Outcome in 53 patients with spinal cord cavernomas

Available online at www.sciencedirect.com

Surgical Neurology 70 (2008) 176 – 181 www.surgicalneurology-online.com

Spine

Outcome in 53 patients with spinal cord cavernomas Pierre Labauge, MD, PhD a,⁎, Stéphane Bouly, MD a , Fabrice Parker, MD, PhD b , Sophie Gallas, MD c , Evelyne Emery, MD, PhD d , Hugues Loiseau, MD, PhD e , Jean Paul Lejeune, MD f , Michel Lonjon, MD g , François Proust, MD h , Sergio Boetto, MD i , Sophie Coulbois, MD j , Jean Auque, MD, PhD j , Jacques Boulliat, MD k , on behalf of the French Study Group of Spinal Cord Cavernomas a Service de Neurologie, CHU de Nîmes, 30029 France Services de Neurochirurgie, CHU de Bicêtre, 94275 France c Services de Neurochirurgie, CHU de Reims, 51000 France d Services de Neurochirurgie, CHU de Caen, 14011 France e Services de Neurochirurgie, CHU de Bordeaux, 33000 France f Services de Neurochirurgie, CHU de Lille, 59000 France g Services de Neurochirurgie, CHU de Nice, 06100 France h Services de Neurochirurgie, CHU de Rouen, 76000 France i Services de Neurochirurgie, CHU de Toulouse, 31033 France j Services de Neurochirurgie, CHU de Nancy, 57000 France k Service de Neurologie, CH de Bourg en Bresse, 01100 France Received 10 May 2007; accepted 13 June 2007 b

Abstract

Background: Prevalence of cerebral cavernomas in the general population is close to 0.5%. In contrast, SCCs are rare. The aim of this study was to determine the outcome of SCC in a large sample of patients. Methods: Clinical and neuroradiologic findings were retrospectively collected in a multicentric study. Diagnosis was based on pathologic criteria or MR findings. Results: Fifty-three patients were included (26 males, 27 females). Mean age at onset of symptoms was 40.2 years (11-80 years). Initial symptoms were progressive (32) or acute myelopathy (20). One case was asymptomatic. Triggering factors were found in 14 of the patients (26%). Clinical symptoms were related to spinal cord compression (27) and hemorrhage (22). Spinal cord cavernoma was thoracic in 41 cases and cervical in 12. Mean size of the lesions was 16.3 mm (3-54 mm). In the 40 surgical patients, long-term follow-up was available in 37 cases for a mean time of 7.3 years (0.4-50 years). During the follow-up period, 20 patients improved, 6 remained on their preoperative baseline, and 11 got worse. Surgical improvement was more often found in posterior rather than anterior location. Using McCormick classification, 22 patients were autonomous (grades 1-2), 12 handicapped (grade 3), and 3 bedridden (grade 4) at the end of the follow-up. Conclusions: This study has defined clinical and MR patterns of spinal cavernomas. Surgery lastingly improved more than half of the patients. © 2008 Elsevier Inc. All rights reserved.

Keywords:

Cavernous angioma; Magnetic resonance imaging; Intramedullary; Spinal cord; Neurosurgery

Abbreviations: CCM, cerebral cavernous malformations; MR, magnetic resonance; SCC, spinal cord cavernoma. ⁎ Corresponding author. Service de Neurologie, CHU MontpellierNîmes, Hôpital Caremeau, 30029 Nîmes cedex 4, France. Tel.: +33 4 66 68 32 61; fax: +33 4 66 68 40 16. E-mail address: [email protected] (P. Labauge). 0090-3019/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.surneu.2007.06.039

1. Introduction Cavernomas are defined by abnormal enlarged cavities without any nervous tissue. Prevalence is close to 0.5% in the general population, of which 10% are familial [20].

P. Labauge et al. / Surgical Neurology 70 (2008) 176–181

177

Spinal cord cavernomas are considered rare, approximately 5% of spinal cord lesions. Previous publications studied limited numbers of patients. The most important overlooked 26 surgically treated patients [19]. Review of literature by Deutsch et al [11] found 94 published cases. This poor information explains natural history, and prognostic factors after surgery are poorly understood. The aims of this nationwide study were to define (1) clinical and MR patterns and (2) outcome in surgical patients. 2. Methods We led a retrospective multicentric study including 11 hospital neurosurgical centers (see Appendix A) from 1996 to 2006. Inclusion criteria were patients with definite diagnosis of spinal cavernoma. Diagnosis was based on MR or neuropathologic findings. Exclusion criteria were cervicomedullary and spinal epidural cavernomas [14,21, 27], other types of malformations, or unavailable clinical or MR data. Patients were clinically classified as asymptomatic, sudden onset (b1 day), or progressive onset (N1 day).

Fig. 2. A: Sagittal MR T2 sequence. B: Axial MR T1 sequence. Intralesional hemorrhage of a thoracic SCC.

Fig. 1. Sagittal MR T2 sequence. Hematomyelia revealing a thoracic SCC.

Symptoms were due to acute hemorrhage, including diffuse hemorrhage (hematomyelia) (Fig. 1); intralesional hemorrhage (Fig. 2); or compression, without acute hemorrhage (Fig. 3). Diagnosis of acute hemorrhage was defined by histologic or MR findings (hyperintense signal on T1 sequences). Magnetic resonance consisted of high-field MR, including sagittal T1-weighted, sagittal and axial T2weighted, gradient echo sequences. The size of the lesions and their spinal location (cervical, dorsal, and lumbar) were recorded. Spinal cord location was defined in a horizontal plane (anterior, central, posterior, or lateral location) compared with the midline and dentatus ligament planes. Natural history was based on follow-up of the patients from the onset of the disease until surgery or until the last medical information in the absence of surgery. The long-term surgical outcome was based on follow-up from surgery to the last medical information.

178

P. Labauge et al. / Surgical Neurology 70 (2008) 176–181

McCormick classification [23] distinguishes 4 functional stages (grade 1: neurologically normal, mild focal deficit not significantly affecting function of involved limb, normal gait; grade 2: sensorimotor deficit affecting function of involved limb, mild to moderate gait difficulty, still independent walk; grade 3: more severe neurologic deficiencies, requires cane/brace for ambulation; may or may not be independent; grade 4: severe deficit, requires wheelchair or has bilateral upper extremity impairment, usually not independent). All investigators examined their own patients. Clinical data were retrospectively recorded by PL and SB. Magnetic resonance analysis, including the size, site, and signal characteristics, was reviewed by an independent neuroradiologist (SG). We used the nonparametric Mann-Whitney U test to compare quantitative values and χ2 for qualitative values. A P value of less than .5 asserted a significant difference. 3. Results Fifty-three patients were included (26 males and 27 females). Mean age at onset was 40.2 years (11.7-80.2 years). Forty patients underwent surgery. 3.1. Initial symptoms and their mechanisms Neurologic symptoms were acute in 20 cases (38%), progressive in 32 (60%). Only 1 patient was asymptomatic (2%). Triggering factors were found in 14 of the patients (26%): pregnancy (3 cases), trauma (1 case), and strenuous activity (10 cases). Clinical symptoms were due to hemorrhage in 22 cases (39%), including hematomyelia in 19 (36%), intralesional hemorrhage in 3 (6%), and compression in 27 cases (51%). In 3 patients (6%), the exact mechanism was not sure because of incomplete data. Clinical symptoms were sensitive (24), motor (15), and sensitive and motor (13). By McCormick classification, 20 patients were of grade 1, 17 of grade 2, 10 of grade 3, and 6 of grade 4 at the onset of symptoms. 3.2. Magnetic resonance data Spinal MR was done for all patients. Lesions were thoracic in 41 (77%) and cervical in 12 (23%). In a horizontal plane, the lesion was central in 18, posterior in 12, lateral in 9, and anterior in 5. Determining the precise location was not possible in 9 cases. The average size was 16.3 mm (3-54 mm) in height and 7 mm (2-15 mm) in width. Of the 38 patients who underwent brain MR examination, brain cavernomas were found in 14 (37%), including 9 known familial forms. In the remaining 24 patients (63%), no lesion was found.

Fig. 3. A: Sagittal MR T1 sequence. B: Sagittal MR T2 sequence. C: Axial MR gradient echo sequence. Cervical SCC. Absence of acute hemorrhage.

P. Labauge et al. / Surgical Neurology 70 (2008) 176–181 Table 1 Long-term results of surgery (N = 37)

Table 3 Comparison of patients who improved or got worse after surgery

McCormick

1

2

3

4

Initial Final

13 16

11 6

10 12

3 3 Site

Only 1 patient had multiple SCCs, with 2 thoracic lesions. Surprisingly, extensive investigation with brain MR, including gradient echo sequences, and thorough questioning of the relatives did not yield information.

Mechanism

3.3. Natural history Of the 53 included patients, 20 (38%) had acute initial symptoms, and 32 (60%) had progressive onset. One patient (2%) had a fortuitous diagnosis. The mean follow-up lasted for 7.3 years (0.4-5.0 years). 3.3.1. Sudden onset (20 patients) In 20 of these patients, clinical symptoms were related to hematomyelia for 14 (70%), intralesional hemorrhage in 1 (5%), and compression in 3 (15%). Data were not available for 2 cases (10%). • Of the 14 patients with hematomyelia, 3 (21%) did not undergo surgery. The average follow-up for these patients was 5.2 years (3.7-7 years). There was spontaneous improvement in 1 and unchanged status in 2 cases. Initial McCormick scores were 2 in the first patient and 1 in the other 2 patients. There was no hemorrhagic relapse during follow-up. • Of 14 patients with hematomyelia, 11 (79%) underwent surgical resection at the onset of symptoms. Mean follow-up lasted for 1.9 years (0.2-11 years). Six patients (55%) improved, 3 (27%) remained stable, and 1 (9%) was functionally worse after surgery. Clinical evolution was not known in 1 case (9%). 3.3.2. Progressive symptoms (32 patients) Of these 32 patients, clinical symptoms were due to compression in 24 cases (75%). Hematomyelia was found in

Table 2 Comparison of patients who improved or got worse after surgery

Sex Mean age (y) Follow-up (y) Size (mm) Location

Total: 31

Improved (n = 20) n

%

n

%

M: 18 F: 13

10 10 38.7 2.5 16.2 6.1 15 5

56 77

8 3 41.4 5.1 14.1 6.9 8 3

44 23

Height Width Thoracic: 23 Cervical: 8

179

Worse (n = 11)

65 63

Initial McCormick

Anterior: 4 Posterior: 8 Central: 11 Lateral: 3 ND: 5 Compression: 18 Hematom: 11 Hemor intr: 1 Not determined: 1 1: 11 2: 10 3: 7 4: 3

Improved (n = 20)

Worse (n = 11)

n

%

n

%

1 7 7 2 3 10 8 1 1 5 7 6 2

25 85 64 66 60 56 73% – – 45 70 86 66

3 1 4 1 2 8 3 0 0 6 3 1 1

75 15 36 34 40 44 27 – – 55 30 14 33

Hematom indicates hematomyelia; Hemor intr: intralesional hemorrhage.

5 cases (16%), intralesional hemorrhage in 2 (6%). The precise mechanism remained unknown in 1 patient. • Of these 24 nonhemorrhagic patients, 5 (21%) did not undergo surgery. Mean follow-up was 2.4 years (1-4.2 years). Clinical evolution was unchanged in 3 cases (60%), worse in 1 case (20%), and not available in 1 case (20%). • For the 19 patients who underwent surgery, 9 (47%) improved, 8 were worse (42%), and 1 was unchanged (5%). Follow-up had a mean length of 3.9 years (124 years). Follow-up was not available in 1 case (5%). 4. Postoperative course and prognostic factors Data for long-term postsurgery evolution were available for 37 of the 40 surgically treated patients (Tables 1-3). Clinical status was unchanged in 6 patients (16%); mean follow-up was 3.9 years (1-11 years). Surgery durably improved 20 (54%; mean follow-up, 2.5 years [0.2-14.8 years]); 11 were worse (30%; mean follow-up, 5.1 years [0.3-24 years]). Comparing age, sex, delay from onset to surgery, lesion size, cervical or thoracic location, mechanisms (hematomyelia or compression), initial McCormick evaluation, we found no statistical difference between the improved patients and those who got worse. Surgical efficiency depended on location of the lesions, good for 1 of 4 patients with anterior location and for 7 of 8 patients with posterior location (P b .005) (Tables 2 and 3). 5. Final evaluation

35 37

The mean follow-up of the 37 surgical patients was 3.5 years (1-24 years). Of the 37 patients, 22 were of grades 1 to 2, 12 were of grade 3, and 3 were of grade 4.

180

P. Labauge et al. / Surgical Neurology 70 (2008) 176–181

6. Discussion Our study may be the largest we know concerning the outcome of SCC. Spinal cord cavernomas are considered rare, with a prevalence of close to 5% of spinal cord lesions. Most reports concerned isolated cases or small series [1,4,6,8,13,19,22,25,26,28,30,31,33,34]. So, clinical symptoms, evolution, and prognosis are not well known [3,7,8,10,19]. The large number of SCC in our series allowed us to specify clinical and MR characteristics of intramedullary cavernomas. The average age of onset was 40 years, confirming that SCC becomes clinically apparent from the third to the fifth decade [11]. As Cristante et al [9], we found a female-male ratio of 1:1. The 2:1 female-male ratio previously reported was based on limited series and literature analysis. It suggested that clinical symptoms were more frequent in female patients because of possible hormonal effects. Clinical onset followed mechanical factors in 20% of cases, as was already noted by McCormick et al in 2 of their 6 cases. These mechanical factors are of many types, including pregnancy, minor trauma, and strenuous exercise [23]. By contrast, brain cavernomas are less sensitive to environmental factors, except for anticoagulant or hormonal treatment. Histologic analysis of brain and spinal cavernomas are similar [5], apart from associated venous angioma that was never found in SCC [12,32]. This suggests a mechanical trigger is correlated to spinal cord location rather than anatomopathologic differences. We could describe spinal lesions, their symptoms, size, and spinal cord location. As previously seen [11], 77% (41/ 53), are thoracic. Cervical are less frequent and lumbar medullary conus, exceptional. Only 1 of our patients (2%) had multiple spinal lesions with a normal brain MR [2]. Multiple SCCs are rare, with an estimated incidence of around 2% of SCC [20,34]. We did not find any multiple SCC in patients with known familial and multiple brain cavernomas. This suggests spinal lesions are essentially single whether they belong or not to a familial form. Our study, including more than 50 consecutive patients, allowed us to highlight the outcome of SCC. Clinical onset can be sudden (38%) or progressive (60%). Acute symptoms are mostly (70% of the cases) linked to acute hemorrhage [11]. Unlike McCormick et al [23], we did not find that thrombosis was a cause of acute symptoms. The remaining patients (60%) have progressive symptoms with a tumoral or multiple sclerosislike evolution. Many mechanisms have been suggested: microhemorrhages and resultant gliosis, neurotoxic effect of hemosiderin, impaired microcirculation because of local pressure, and progressive enlargement of the lesion [25]. But, we did not observe subarachnoid hemorrhage, which has been only described in extramedullary spinal cord lesions [15-17]. Different clinical symptoms have been previously described in literature [14,19,23-25,34]. Ogilvy et al [25]

underlined distinct patterns of presentation: (1) discrete episodes of neurologic deterioration punctuated by varying degrees of recovery; (2) slow progression of neurologic decline; (3) sudden onset of symptoms with rapid decline; (4) sudden onset of mild symptoms followed by gradual decline. Natural history of SCC is not well known. Bleeding rate was estimated between 1.4% and 4.5% per year [34,29]. Of the 13 patients in our series who were not surgically treated, 4 had relapse of symptoms (hemorrhage in 3 and compression in 1). Surgery lastingly improved half of our patients, as in previous reports [1,13,34]; Deutsch et al [11] found that surgery improved all of their 16 patients but 1. The only predictive factor of postsurgical improvement in our patients was the spinal cord location in a horizontal plane. Surgery improved 7 of 8 anterior SCCs and only 1 of the 4 posterior SCCs. Other good prognosis was suggested in the literature as degree of initial disability and short delay of surgery [34]. Surgery is very efficient on pain, as suggested by Kim et al [18], who showed that surgery lastingly improved pain in half of their patients. Transient postsurgical neurologic worsening is seen in half of the patients [1,11,19], mostly followed by a long-term improvement. Complete resection is achieved in most of the patients and explains the long-term efficiency [23,30]. To conclude, our study allows to define clinical and MR findings and the outcome in patients with SCC. Surgery is more often efficient in posterior cavernomas. Appendix A French Study Group of Spinal Cord Cavernomas included Pierre Labauge, Stéphane Bouly, Pablo Bouillot, Department of Neurology (CHU de Nîmes); Fabrice Parker, Department of Neurosurgery (CHU de Bicêtre); Sophie Gallas, Department of Neuroradiology (CHU de Reims); Evelyne Emery, Department of Neurosurgery (CHU de Caen); Hugues Loiseau, Department of Neurosurgery (CHU de Bordeaux); Jean Paul Lejeune, Department of Neurosurgery (CHU de Lille); Michel Lonjon, Philippe Paquis, Department of Neurosurgery (CHU de Nice); François Proust, Department of Neurosurgery (CHU de Rouen); Sergio Boetto, Michel Tremoulet Department of Neurosurgery (CHU de Toulouse); Sophie Coulbois, Jean Auque Department of Neurosurgery (CHU de Nancy); and Jacques Bouillat (CHU de Bourg en Bresse). References [1] Anson JA, Spetzler RF. Surgical resection of intramedullary spinal cord cavernous malformations. J Neurosurg 1993;78:446-51. [2] Balaban H, Sener HO, Erden I, Caglar S, Sahin A, Yucemen N. Multiple spinal intramedullary cavernous angioma: case report. Clin Neurol Neurosurg 2001;103:120-2. [3] Barnwell SL, Dowd CF, Davis RL, Edwards MSB, Gutin PH, Wilson CB. Cryptic vascular malformations of the spinal cord: diagnosis by

P. Labauge et al. / Surgical Neurology 70 (2008) 176–181

[4]

[5]

[6]

[7]

[8]

[9] [10]

[11]

[12]

[13] [14]

[15]

[16]

[17]

[18]

[19]

[20]

[21] [22]

magnetic resonance imaging and outcome of surgery. J Neurosurg 1990;72:403-7. Barrena MR, Guelbenzu S, Mayayo-Sinues E, Rivero-Celada D, Fayed-Miquel N, Gomez-Perun J. A clinico radiological study of patients with cavernous angiomas of the spinal cord. Rev Neurol 2003; 37:1-7. Bourgouin PM, Tampiere D, Johnston W, Stewrad J, Melancon D, Ethier R. Multiple occult vascular malformations of the brain and spinal cord: MRI diagnosis. Neuroradiology 1992;2:110-1. Bucciero A, del Basso de Caro M, Carangelo B, Vizioli L. Intramedullary cavernoma: a case report and review of the literature. Neurol (Napoli) 1994(4):162-9. Canavero S, Pagni CA, Duca S, Bradac GB. Spinal intramedullary cavernous angioma: a literature meta-analysis. Surg Neurol 1994;5: 381-8. Cantore G, Delfin R, Cervoni L, Innocenzi G, Orlando ER. Intramedullary cavernous angiomas of the spinal cord: report of six cases. Surg Neurol 1995;43:448-51. Cristante L, Hermann HD. Radical excision of intramedullary cavernous angiomas. Neurosurgery 1998;43:424-30. Del Curling Jr O, Kelly Jr DL, Elster AD, et al. An analysis of the natural history of cavernous angiomas. J Neurosurg 1991;75: 702-8. Deutsch H, Jallo GL, Faktorovich A, Epstein F. Spinal intramedullary cavernoma: clinical presentation and surgical outcome. J Neurosurg 2000;93(1 Suppl):65-70. Fontaine S, Melanson D, Cosgrove R, Betrand G. Cavernous hemangiomas of the spinal cord: MR imaging. Radiology 1988;166: 839-41. Gordon CR, Crockard HA, Symon L. Surgical management of spinal cord cavernoma. Br J Neurosurg 1995;9:459-64. Harrington Jr JF, Khan A, Grunnet M. Spinal epidural cavernous angioma presenting as a lumbar radiculopathy with analysis of magnetic resonance imaging characteristics: case report. Neurosurgery 1995;36:581-4. Heimberger K, Schnaberth G, Koos W, Pendl G, Auff E. Spinal cavernous haemangioma (intradural-extramedullary) underlying repeated subarachnoid haemorrhage. J Neurol 1982;226:289-93. Hida K, Tada M, Iwasaki Y, et al. Intramedullary disseminated capillary hemangioma with localized spinal cord swelling: case report. Neurosurgery 1993;33:1099-101. Kim CH, Kim HJ. Subarachnoid haemorrhage malformation presenting with recurrent cervical subarachnoid floating cavernous. J Neurol Neurosurg Psychiatry 2002;72:668. Kim LJ, Klopfenstein JD, Zabramski JM, Sonntag VK, Sptezler RF. Analysis of pain resolution after surgical resection of intramedullary spinal cord cavernous malformations. Neurosurgery 2006;1:106-11. Jallo GI, Freed D, Zareck M, Epstein F, Kothbauer KF. Clinical presentation and optimal management of intramedullary cavernous malformations. Neurosurg Focus 2006;21:E10. Labauge P, Laberge S, Brunereau L, et al. Hereditary cerebral cavernous angiomas: clinical and genetic features in 57 French families. Société Française de Neurochirurgie. Lancet 1998;352: 1892-7. Lee JP, Wang ADJ, Wai YY, et al. Spinal extradural cavernous hemangioma. Surg Neurol 1990;34:345-51. Lopate G, Black JT, Grubb RL. Cavernous hemangioma of the spinal cord: report of 2 unusual cases. Neurology 1990;40:1791-3.

181

[23] McCormick PC, Michelsen WJ, Post KD, et al. Cavernous malformations of the spinal cord. Neurosurgery 1988;23:459-63. [24] Mehdorn HM, Stolke D. Cervical intramedullary cavernous angioma with MRI-proven hemorrhage. J Neurol 1991;238:420-6. [25] Ogilvy CS, Louis DN, Ojemann RG. Intramedullary cavernous angiomas of the spinal cord: clinical presentation, pathological features, and surgical management. Neurosurgery 1992;31:219-30. [26] Padovani R, Acciarri N, Giulioni M, Pantieri R, Foschini MP. Cavernous angiomas of the spinal district: surgical treatment of 11 patients. Eur Spine J 1997;6:298-303. [27] Roncaroli F, Scheithauer BW, Krauss WE. Hemangioma of spinal nerve root. J Neurosurg 1999;91(Spine 2):175-80. [28] Roncaroli F, Scheithauer BW, Krauss WE. Capillary hemangioma of the spinal cord. Report of four cases. J Neurosurg Spine 2000;93: 148-51. [29] Sandalcioglu IE, Wiedemayer H, Gasser T, Asgari S, Engelhorn T, Stolke D. Intramedullary spinal cord cavernous malformations: clinical features and risk of hemorrhage. Neurosurg Rev 2003;26:253-6. [30] Santoro A, Piccirilli M, Frati A, Salvati M, Innocenzi G, Ricci G, Cantore G. Intramedullary spinal cord cavernous malformations: report of ten new cases. Neurosurg Rev 2004;27:93-8. [31] Vaquero J, Martinez R, Martinez P. Cavernomas of the spinal cord: report of two cases. Neurosurgery 1988;22:143-4. [32] Vishteh AG, Sankhla S, Anson JA, et al. Surgical resection of intramedullary spinal cord cavernous malformations: delayed complications, long-term outcomes, and association with cryptic venous malformations. Neurosurgery 1997;41:1094-101. [33] Zentner J, Hassler W, Gawehn J, et al. Intramedullary cavernous angiomas. Surg Neurol 1989;31:64-8. [34] Zevgaridis D, Buttner A, Weis S, Hamburger C, Reulen HJ. Spinal epidural cavernous hemangiomas. Report of three cases and review of the literature. J Neurosurg 1998;5:903-8.

Commentary This is an important contribution from a French cooperative retrospective study on a large series of 53 patients with SCCs. The rate of related symptoms to hemorrhage is high (22 patients), which is a strong argument in favor of surgery as soon as the lesion is symptomatic. I am impressed by the level of postoperative improvement (20 patients), which is unusual in SCCs and different from our experience based on 24 patients among whom 22 remained neurologically stable except for a favorable benefit on pain. The cooperative article confirms the indication to operate those lesions as soon as they are symptomatic. Jacques Brotchi, MD, PhD Department of Neurosurgery Erasme Hospital Univ. Brussels B-1070 Brussels, Belgium