Acta Neurochir (2003) 145: 411–415 DOI 10.1007/s00701-003-0025-2

Case Report Spinal subdural haematoma as a complication of cranial surgery J. I. Lee and S. C. Hong Department of Neurosurgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea Published online May 19, 2003 # Springer-Verlag 2003

Summary Background. Spinal subdural haematoma is a rare condition usually associated with several precipitating factors including coagulopathy, lumbar puncture, trauma, vascular malformation and previous spinal surgery. In this paper we report spinal subdural haematoma related to cranial surgery which is a previously unknown precipitating factor. Method. The medical records of six patients in whom spinal subdural haematoma developed after cranial surgery was reviewed retrospectively for clinical presentation, radiological findings, treatment, and outcome. Findings. Six patients presented with low back pain and radiculopathy in the lower extremity after surgery for intracranial lesions. Symptom onset was between 2 and 9 days after cranial surgery. Initial cranial procedures were craniotomy and tumour removal in 1 patient, clipping of aneurysm in 1, temporal lobectomy for epilepsy in 4. None of the patients had previously known precipitating factors for spinal subdural haematoma. In all of them, the diagnosis was confirmed by magnetic resonance (MR) imaging and the spinal segment involved was the lower lumbar and sacral level except for one patient with a wide distribution of haematoma over the thoracolumbar region. All patients recovered completely without surgical intervention. Interpretation. Spinal subdural haematoma is a rare but possible complication of cranial surgery. It should be considered in patients with back pain and radiculopathy in the lower extremity developing after surgery for intracranial lesions. Unlike spontaneous spinal subdural haematoma with other precipitating factors, spinal subdural haematoma developing after cranial surgery takes a benign clinical course and resolves spontaneously over several days to 2 weeks without surgical intervention. Keywords: Spinal subdural haematoma; craniotomy; complication.

Introduction Nontraumatic spinal subdural haematoma occurs rarely and the most common causative factor is coagulopathy due to haematological diseases or anticoagulant therapy [1]. Iatrogenic factors such as lumbar puncture or epidural anesthetic procedures contribute secondly to coagulopathy. Extremely rare cases of spinal subdural

haematoma following intracranial surgical procedures have also been reported [3, 4, 6, 7]. We describe 6 cases of spinal subdural haematoma which developed after intracranial surgery. The clinical features are discussed. Patients and methods Between 1995 to 2001 2,689 craniotomies for intracranial conditions excluding traumatic lesions were performed at our institution. Six patients who developed spinal subdural haematoma after cranial surgery were studied retrospectively for clinical presentation, time course, radiological findings, treatment, and outcome.

Results Six patients presented with acute low back pain, radicular pain and=or paresthesia after cranial surgery. Diagnosis of spinal subdural haematoma was confirmed by the findings on MR study. Their demographic characteristics as well as some clinical data are summarized in Table 1. None of the patients showed abnormal findings in pre- and postoperative laboratory tests for coagulopathy. Initial cranial procedures were craniotomy and tumour removal in 1 patient, clipping of aneurysm in 1, temporal lobectomy for epilepsy in 4. All the procedures were performed in a supine position under general endotracheal anesthesia and the head was not elevated more than 30 degrees from the horizontal plane. There were no unexpected events during the operation or immediate postoperative recovery period. Lumbar puncture or other surgical procedures around the spine were not done before the development of symptoms due to spinal subdural haematoma.

412

J. I. Lee and S. C. Hong

Table 1. Clinical summary of 6 patients with spinal subdural haematoma after cranial surgery Patient number

Age=sex

Diagnosis

Cranial surgery

Location of haematoma

Interval to symptom onset (days)

Treatment

Interval to recovery (days)

1 2 3 4 5 6

39=M 57=F 26=M 24=M 36=F 20=M

right frontal astrocytoma aneurysm epilepsy epilepsy epilepsy epilepsy

tumor removal clipping temporal lobectomy temporal lobectomy temporal lobectomy temporal lobectomy

T-L-S L4-S L3-S L5-S L4-S L3-S

4 4 8 4 7 9

lumbar – lumbar – lumbar lumbar

20 10 16 10 8 3

puncture puncture puncture puncture

Fig. 1. Case 5. Sagittal T1-(a) and T2-(b) weighted MR images of lumbar spine reveal high-signal haematoma over L4 to S1 level within the dural sac effacing subarachnoid space. Axial T1 -weighted image reveals two compartments of biconvex haematoma compressing subarachnoid space from dorsal aspect

All six patients presented with low back pain and radicular pain on both sides of the lower extremity. There were limitations in the straight leg-raising test and they complained of paresthesia in the lower limbs. Other symptoms included neck stiffness and headache. Onset of symptoms was between 4 and 9 days after the initial cranial procedures. All patients underwent MR imaging of the thoracolumbar spine and revealed spinal subdural haematoma within 2 days from the symptom

onset. A representative case is shown in Fig. 1. In sagittal MR images haematoma was distributed mainly along the dorsal aspect of dural sac. Axial MR image revealed two or three compartments of biconvex collection of haematoma clearly separate from the subarachnoid space. The high signal intensity in both T1- and T2-weighted images suggested that haematoma was in the subacute or chronic phase. Haematoma showed low signal intensity in none of the patients which could be

Spinal subdural haematoma as a complication of cranial surgery

413

Fig. 2. Case 1. Sagittal T1-weighted MR image of lumbar (a) and thoracic (b) spine reveal high signal lesion suggesting haemorrhage in subacute stage. MR image at 3 weeks after symptom onset (c) shows that the lesion disappeared completely

interpreted as pure blood clot in the acute stage. In 5 patients, haematoma was located in the lumbar or sacral level, but a more extensive spread including the thoracic level was found in 1 patient (Fig. 2). In four patients, lumbar puncture and drainage of spinal subdural haematoma was done between 1 and 5 days after the onset of symptoms. At lumbar puncture, old brown haemorrhagic fluid was obtained and there was more or less immediate improvement of symptoms after drainage. The rest of the patients were treated on bed rest and symptomatic medication only. Symptoms and signs improved over 3 days to 3 weeks after clinical onset, resulting in complete recovery without any neurological deficit in all patients. Lumbar puncture and drainage of haematoma seemed to have the more or less immediate effect of pain relief, however, symptoms and signs disappeared completely within 2 weeks in all patients regardless of the procedure. Follow up MR study in a patient confirmed complete resolution of the haematoma (Fig. 2). In this case 4 ml of hemorrhagic fluid was drained by lumbar puncture on the day after the onset of symptoms. It is thought that lumbar drainage could remove the subdural haematoma partially and most of haematoma resolved spontaneously. Discussion Nontraumatic spinal subdural haematoma was first described more than 50 years ago [3] and a recent extensive review of the literature revealed that more than half of nontraumatic spinal subdural haematomas were associa-

ted with a coagulopathy and also iatrogenic factors such as a diagnostic spinal puncture or peridural anesthesia contributed in two thirds of the cases with spinal subdural haematoma and coagulopathy [1]. Only two cases related with hydrocephalus and ventriculoperitoneal shunt were identified in that review. Recently a few more cases related to surgical procedures for intracranial lesions were reported. Those are the cases with ventriculoperitoneal shunt for von Recklinghausen’s disease associated with aqueductal stenosis [4], craniotomy for clipping an aneurysm [6], and resection of an acoustic neuroma [3]. In those cases, downward migration of the cranial subdural haematoma was commonly suggested as a pathogenetic mechanism of spinal subdural haematoma. Our six cases comprise only 0.002% of all intracranial surgical procedures at our institution excluding traumatic lesions. However, the actual incidence may be higher because cases with mild symptoms over a short duration may have been missed without confirmative imaging diagnosis. Our observation of the cases supports the hypothesis suggested by the authors of previous case reports that spinal subdural haematoma is due to migration of haematoma from the cranial subdural space. Haematoma was present consistently in the lower part of the spinal canal with only one variation in respect of the rostral extent of the legion more likely to be in the dorsal part of the dural sac. These findings suggest the effect of gravity on distribution of the haematoma. Signal intensity of haematoma on MR imaging suggested the subacute or chronic stage. This means that haematoma is in the liquefied form which can migrate through the

414

potential subdural space and is not from de novo bleeding in the spinal canal which will result in acute blood clot. Onset of symptoms occurred between 4 and 9 days after the initial cranial surgery when the patients were beginning to freely ambulate and this was already pointed out by other authors [6]. It is thought that decompression of the intracranial space with removal of large amounts of tissue or CSF during any operation makes the cranial space vulnerable to the formation of subdural hemorrhage or fluid collection and its migration results in spinal subdural haematoma. Though our findings are consistent with this hypothesis, alternatively CSF hypotension in the spinal subdural space may also be a plausible explanation. Formation of intracranial subdural haematoma after excessive drainage of cerebrospinal fluid (CSF) is a well-known complication of ventriculoperitoneal shunt [5] and the spinal subdural haematoma may be produced by the same mechanism. After all, whatever the exact mechanism is, the important factor for haematoma formation should be decompression of intracranial space during surgery. It is supported by the fact that temporal lobectomy was the initial cranial procedure in four among six patients described here in whom the temporal horn of the lateral ventricle was routinely opened and large amounts of CSF had typically flowed out. In general acute spinal subdural haematoma is a potentially dangerous condition and may have disastrous consequences. Though there are several case reports of spinal subdural haematoma with spontaneous resolution [2], early surgical treatment is recommended when the neurological state progressively deteriorates and conservative treatment is an option for only a few selected patients with minimal neurological impairment [1]. In contrast, all six cases in this report showed a benign course with complete recovery within 3 weeks without permanent neurological deficit. We did lumbar spinal puncture and drainage of haematoma in four patients. However, drainage of haematoma does not seem to contribute to the outcome because all the patients recovered completely in the end. The benign clinical course may be explained also based on the pathogenetic hypothesis. Because there is no actively bleeding site within the spinal canal, the collection of haematoma does not give significant pressure to the neural structures. If haematoma accumulation is mainly due to the effect of gravity causing downward migration from the intracranial space, it is possible that simply placing patients in a lying position after symptom development may prevent further haematoma accumulation before serious neurological deterioration occurs. The main distribution of haematoma over

J. I. Lee and S. C. Hong

the lower lumbar level where the intradural space is wide and does not contain the spinal cord but only the cauda equina may be another reason for the mild neurological features and favorable outcomes.

Conclusions In our series of six patients, we document that spinal subdural haematoma is a possible complication of any kind of cranial surgery accompanied by decompression of the intracranial space. It should be considered when a patient after craniotomy shows signs and symptoms of cauda equina compression. Unlike spontaneous spinal subdural haematoma with common precipitating factors such as coagulopathy or spinal puncture, spinal subdural haematoma following cranial surgery is a benign condition which resolves spontaneously without surgical intervention.

References 1. Domenicucci M, Ramieri A, Ciappetta P, Delfini R (1999) Nontraumatic acute subdural haematoma. Report of five cases and review of the literature. J Neurosurg (Spine 1) 91: 65–73 2. Kulkarni AV, Willinsky RA, Gray T, Cusimano M (1998) Serial magnetic resonance imaging findings for a spontaneously resolving spinal subdural haematoma: case report. Neurosurgery 42: 398–401 3. Marx SV, Roberson DW, Coates G, Langman AW (1999) Spinal subdural haematoma after resection of an acoustic neuroma. Otolaryngology Head Neck Surg 120: 540–542 4. Ohta H, Ottomo M, Nakamura T (2001) A case of the spinal subdural haematoma formation following ventriculoperitoneal shunting for von Recklinghausen’s disease associated with aqueductal stenosis. No Shinkei Geka 29: 53–57 5. Samuelson S, Long DM, Chou SN (1972) Subdural haematoma as a complication of shunting procedures for normal pressure hydrocephalus. J Neurosurg 37: 548–551 6. Shimizu S, Tachibana S, Maezawa H, Fujii K, Kan S (1999) Lumbar spinal subdural haematoma following craniotomy – case report. Neurologia Medico-Chirurgica 39: 299–301 7. Silver JM, Wilkins RH (1991) Spinal subdural haematoma formation following ventriculoperitoenal shunting for hydrocephalus: case report. Acta Neurochir (Wien) 108: 159–162

Comments The authors report 6 cases of spinal subdural haematoma following intracranial surgery. This report is interesting because it addresses an uncommon complication which many neurosurgeons would not suspect. Previous reports in the literature, summarized also by the authors, mainly concern single case reports. This manuscript makes a fairly strong case that the occurrence of a spinal subdural haematoma following intracranial surgery may not be as rare a complication as presumed. The clinical importance of the findings may however not be so great as all cases show a benign course with symptoms and signs disappearing completely within 2 weeks regardless of treatment. Operative procedures were never necessary in any of the reported cases. It is important to recognize that in the

Spinal subdural haematoma as a complication of cranial surgery cases described, lumbar puncture or surgical procedures around the spine were not performed prior to the development of symptoms and detection of spinal subdural haematomas. Further none of the patients showed abnormal findings in pre- and postoperative laboratory tests for coagulopathy. It remains difficult to identify a clear cause for the occurrence of a spinal subdural haematoma following intracranial surgery. The hypothesis stated by the authors that the spinal subdural haematoma may have migrated from the cranial subdural space is difficult to accept. Alternatively CSF hypotension in the spinal subdural space may be a more plausible explanation. A. Maas The authors describe a complication of intracranial surgery, that might be more frequent than generally expected. Although this is a retrospective study, the collection of 6 cases, reflecting about 0.2% of all craniotomies performed during a certain time interval, is unique in comparison to the rare case reports, that have been published until now. It is interesting to notice that 4 of the 6 patients were treated for epilepsy and underwent a temporal lobectomy with opening of the temporal horn

415 of the lateral ventricle. As a consequence, the authors speculate on the pathogenesis of the spinal subdural haematoma. They consider the time interval of 4–9 days between the operation and the onset of symptoms as well as the liquefaction of the haematoma to be arguments in favour of a migration of the haematoma from the intracranial compartment down to the lumbar subdural space. However, it seems to me that nothing has been proven yet with regard to this pathogenesis. This postcraniotomy subacute spinal subdural haematoma seems to be a benign lesion. However, 4 of the 6 cases were treated by a lumbar puncture. One might speculate on what the evolution might have been if all 6 patients would have been followed purely conservatively. More and prospective research on the clinical circumstances of this interesting side-effect of a number of craniotomies is necessary. J. Goffin Leuven

Correspondence: Jung-Il Lee, M.D., Department of Neurosurgery, Samsung Medical Center, 50 Ilwon-Dong, Kangnam-Gu, Seoul 135710, Korea. e-mail: [email protected]