patient I A 2 1-year-old male, with a previous history of cellulitis in the left leg 6 months before, suddenly developed progressively worsening low-back pain and fever on October 4. 1990. and subsequently was admitted to a local hospital. An epidural block was successful in con trolling the back pain, but on the following morning, the patient noticed severe nuchal pain that limited neck movements and numb ness of ankles and feet. As lumbar puncture revealed gray-yellowish pus from which Staphylococcus aureus was islolated. a tentative diag nosis of suppurative myelitis was established. Although intravenous ampicillin sodium was started immediately, the patient continued to have high fever and rapidly developed quadriparesis. numbness
Toshiya Fukui. MD Department of Neurology Showa University School of Medicine 1-5-8 Hatanodai. Shinagawa-ku Tokvo 142 (Jaoan)
© 1992 S. Karger AG. Basel 0014-3022/92/0326-0328 $2.75/0
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
Toshiya Fukui Hiroo Ichikawa Nobuyuki Kawate Taneyoshi Nozawa Koujiro Sugita
below the sensory level C-5 dermatome, and urinary retention. At this point he was transferred to our hospital for further investigation and treatment. Physical examination on admission revealed an acutely ill obscse male with a body temperature of 37.9 °C and severe spontaneous pain at spinous processes of C-6 and 7, with marked tenderness. There was no evidence of concurrent dermal, dental, or rhino-oph thalmologic infections. Neurological examination revealed that the patient was alert and oriented. The cranial nerves were unremark able. He showed flaccid tetraparesis, more prominent in the lower as well as the left extremities, total areflexia with positive bilateral Babinski signs, hypocsthesia of all modalities at C-5 level extending caudally. and bowel and bladder disturbances. Acute inflammatory reactions such as leukocytosis ( 13.400/mm3), increased erythrocyte sedimentation rate (120mm/h). and strongly positive C-reactive protein (25.5 mg/dl) were noted on admission. Lumbar tap at L-4/5 again produced purulent fluid with countless numbers of polymorphonuclear leukocytes, protein 4.058 mg/dl. and glucose 7 mg/dl. with the culture growing S. aureus. The roentgenograms and nonenhanced computerized tomograms (CT) of the cervical spinal column were unremarkable, and there was no sign of osteomyelitis. Myelography or myelo-CT was not per formed because it was considered to be inappropriate when the pres ence of an acute paraspinal infection w'as highly suspected. Sagittal cervical MRI (Toshiba MRT-50A. 0.5 T) revealed a sharply delineated mass of increased signal intensity with gadopcntctate dimegluminc (Gd-DTPA) injection on Tl-weighted image, lo cated around and mainly dorsal to the spinal cord. The lesion was situated in the epidural space between C-l and C-7, without invasion into the intracranium. Thoracic and lumbar MRI records were too poor, due to motion artifacts, for the lower end of the abscess to be evaluated. The epidural fat tissue, which would appear as a highintensity band in epidural space, could not be isolated from the abscess in the low'er cervical region (fig. I A). T2-weighted scans were not available. Based on the clinical and laboratory manifestations, he was diagnosed as having an acute spinal epidural abscess. The patient was soon responding to intensive intravenous antibi otic therapy with piperacillin sodium while waiting for surgical inter vention, which turned out to be unnecessary once quick and promis ing recovery' commenced. The abscess had disappeared on follow-up MRI I month after the onset of the symptoms (fig. IB) w'hen the patient did not have fever, nuchal pain, or paresis of the upper extremities and was capable of ambulating with assistance, albeit he had remaining spastic paraparesis and moderate sensory deficits in the legs.
Fig. 1. Patient l. With Tl-weighted (spin-echo technique, TR 300 ms. TE 14 ms) sagittal MRI with Gd-DTPA injection, the abscess is depicted in the epidural space as a sharply demarcated linear lesion of high signal intensity (arrows in A). The lesion is not seen after successful antibiotic treatment (B). The high-intensity area anterior to the spinal cord suggests the presence of a small amount of pus (A).
Patient 2
Fig. 2. Patient 2. Lumbar myelography shows a partial block of the subarachnoid space at T-9 and L-l due to an extramedullary lesion. It is inconclusive whether the lesion is located intradural or not.
329
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
An 82-ycar-old female, without a previous medical history of dia betes mcllilus or any suppurative diseases, noticed vague low-back pain while cleaning her house on December 31. 1989. The next morni ng, her back pain was also accompanied by weakness of the legs and fever of 38.0 °C. The paraparesis of the lower extremities rapidlyworsened. followed by mild clouding of consciousness, total loss of all sensations in the lower extremities, and loss of bladder and rectal control within days. She was admitted to a hospital with a suspicion of acute meningomvclitis. Myelography performed there showed a partial block of the spinal column between T-9 and L-l by an extra medullary lesion which was either intra- or extradural (fig. 2). On myelo-CT. the spinal cord appeared to be compressed anteriorly by localized masses which projected into the enhanced subarachnoid
A
B
Fig. 4. Patient 2. A Tl-weighted (spin-echo technique. TR 300 ms. TE 14 ms) sagittal MR1 (A) and a transaxial image at T-10 (B) with Gd-DTPA. The lesion (arrows in A) is located dorsal to the spi nal cord and anterior to the epidural fat from T-9 through L-l, but the lesion seems to extend further rostrally and caudally. A mass lesion of low signal intensity which is surrounded by a high-intensity band (black arrow in B) is present behind the spinal cord (asterisk in B). Epidural fat tissue (white arrow in B) can be recognized clearly separate from the mass lesion. The band of low signal intensity between them represents dura mater and arachnoid membrane.
330
Fukui/ichikaw'a/Kawate/Nozaw'a/Sugiia
Neuroradiological Study in Paraspinal Infections
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
Fig. 3. Patient 2. Myelo-CT at L-l shows a few extramedullary masses in the subarachnoid space. An outer band of low density cor responding to the epidural fat tissue is clearly recognized.
space in the same segment as shown on myelography (fig. 3). She was then on ampicillin sodium for a few days which did not bring about any amelioration of the symptoms. When transferred to our hospital on the 9th day. she was febrile (38.6 °C) and slightly disoriented, complaining of nausea and head ache with meningeal irritation signs, but was devoid of spinal percus sion tenderness or any other noticeable skin infection. Neurological examination revealed that the patient had total paraplegia in the legs, all modal sensory disturbances of a partial degree at L-l to L-3 and total sensory loss below L-3, brisker but equal tendon reflexes in the upper extremities, areflexia in the legs with positive bilateral Babinski signs, and sphincter disturbances. However, she did not show any cranial nerve palsy, weakness, or cerebellar ataxia of the upper extremities. Acute inflammatory signs such as leukocytosis (20,400/mm3), increased erythrocyte sedimentation rate (30 mrn/h). and positive Creactive protein (1.3 mg/dl) were remarkable among the laboratory Endings on admission. Urine culture grew Candida species under use of ampicillin sodium. A lumbar tap yielded purulent fluid which con tained more than 170.000 cells/mm1. 97% of which polymorphonu clear leukocytes, 5,000 mg/dl protein. 44 mg/dl glucose, and culture of the fluid grew Escherichia coii. The roentgenograms of the vertebral column showed no sign of osteomyelitis. Tl-weighted sagittal MR1 with Gd-DTPA of the tho racic region disclosed an ill-defined mass of the same or slightly higher signal intensity as the spinal cord, displacing it anteriorly from the level of T-7 at least to L-2. The lesion was recognized to have much more extensive rostral-caudal distribution than was inferred from myelography and myclo-CT (fig. 4A). The transaxial T lweighted MRI scan at T-10 level showed more or less the same find ings as myelo-CT. i.e.. the spinal cord was anteriorly dislocated by a few blocks of mass lesions of the same intensity as the spinal cord, some of which appeared to be encapsulated by a membrane-like structure enhanced by GD-DTPA. apparently inside a circular band of low intensity which would correspond to dura mater and arach noid membrane. Epidural fat was conspicuously depicted outside the dural sac and characterized by its markedly high signal intensity (fig. 4B). Combined with lumbar tap results, these mass lesions were regarded as subarachnoid pus. No abnormalities were detected in the spinal cord parenchyma on Tl-weighted image with Gd-DTPA. and a T2-weighted scan was not available. From these observations, the patient was diagnosed as having an acute spinal subarachnoid abscess or spinal leptomeningitis, caused by E. coli at the lower thoracic and upper lumbar regions. The pres ence of clouding of consciousness suggested concomitant purulent meningitis as a result of extension of the inflammatory process into the intracranium as inferred from the MRI findings, although mycloCT suggested otherwise. Candida in the urinary tract suggested prior infection by suppurative bacteria including E. coli which may have been adventitiously removed by antibiotic therapy. Because of her high age. the severity of paraplegia, and a poor nutritional state, she was rendered to be a poor candidate for surgery' and was immediately placed on piperacillin sodium and latamoxef sodium in place of ampicillin sodium. This regimen was effective for her meningitis and back pain, but no neurological improvement in myelopathy was noted. She was on the same regimen w'ithout any neurological improvement and expired on the 21st day because of concomitant cerebral hemorrhage and hematemesis from disseminated intravas cular coagulopathy.
At autopsy, no purulent exudate was encountered on the surface of the spinal dura mater, but once it was opened, there was sticky purulent material covering the spinal cord from C-2 to cauda equina, the thickest pus being present at T-9 to L-3. There was no gross com pression or deformity of the spinal cord. Microscopically, the puru lent material was confirmed as an aggregation of polymorphonuclear cells confined to the subarachnoid space. The spinal cord per sc did not show any infiltration with acute inflammatory cells in its paren chyma. Abscess masses had no encapsulating membrane against the inference from MRI findings. The most striking pathological change was a spinal cord infarction most prominently seen at T-12 and L-l (fig. 5). The whole gray matter and patchy areas of the white matter at T-12 (fig. 5A) and the anterior two thirds of the spinal cord at L-l (fig. 5B) were completely infarcted and microscopically infiltrated by macrophages, the presence of which suggested that the infarction was in a subacute to early chronic stage. A mild ischemic change was also noted in anterior and lateral columns between T-10 and T-12. The anterior spinal artery was patent, and there was no engorgement of the spinal venous plexus.
Discussion
Fig. 5. Patient 2. Mainly the gray matter and patchy areas of the white matter at T-12 (A) and an extensive area of the ventral two thirds of the spinal cord at L-l (B) are grossly infarcted. These areas may correspond to the territories irrigated by the anterior spinal artery. Note a large amount of abscess underneath the arachnoid membrane (arrowheads) and the absence of any direct compression to the spinal cord. Kliiver-Barrera. X 8.
fat, dura mater, arachnoid membrane, spinal cord, and abscess if any. If an abscess is to be distinguished from the normal structures in the spinal canal, it would necessitate intrathecal injection of contrast medium; therefore, myelo-CT would carrv the same risks as mvelography [ 6], MRI is being recognized as a noninvasive, repeatable, and potentially capable examination tool for delineating the location and the extent of an epidural abscess [2, 3, 6, 7], The MRI presentation of an epidural abscess in patient 1 may be characterized by a sharp margin, inabil ity of the abscess to extend into the intracranium, and dis appearance of normally bright high intensity of epidural fatty tissue. The fact that a cervical abscess has sharp
331
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
As Heusner [8] and others [4, 5. 9] described, the clini cal features of both epidural and subarachnoid abscesses follow similar courses and are characterized by back pain, radicular pain, limb paresis, and rapidly progressing quadriplegia. Therefore, these two entities are essentially indistinguishable from each other on clinical grounds. It is discussed that coexisting cerebral meningitis, ventricu litis, and spinal vascular lesions are more common in subarachnoid abscesses and that spinal percussion tender ness is more common in epidural abscesses [4. 10], but these are not absolute distinctive differences [11], Al though some authors maintain that the distinction be tween the two is of little practical significance, since surgi cal drainage and antibiotic treatment are all-important in both situations [12], it would be of neuroradiological interest as well as an adjunct help in selecting candidacy and timing for surgical treatment to know the exact loca tion and the extent of a lesion. Myelography is in some cases a successful method in distinguishing an epidural from a subarachnoid abscess [ 13.14], but in others it does not suffice [ 12, 15, 16], Some potential risk is also suggested by implanting pathogenic organisms into the subarachnoid space from an epidural abscess via spinal tapping [16]. Since patient 1 had an accelerated exacerbation of neurological symptoms after epidural block to control the back pain, it is possible that an abscess originally localized in the lumbar region might have been disseminated into the cervical area. With noninvasive plain CT. it is impossible to distin guish the soft tissues within the spinal canal, i.e., epidural
332
abscess per se. As the lower thoracic to upper lumbar spi nal cord is vulnerable to ischemia because this region cor responds to a borderzone between two adjacent radicular arteries [20, 21 ]. it is possible that radicular arteries were compressed by abscess or that arterial spasms were in duced by the irritative nature of infection, resulting in spi nal infarction at T-12 and L-l. It is uncertain why the spinal cord infarction escaped detection by the most sensitive diagnostic tool of MRI. but presumably may it be either because spinal infarction at a specific stage of disease progression may appear isosignaled on Tl-weighted image as noninjured normal spi nal parenchyma or because the spinal cord imaging in the thoracolumbar region is apt to be obscured by artifacts from surrounding structures. For more accurate diagnosis of a spinal infarction, spinal imaging not only on Tlweighted but also on T2-weightcd MRI setting would be recommended. Early detection of spinal infarction which may conse quently accompany spinal leptomeningitis would be of great importance in determining a conservative treatment strategy in lieu of surgical intervention, since decompres sion and removal of abscess may not be effective or suit able in reducing neurological impairments of those af flicted with this ailment once accompanied by spinal infarction. According to the previous literature [4, 5, II. 12. 19], the most common causative organism in both spinal epi dural and subarachnoid abscess is S. aureus. Patient 1 with an epidural abscess was not an exception, but patient 2 was one of the rare cases whose spinal leptomeningitis was caused by an organism other than this common bac terium. E. coli is regarded as a rare cause of paraspinal infections and shares only 2% of all causative organisms in a series of patients reviewed by Takenaka et al. [11], In either case, it was impossible to ascertain the exact entry route of infection to the paraspinal regions, but it was speculated that hematogenic dissemination of a causative organism from the respiratory or urinary tract, for in stance. might be more probable than direct extension of infection from nearby structures to the spinal canal.
Acknowledgements The authors would like to express sincere gratitude to Dr. Nobutaka Arai for providing pathological findings of the spinal cord and to Ms. Ramona Severson for syntactic revision of the manuscript.
Fukui/Ichikawa/Kawate/Nozawa/Sugita
Neuroradiological Study in Paraspinal Infections
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
boundaries and that it does not extend into the intracra nium. even though it is closely adjacent to the foramen magnum, would suggest that the abscess was located in the tight and closed epidural space which exists between the spinal dura mater proper and the periosteum of spi nous processes, both being joined to form one sheet of intracranial dura mater. Since the disappearance of epi dural fat on CT is regarded as a pathognomonic feature in epidural abscess [17], and whether or not the abscess located in the epidural space can be identified on myeloCT with ease, transaxial MRI seems to have no additional diagnostic advantage over myelo-CT except for its low ered risk due to the absence of contrast medium. Sagittal MRI. however, would be far more superior to myelo-CT in depicting the longitudinal extent of the disease. Spinal subarachnoid abscess or leptomeningitis, on the other hand, is a much rarer disease and considered to be one of the subtypes of subdural abscess. Although there have been few reports on the MRI findings in subarach noid abscess, Kurokawa ct al. [ 18] recently reported a case of spinal 'subdural' abscess which is depicted 'intrathecally’ as an isointensity on T1 -weighted and as a strongly high-signaled band on T2-weighted scans. MRI findings of the subarachnoid abscess in patient 2 are remarkable in that the demarcation of the abscess is more ill-defined, that it is easy to identify epidural fat apart from an abscess, and that the rostral-caudal extent of the abscess is more precisely correlated with autopsy findings than myelo-CT. As mentioned above, transaxial myelo-CT may be as efficacious as MRI in defining size, configuration, and location of an abscess in relation to dura mater and spinal cord; the superiority of MRI to myelo-CT in the diagnosis of spinal subarachnoid ab scess. therefore, may exist in its availability of diagnosti cally valuable sagittal scans as well as its being noninvasive and easily repeatable. In this case, enhancement by Gd-DTPA was not much help in separating the lesion from the normal structures, and also careful interpreta tion of ‘ringed enhancement' is needed because the au topsy examination did not prove any capsular membrane around abscess masses. Without similar cases, it remains uncertain as to what it may represent. Paraplegia and sensory loss in the lower extremities observed in patient 2 were most likely to be ascribed to spinal infarction at T-12 and L-l rather than to nonsegmental and diffuse compression, if any, of the spinal cord by abscess. This may quite well coincide with the estima tion of Freedman and Alpers [19] that spinal vascular lesion is more responsible for the presentation of neuro logical symptoms than the compression of spinal cord by
References 8 Heusner AP: Nontubcrculous spinal epidural infections. N Engl J Med 1948;239:845-854. 9 Hirson C: Spinal subdural abscess. Lancet 1965;ii: 1205—1217. 10 Dus V: Spinal peripachymeningitis (epidural abscess). J Neurosurg 1960:17:972-983. 11 Takcnaka K, Kobayashi H. Niikawa S. Hattori T. Ookuma S. Nokura H. Sakai N. Yamada H. Sasaoka I: Spinal subdural abscess. Report o f a case and a review of the literature of 43 cases. Brain Nerve (Tokyo) 1989:41:331-336. 12 Schiller E, Shadlc OW: Extrathecal and intra thecal suppuration. Arch Neurol 1962:7:33— 36. 13 Patronus NJ, Marx WJ, Duda EE: Radiographic presentation of spinal abscess in the subdural space. AJR 1979:132:138-139. 14 Danner RL. Hartman B.I: Update of spinal epi dural abscess: 35 cases and review of the litera ture. Rev Infect Dis 1987:9:265-274. 15 Messer HD, LenchnerGS. Brust JCM. RcsorS: Lumbar spinal abscess managed conservative ly. J Neurosurg 1977:46:825-829.
16 Knudsen LL. Voldby B. Stagaard M: Com puted tomographic myelography in spinal sub dural empyema. Neuroradiology 1987:29:99. 17 Leys D. Lesoin F. Viaud C. Pasquier F. Rous seaux M. Jomin M. Petit H: Decreased morbid ity from acute bacterial spinal epidural abscess using computed tomography and nonsurgical treatment in selected patients. Ann Neurol 1985:17:350-355. 18 Kurokawa Y. I lashi K. Fujishigc M. Tsuchida M. Maeda K, Kaneko M: Spinal subdural em pyema diagnosed by MRI and recovered by conservative treatment. Brain Nerve (Tokyo) 1989:41:513-517. 19 Freedman H. Alpers BJ: Spinal subdural ab scess. Arch Neurol Psychiatry 1948:60:49-60. 20 Lazorthcs G. Gouaze A, Zadeh JO. Santini JJ. Lazorthes Y. Burdin P: Arterial vascularization of the spinal cord. Recent studies of the ana tomic substitution pathways. J Neurosurg 1971:35:253-261. 21 Turnbull IM: Microvasculature of the human spinal cord. J Neurosurg 1971:35:141-147.
333
Downloaded by: Stockholm University Library 130.237.165.40 - 12/16/2018 10:25:13 AM
1 Baker AS. Ojemann RG. Swartz MN. Richard son EP Jr: Spinal epidural abscess. N Engl J Med 1975:293:463-468. 2 Hlavin ML. Kaminski HJ. Ross JS. Ganz E: Spinal epidural abscess: A ten-year perspective. Neurosurgery 1990:27:177-184. 3 Curling OD. Gower DJ. McWhorter JM: Changing concept in spinal epidural abscess: A report of 29 cases. Neurosurgery 1990:27:185— 192. 4 Fraser RAR. Ratzan K. Wolpert SM. Wein stein L: Spinal subdural empyema. Arch Neu rol 1973:28:235-238. 5 Dacey RG. Winn HR. Jane JA. Butler AB: Spi nal subdural empyema: Report of two cases. Neurosurgery 1978:3:400-403. 6 Erntell M. Hollas S. Norlin K. Dahlquist E. Nilsson-Ehle I: Magnetic resonance imaging in the diagnosis of spinal epidural abscess. Scand J Infect Dis 1988:20:323-327. 7 Hanigan WC. Asner NG. Elwood PW: Mag netic resonance imaging and the nonoperative treatment of spinal epidural abscess. Surg Neu rol 1990:34:408-413.
