Cerebellar Hemorrhage after Supratentorial Surgery for Treatment of Epilepsy: Report of Three Cases Elza Marcia Yacubian, M.D., Marli M. de Andrade, M.D., Carmen Lisa Jorge, M.D., Rosa Maria Valerio, M.D. Epilepsy Program of the Hospital das Clinicas da Faculdade de Medicina da Universidade deSao Paulo, Sao Paulo, Brazil OBJECTIVE AND IMPORTANCE: W e report three cases of cerebellar hemor­ rhage complicating supratentorial craniotomies for the treatment of epi­ lepsy. In a literature review, we identified only four similar cases of cerebel­ lar hemorrhage after temporal lobectomy for the treatment of epilepsy. CLINICAL PRESENTATION AND RESULTS: Three young and otherwise healthy patients underwent frontal, occipital, and temporal resections for the treat­ ment of refractory epilepsy. The hemorrhage manifested as peduncular tremor, ataxia, and decerebrate posturing presenting early in the postoperative period. The diagnosis was established by computed tomography and/or mag­ netic resonance imaging. Benign outcomes were observed for all patients. CONCLUSION: Based on the available data, it is our opinion that brain dislocation resulting from excessive intraoperative cerebrospinal fluid drainage is a possible mechanism for this rare complication of supraten­ torial craniotomy. The overdrainage seems to be less hazardous when the procedure is performed for the removal of space-occupying mass lesions. In contrast, the resection of nonexpanding tissues, such as in lobectomies for the treatment of epilepsy, may be an additional risk factor, because the incidence of this complication seems to be higher in these situations. (Neurosurgery 45:1 59-1 62, 1999) Keywords: Cerebellar hemorrhage, Epilepsy surgery, Supratentorial craniotomies postoperative parenchymal hemor* rhage at a remote site in the brain is a well-recognized, albeit rare, complica­ tion of craniotomies. In 1987, Konig et al- (6) reported four cases of postopera­ tive cerebellar hemorrhage in a series of 1350 supratentorial craniotomies. In 1^88, Kalfas and Little (5), in a survey of 9992 intracranial procedures performed during a 11-year period, found only t°ur cases of postoperative parenchy­ mal hemorrhage remote from the surgical site. None of these cases involved cerebellar hemorrhage after a supraten­ torial craniotomy. In 1996, Toczek et al. (9) reP°rted four cases of cerebellar hemor­ Downloaded from https://academic.oup.com/neurosurgery/article-abstract/45/1/159/2846114 by Tufts University user on 17 June 2018 rhage complicating supratentorial sur­ gery for the treatment of epilepsy. All cases followed temporal lobectomies, oc­ curring in 4.9% of their series of temporal lobectomies performed in a 5-year period. We report three cases of similar compli­ cations, occurring after frontal, temporal, and occipital resections for the treatment of refractory epilepsy. Examination The neurological examination was unremarkable. Neuropsychological test­ ing revealed a verbal intelligence quotient of 65 and a performance intelligence quo­ tient of 71. The electroencephalograms showed basal left temporal epileptiform discharges, and mesial temporal seizure onset w as d o cu m en ted by v id e o ­ electroencephalography. Magnetic reso­ nance imaging (MRI) showed left mesial temporal sclerosis. Lnterictal single pho­ ton emission computed tomography re­ vealed left temporal hypoperfusion. The patient was normotensive, and there was no laboratory evidence of coagulopathy. The patient was receiving phenytoin (PHT). Operation On March 3, 1993, with the patient supine and his head turned to the right, a standard temporal craniotomy for a left 4.5-cm anterior temporal lobectomy with amygdalohippocampectomy was per­ formed without complications. The tempo­ ral hom was opened. The patient under­ went drainage via subgaleal suction for 2 days after surgery. Histopathological analysis confirmed mesial temporal sclerosis. Postoperative course The patient did not awaken from an­ esthesia in the immediate postoperative period and exhibited decerebrate pos­ turing. Postoperative blood pressure and coagulation test results were nor­ mal. Computed tomographic scans ob­ tained 6 hours after surgery showed hemorrhage involving the vermis and both cerebellar hemispheres (Fig. 1). The A 29-year-old man presented with a history of complex partial seizures since he was 10 years of age. FIG U RE 1. Patient 1. Extensive hemorrhage involving the vermis and both cerebellar hemispheres was demonstrated by postoperative computed tomography. N e u ro su rg e ry , Vol. 45 , No. 7, July 1999 159 CASE REPORTS Patient 1 160 Yacubian et al patient was treated conservatively and exhibited gradual improvement during a 4-day period. Cerebellar ataxia per­ sisted for approximately 10 days. The patient has been seizure-free since the operation. Patient 2 A 10-year-old boy presented with medically refractory epilepsy character­ ized by up to 20 hypermotor seizures per day since he was 19 months of age. Examination The neurological examination results were unremarkable, and the intelligence quotient was in the borderline range. Video-electroencephalographic moni­ toring showed epileptiform discharges and seizure onset in the frontal lobes. MRI suggested mesial right frontal focal cortical dysplasia in the mesial frontal lobe on the right. Ictal single photon emission computed tomographic scans showed hyperperfusion in the same re­ gion. The patient received sodium val­ proate (VPA), carbamazepine (CBZ), and clonazepam before surgery. The plasma level of VPA was 42 /xg/ml, and the CBZ level was 7.8 jag/ml. The pro­ thrombin time, activated partial throm­ boplastin time, thrombin time, and platelet count at the time of the preop­ erative laboratory assays were normal. Operation The mesial frontal lesion was resected via a right frontal craniotomy per­ formed on March 23, 1993, with the pa­ tient in the supine position and his head slightly extended. Surgery was com­ pleted without complications. The ven­ tricle was not opened, and no drain was placed. The patient received PHT dur­ ing the procedure. Histological exami­ nation of the surgical specimen (5 X 4 X 1 cm) revealed Taylor's type cortical dysplasia. Postoperative course Shortly after the patient recovered from anesthesia, bilateral peduncular tremor was observed; the tremor per­ sisted for 3 days. The patient remained fully alert, without other neurological signs. His blood pressure remained within the normal range, and there was no laboratory evidence of coagulopathy. The serum PHT level was within the therapeutic range. Computed tomo­ graphic scans obtained 12 hours after surgery showed bilateral vermian hem­ orrhage. MRI scans confirmed two sym­ metrical nodular hemorrhage areas in the superior cerebellar peduncles (Fig. 2). One week later, the neurological ex­ amination results were normal. The pa­ tient has been seizure-free since the operation. Patient 3 A 20-year-old man experienced fre­ quent left hemifield visual paroxysms followed by complex partial seizures since the age of 8 years. At admission, the patient was receiving CBZ, VPA, vigabatrin, and clobazam. The CBZ plasma level was 12 /xg/ml, and the VPA level was 56 /xg/ml. Examination The neurological examination results were normal. Electroencephalograms showed continuous epileptiform dis­ charges, which were partially abolished by eye-opening, occurring over the temporoparieto-occipital areas and be­ ing more prominent on the right, where fast activity was seen during the sei­ zures. The MRI results were normal, and interictal and ictal single photon emission computed tomographic scans showed hyperperfusion of the posterior areas of the right hemisphere. Preoper­ ative laboratory values, including pro­ thrombin time, activated partial throm­ boplastin time, thrombin time, and platelet count, were normal. Operation On April 19,1994, the patient underwent extensive electrocorticography-guided re­ section in the right temporoparietoFIGURE 2. Patient 2. This coronal T1weighted MRI scan shows hemorrhage as a bilateral nodular hypersignal lesion in both superior cerebellar peduncles. Neurosurgery, Vol. 45 , No. /, July 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/45/1/159/2846114 by Tufts University user on 17 June 2018 occipital region. The patient was placed in a left lateral decubitus position with his head slightly turned to the left. The procedure was uneventful, and the pa­ tient received PHT intraoperatively. The right occipital horn of the lateral ventri­ cle was opened. A subgaleal suction drain was left in place for 4 days. Histopathological analysis showed mild gliosis. Postoperative course Postoperatively, the patient was fully conscious and experienced mild head­ ache, dizziness, and one episode of vomiting. The examination showed left homonymous hemianopsia, mild gait ataxia, and nystagmus. The PHT levels were within the therapeutic range. In the postoperative period, the patient's systolic blood pressure ranged between 130 and 150 mm Hg, with one pressure peak of 180/80 mm Hg. Laboratory tests did not indicate coagulopathy. Com­ puted tomographic scans obtained 12 hours after surgery revealed extensive hemorrhage in both cerebellar hemi­ spheres, which was confirmed by MRI (Fig. 3). The evolution was benign. One week later, a neurological examination revealed the expected left homonymous hemianopsia and no cerebellar signs. The patient adapted to the visual-field deficit within a few months and has been seizure-free since the operation. D ISCUSSIO N Postoperative brain hemorrhage at sites distant from the craniotomy is a FIGURE 3. Patient 3. A, postoperative MRI scan showing extensive hemorrhage involving both cerebellar hemispheres. B, T1 -weighted MRI scan (sagittal view) showing the extent of the parieto­ occipital resection and the area of hemorrhage in the cerebellum. Cerebellar Hemorrhage after Epilepsy Surgery rare occurrence (1, 6, 8-11, 13). The pre­ cise mechanisms underlying these bleeding episodes are not known. Com ­ mon predisposing factors for cerebral hemorrhage, such as arterial hyperten­ sion, underlying occult brain lesions, or coagulopathy, were lacking in most of the reported cases. The mechanisms for infratentorial hemorrhage after infratentorial craniotomies, infratentorial hem­ orrhage after supratentorial cranioto­ mies, and supratentorial hemorrhage after infratentorial craniotomies may not be the same. Some authors suggest that multiple mechanisms may play a role in these complications (6, 9,13). The single most common finding in the lit­ erature is the fact that the bleeding pre­ sents early in the postoperative course, suggesting that it develops during or soon after surgery (1, 5, 8, 9). Patients subjected to posterior fossa procedures in the sitting position may develop supratentorial hemorrhage be­ cause of downward displacement of the brain and disruption of subcortical ves­ sels or cortical bridging veins. When the patient lies supine, hyperperfusion may cause intracerebral hemorrhage. The ma­ jority of the supratentorial hemorrhagic episodes after infratentorial craniotomies that were reviewed by Brisman et al. (1) were observed in patients who were sur­ gically treated in the sitting position, al­ though Kalfas and Little (5), who ana­ lyzed the incidence of postoperative hemorrhage in the largest published se­ ries, concluded that the position of the patient is of little importance. The mechanisms underlying posterior fossa/cerebellar hemorrhage after su­ pratentorial craniotomies are more dif­ ficult to understand. In 1996, Brisman et al- (1) found 13 cases in a thorough review of the data published between 1978 and 1993, and they added 3 other cases of their own. In the same year, Papanastassiou et al. (8) reported 5 ad­ ditional cases. These data (1, 8) showed that the hemorrhage occurred after frontotemporal craniotomy for aneurysm clipping in 11 of 21 cases and after sur­ gical treatment of parasellar tumors in a cases. In 1996, Toczek et al. (9) reported four cases of cerebellar hemorrhage complicat­ es temporal lobectomies for treatment of ePilepsy, which represented 4.9% of their 5-year series of temporal lobectomies. Our three cases included temporal, frontal, and occipital lobectomies for treatment of epilepsy and represented 3.5% of the lo­ bectomies or corticectomies performed during the previous 5 years. These rates are higher than the 0.3 to 0.5% incidence of cerebellar hemorrhage after supraten­ torial craniotomy reported by Konig et al. (6) and by Papanastassiou et al. (8). The overall incidence of remote hemorrhage in a survey of 4992 intracranial proce­ dures published by Kalfas and Little (5) was 0.15%. That series included subdural and extradural hematomas. The currently available data suggest that there is a greater likelihood of post­ operative hemorrhage in the cerebellum after a supratentorial craniotomy when cerebrospinal fluid has been extensively drained during surgery. This condition occurs, for example, when the ventri­ cles, basal cisterns, or sylvian fissure is widely opened (1, 6, 8, 9). Postoperative drainage may provide an additional risk. Dacey (2), commenting on the se­ ries reported by Brisman et al. (1), sug­ gested that medial sylvian/tentorial ap­ proaches to the basal cisterns might disturb the venous drainage of the cer­ ebellar vermis. We think that mechanical shifting of the cerebellum, resulting from cerebro­ spinal fluid drainage, is the mechanism most frequently involved in tearing of the superior vermian veins or their trib­ utaries or increases in their transmural pressure (6, 13), leading to hemorrhage. The overdrainage seems less hazardous when the procedure is performed for the removal of space-occupying lesions (6, 13). We think that in these situations the displacement of the brain back to its normal position rectifies the disturbed venous circulation, rather than worsen­ ing it. The resection of nonexpanding lesions, such as in lobectomies for the treatment of epilepsy, causes brain dis­ location. This could explain the higher incidence of this complication in epi­ lepsy series. Other mechanisms men­ tioned in the literature seem to be less important in cerebellar hemorrhage af­ ter supratentorial craniotomies. All patients in our series were young and otherwise healthy. The isolated post­ operative hypertensive peak recorded for Patient 3 is more likely a consequence N e u ro su rg e ry ; Vol. 45, No. 7, July 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/45/1/159/2846114 by Tufts University user on 17 June 2018 161 than the cause of the intracranial bleed­ ing. At the time of this recording, the pa­ tient already showed signs of intracranial hypertension. Administration of VPA may eventually increase the risk of surgi­ cal bleeding (3, 4) but is unlikely to be the cause of the bleeding in our patients. Ward et al. (12) verified that preoperative VPA administration does not increase blood loss during temporal lobectomies. Postoperative cerebrospinal fluid drainage from epidural/subgaleal suc­ tion drains may play an additional role. However, in both our series and other published series, the placement of a drain was not always performed (1). Ex­ treme rotation or flexion of the neck, leading to hemorrhage (1, 8) attributable to venous obstruction, was not seen in our series. Head positioning that might jeopardize jugular blood flow could be promptly detected by the surgeon on the basis of brain swelling and easily corrected by appropriate head reposi­ tioning early in the operation. Our pa­ tients experienced a benign course and full recovery. No single factor could predict a benign outcome (7). ACKNOW LEDGM ENT We thank Dr. Milton K. Shibata for sug­ gestions and review of the manuscript. Received, August 11, 1998. Accepted, February 26, 1999. Reprint requests: Elza Marcia Targas Yacubian, M.D., Department of Neurology, Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Av. Dr. Eneas de Carvalho Aguiar 255, Room 5011, 05403.000 Sao Paulo, Sao Paulo, Brazil. REFERENCES 1. Brism an M H , B ederson JB, Sen C N , G e rm a n o IM , M o o r e F, Post K D : Intracerebral h e m o r ­ rh age o cc u rrin g re m o te fro m the c ra n io to m y site. N e u r o s u r g e r y 39:1114-1122, 1996. 2. D a cey RG Jr: Intracerebral h e m o r rh a g e o cc u r­ rin g re m o te fro m the c ra n io to m y site. N e u r o ­ s u rg e ry 39:1121, 1996 (c o m m e n t). 3. D e lg a d o MR, R iela AR, M ills J, B ro w n e R, Roach ES: T h r o m b o c y to p e n ia s e c o n d a ry to h igh v a lp ro a te le v e ls in ch ild ren J C h ild N e u r o l 9:311-314, 1994. w ith e p ile p s y . Yacubian et al. 162 4. GidaJ B, Spencer N , M a ly M , Pitterle M , W illiam s E, 12. W a r d M M , Barbaro N M , L axer K D , R a m p il IJ: Collins M , Jones J: Valproate-mediated disturbances P re o p e ra tiv e v a lp ro a te ad m in istra tion d o e s not o f hemostasis: Relationship to dose and plasma con­ increase b lo o d loss d u rin g tem p o ra l lo b e c to m y . centration. N e u ro lo g y 44:1418-1422, 1994. E p ile p s ia 37:98-101, 1996. 5. Kalfas IH , Little JR: Postoperative hem orrhage: A 13. Y o s h id a S, Y o n e k a w a Y, Y am ash ita K, Ihara I, s u rv e y o f 4992 intracranial p roced u res. N e u r o ­ M o r o o k a Y: C ereb e lla r h e m o rrh a g e after su pra­ s u rg e ry 23:343-347, 1988. tentorial 6. K onig A, Laas R, Herrmann H-D: Cerebellar haem­ cra n io to m y: R e p o rt of three cases. N e u r o l M e d C h ir ( T o k y o ) 30:738-743, 1990. orrhage as a complication after supratentorial crani­ otomy. Acta Neurochir (W ien ) 88:104-108, 1987. 7. M e la m e d N , S atya-M u rti S: C ereb e lla r h e m o r­ rhage: A r e v ie w and reappraisal o f b en ign cases. A rc h N e u r o l 41:423-428, 1984. 8. Papanastassiou V, Kerr R, A d am s C: Contralateral cerebellar hemorrhagic infarction after pterional cra­ niotomy: Report o f five cases and review o f the literature. Neurosurgery 39:841-852, 1996. 9. T o c z e k M T , M o rre ll MJ, S ilv e rb e rg G A , L o w e G M : C ereb ella r h em o rrh a g e c o m p lic a tin g te m ­ p oral lo b ecto m y. J N e u r o s u r g 85:718-722, 1996. 10. van Calenbergh F, Goffin J, Piets C: Cerebellar hem ­ orrhage complicating supratentorial craniotomy: Report o f tw o cases. Surg N eu rol 40:336-338, 1993. 11. W a g a S, Shim osaka S, Sakakura M : Intracere­ bral h em o rrh a g e rem ote fro m the site o f initial neurosurgical procedure. N e u ro s u rg e ry 13:662665, 1983. COMMENTS This reports adds more cases to the growing number of hemorrhagic epi­ sodes distant from the site of surgery. We are convinced that this apparently rare complication is more frequent but has not been adequately reported. In our own series of more than 850 resec­ tions for epilepsy treatment, there were two cases of cerebellar hemorrhage after temporal lobe surgery, with excellent postoperative recovery. Interestingly, the computed tomographic scans for our pa­ tients were very similar to the results pre­ Symptomatic Spinal Epidural Lipomatosis after Local Epidural Corticosteroid Injections: Case Report David I. Sandberg, M.D., Michael H. Lavyne, M.D. Division of Neurosurgery, Weill Medical College of Cornell University, New York Presbyterian Hospital-Cornell Campus, New York, New York OBJECTIVE AND IMPORTANCE: Spinal epidural lipomatosis, which causes symptomatic compression of neural elements, is a well known but un­ common complication of Cushing's syndrome. Spinal epidural lipomato­ sis has been reported frequently in association with chronic systemic corticosteroid therapy, but thus far only one case has been attributed to local epidural corticosteroid injections. CLINICAL PRESENTATION: W e report another case of symptomatic spinal epidural lipomatosis after epidural corticosteroid injections. This is the first such case documented by magnetic resonance imaging and con­ firmed with surgical exploration. INTERVENTION: The patient's symptoms of neurogenic claudication resolved after lumbar laminectomy in the region of previous corticosteroid injections and the removal of epidural fat, which was compressing the thecal sac. CO NCLUSIO N: This case should alert clinicians that epidural lipomatosis, which causes symptomatic thecal sac compression, is a possible compli­ cation, not only of systemic glucocorticoid therapy, but also of local epidural corticosteroid injections. (Neurosurgery 45:162-165, 1999) Key words: Corticosteroid injections, Epidural lipomatosis, Lumbar neurogenic claudication Neurosurgery, Vol. 45, No. 1, July 1999 Downloaded from https://academic.oup.com/neurosurgery/article-abstract/45/1/159/2846114 by Tufts University user on 17 June 2018 sented in Figures 1 and 3A of this article. We think that this may be primary hem­ orrhaging into one of the fissures that only later extends into the parenchyma of the cerebellum. Valproate-induced coagu­ lation disorders are routinely treated in our service with preoperative administra­ tion of desmopressin, so clotting prob­ lems are unlikely. Clotting disorders seem unlikely to provide an explanation for hemorrhage distant from the surgical site in any event, because hemorrhage pri­ marily requires a leak in the vessel. Un­ fortunately, the authors were unable to identify clear reasons why these hemor­ rhagic episodes occurred; therefore, this case report makes no further contribution to our understanding of the pathogenesis of this complication. Dirk Van Roost Johannes Schramm Bonn , Germany pinal epidural lipomatosis causing neu­ rological deficits has been most fre quently reported in the setting of Cushing's syndrome. The majority of cases have been reported in association with exoge­ nous systemic corticosteroid treatments for a variety of medical conditions (4), including rheumatoid arthritis (juvenile and adult) (2, 13, 31), radiation pneumo­ nitis (28), Graves's ophthalmopathy (19), dermatomyositis (5), asthma (8,14, 22,24, 25), sarcoidosis (22), Crohn's colitis (22), periarteritis nodosa (10), systemic lupus erythematosus (12, 15), and organ trans­ plantation (8,10 ,1 1 ,1 6 ,1 8 , 30). Symptom­ atic spinal epidural lipomatosis also has been attributed to increased systemic glu­ cocorticoid levels caused by Cushing's disease (6, 19, 27), adrenal tumors (20), and anabolic steroid use by bodybuilders (9). Idiopathic cases of spinal epidural li­ pomatosis in the absence of Cushing's syndrome also have been reported rarely, primarily in obese individuals (3, 17, 21, 23, 26, 29). Spinal epidural lipomatosis second­ ary to exogenous administration of glu­ cocorticoids has been reported almost exclusively in association with systemic treatments. Only one case has previ­ ously been attributed to epidural cortiS