British Journal of Neurosurgery, 2014; Early Online: 1–5 © 2014 The Neurosurgical Foundation ISSN: 0268-8697 print / ISSN 1360-046X online DOI: 10.3109/02688697.2014.987215
ORIGINAL ARTICLE
Traumatic intracerebellar haematoma: To operate or not to operate? Viraat Harsh1*, Anand Prakash1*, James Marcellus Barry2 & Anil Kumar1 1Department of Neurosurgery, Rajendra Institute of Medical Sciences, Ranchi, Jharkhand, India, and 2Department of
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Neurosurgery, Baylor College of Medicine, Houston, TX, USA origin in the absence of posterior fossa subdural or epidural haematoma is even rarer, accounting for only 0.6%– 0.82%.7–9 About one-fourth of all posterior fossa traumatic lesions can be accounted to intra-parenchymal cerebellar haemorrhage.3,10 Only few case reports and series have been documented.10–25 Wright16 recorded 17 posterior fossa haematomas and reported only 6 to have intracerebellar haematomas over a period of 10 years. Though trauma is the obvious cause, specific mechanism for this type of injury remains unclear. The clinical presentation of cerebellar haematoma may be readily apparent with classic signs including ataxia, nystagmus and signs of increased ICP like headache, lethargy and nausea and vomiting. It is critical to note that patients may be totally asymptomatic before rapid deterioration of neurologic status.26 Owing to the rarity of pure intracerebellar haemorrhage cases and the risk of rapid decompensation, Schneider27 emphasized that all patients with occipital trauma should be ruled out for blood collection in both supratentorial and infratentorial spaces. Some patients present with a delayed traumatic intracerebellar haematoma several hours after the trauma and have been reported to be associated with an unfavorable course and poor prognosis in the past.28,29 However, with the increasing use of computed tomography (CT) for screening of all head trauma patients, increasing number of cases have been reported to be diagnosed even before the appearance of symptoms and have thus been treated in a timely manner.7 Treatment options for this pathology are still evolving. The current standard is conservative management for fully conscious patients with a superficial bleed of small diameter ( 3cm) and surgical evacuation for patients with large bleeds ( 3cm), 4th ventricle compression, associated paradural bleeding, or hydrocephalus.30 Outcomes for this type of injury have previously been studied, with initial GCS score ( 7 vs. 7) found to be the most predictive factor.11 Other previously proven predictive factors include haematoma location (superficial vs. deep), haematoma volume, degree of fourth ventricle and cistern deformation, and associated SAH.31 Since very few patients present with this pathology, documentation of every case is important. In this study we
Abstract Background. Pure cerebellar haematoma of traumatic etiology, without associated posterior fossa sub- or epidural haematomas is a rare entity and has been reported to have a poor outcome. We report 23 patients with traumatic intracerebellar haematoma. We sought to study the pattern of such presentations and assess the factors which could be associated with their outcome. Methods. A retrospective review of prospectively collected data for all patients who were admitted for the management of traumatic intracerebellar haematoma at Rajendra Institute of Medical Sciences, Ranchi, India provided data for the 23 consecutive patients admitted for aforesaid over a seven-year study period. Medical records, diagnostic imaging and operative notes were reviewed for all patients. We divided the patient pool in to two groups based on their GCS score at the time of presentation – Group A (GCS 7) and Group B (GCS 7). The association of different allied factors was studied and statistically analyzed. The relevant medical literature was also reviewed. Results. Most Group B patients were found to be associated with poor outcome at hospital discharge. The overall incidence of poor outcome in our study was 69.56%. GCS score at time of admission, allied supratentorial lesions, advanced age, condition of fourth ventricle and chest infection were found to be important factors which could be associated with poor outcome. Conclusion. Surgery in patients with the mentioned risk factors remains debatable and should be approached cautiously. Larger multiinstitutional and meta-analytic studies are required to study and statistically establish the factors which might be associated with poor outcome in these patients. An algorithm which may be used in the management of traumatic intracerebellar haematoma patients is proposed. Keywords: haemorrhage; intracerebellar haematoma; traumatic brain injury
Introduction Haematomas of the posterior fossa are by themselves uncommon and account for only 3.3–6% of all head injuries.1–6 Pure cerebellar haematoma of traumatic
*These authors contributed equally to this work. Correspondence: Anil Kumar MBBS, MS, MCh, Professor and Head, Department of Neurosurgery, Rajendra Institute of Medical Sciences, Bariatu, Ranchi, Jharkhand 834009, India. Tel: 91 943 111 5510. E-mail: [email protected] Received for publication 18 October 2012; accepted 2 November 2014
1
2 V. Harsh et al. endeavor to review the pattern of such cases that we encountered. We have also reviewed the relevant medical literature.
Objectives We conducted this retrospective study to analyze the pattern of traumatic intracerebellar haematoma in our institute and study the factors which could be associated with the outcome.
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Patients and methods We retrospectively screened and (where required) reviewed records of more than 19300 head injury patients who were admitted in our Department of Neurosurgery at Rajendra Institute of Medical Sciences, Ranchi, India between January 2005 and January 2012 and all those who were diagnosed with an intracerebellar haemorrhage were selected for the study. Data on patient age, gender, GCS score, mechanism of injury, CT scan findings, associated supratentorial lesions, management strategy and outcome were gathered and analyzed. Patients were divided in to two groups based on their level of consciousness assessed by the GCS score at the time of presentation – Group A with GCS score greater than 7 and Group B with GCS lesser than or equal to 7. This was done to study the possible factors which could be associated with the outcome of either group. The mechanisms of injury were broadly categorized as road traffic accident (RTA), fall from height and assault. CT scan findings included clot location, size (largest transverse diameter of clot, categorized as 3 cm or 3 cm, or clot volume categorized as 15 ml or 15 ml), other associated findings and status of the fourth ventricle (normal vs. compressed). Surgical treatment (sub-occipital craniectomy or supratentorial craniotomy) and insertion of ventriculoperitoneal (VP) shunt was documented. A pre-decided general management protocol was followed. An algorithm based on the management protocol used by us is proposed and is depicted in Fig. 1. In the absence of an operable supratentorial lesion, conservative management was considered safe and acceptable in patients having GCS 14 or 15 and with clot size having largest diameter less than 3 cm or volume less than 15 ml. Surgical decompression of posterior fossa via a sub-occipital craniectomy was undertaken for patient with deteriorating level of consciousness with clot size measuring greater than or equal to 3 cm in diameter or volume greater than or equal to 15 ml or compression of fourth ventricle causing hydrocephalus. VP shunt was inserted on an emergency basis in patients having cerebellar clots which were near the fourth ventricle and caused compression, thus resulting in acute hydrocephalus but clot size was not big enough ( 15 ml) to merit surgical clot evacuation. Supratentorial craniotomy was undertaken for patients having injury in the supratentorial compartment. Outcome was documented as favorable or poor based on the Glasgow Outcome Scale at the time of hospital discharge (GOS-HD). Good recovery was counted as favorable whereas
all other states (death, vegetative state, severe disability and moderate disability) were counted as poor outcome. Sixmonth follow-up reports of all patients who turned in for the follow-up were also analyzed, however most patients could not be followed up and hence we are unable to discuss data related to six-month follow-up.
Results Retrospective screening and review of medical records, imaging data and operative notes yielded a total of 24 patients who were admitted in our institute for the management of intracerebellar haematoma over a period of seven years. There was 1 patient with firearm injury in occipital region who died prior to investigations and was therefore not included in our study, leaving us with 23 patients. The data related to 23 cases is shown in Table I. Male to female ratio in Group A (GCS score 7) was 2.3:1 and in Group B was 5.5:1. The age of the patients ranged from 6 to 72 years, mean age being 25.7 years in Group A and 38.5 years in group B. Overall mean age was 32.9 years. Most of the patients in our series (13/23, 56.5%) were between the age of 21 and 40. The rest of the patients were almost equally distributed between all age groups (Table II). More patients had poor GCS, i.e. lesser than or equal to 7 (13/23, 56.5%) and fell in Group B. The most common mode of injury was found to be RTA (60.86%) in either group – with 10 out of 14 patients in Group B (71.42%). CT scan records of the patients revealed clot size greater than or equal to 3 cm in 39% (9/23) of patients, most of whom (7/9 77%) were Group B patients. Also six patients out of 14 (42.85%) who had clot size lesser than 3 cm were Group B patients and had poor GCS score because of associated supratentorial lesions. Supratentorial lesions were present in 65.21% (15/23) and most of them (10/15, 66.6%) fell under Group B. Compression of 4th ventricle due to clot or surrounding edema was seen in 56.52% of patients. Two of the 13 cases having compression of 4th ventricle were from Group A. Two of the 10 having no compression of 4th ventricle on initial CT scan were from Group B at the time of admission. All patients had occipital bone fracture and 10 of them also had diffuse axonal injury (DAI). Seven of the patients with DAI were from Group B. Other associated pathology as seen on the CT scan is shown in Table III. Sub-occipital craniectomy with lax duroplasty was done to evacuate clot and brain debris in all the 7 Group B patients having clot size greater than or equal to 3 cm. In 40% cases (6/15) that had supratentorial lesion, extradural or subdural clot was surgically evacuated. VP shunt was inserted in lateral ventricle in three cases to tide over acute rise in intracranial pressure (ICP) due to dilation of ventricles. The overall incidence of poor outcome in our series at the time of hospital discharge was 69.56%. This incidence was reported to be 44.4% by d’Avella et al.11 Poor outcome was higher, 84.6% (11/13) in Group B as against 50% (5/10) in Group A. Mortality was 46.1% (6/13) in Group B and only 10% (1/10) in Group A.
Traumatic intracerebellar haematoma 3 TICH on CT scan Group A GCS > 7
Group B GCS ≤ 7
Clot size ≥ 3 cm or Clot vol. ≥ 15 ml
Clot size < 3 cm or Clot vol. < 15 ml
GCS ≤ 13
GCS> 13
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Clot size < 3 cm or Clot vol. < 15 ml
Clinical and radiological monitoring
No
Clinicoradiological stability and GCS Maintained
Conservative management
Also consider other pathologies as cause of low GCS and treat as per pathology eg. Supratentorial craniotomy for supratentorial hematomas
Extradural haematoma, Subdural haematoma or Brainstem compression present
Clinicoradiological deterioration and GCS decreasing
Yes
Further clinicoradiological or GCS score deterioration
Sub-occipital craniectomy + T/t of associated pathologies
If hydrocephalus is present, perform a VPS or place EVD
Fig. 1. Algorithm for management of traumatic intracerebellar haematoma. CT Computed Tomography; EDH Extradural Haemorrhage; EVD External ventricular drain; GCS Glasgow Coma Scale score; SDH Subdural Haemorrhage; T/t treatment; vol volume; VPS Ventriculo peritoneal shunting. Table I. Demographics and pattern of intracerebellar haematoma. Sl. Factors studied 1 2 3 4
5 6
7 8
9
Number of patients Mean age Male: Female ratio GCS score at admission
3–7 8–13
Group A
Group B
Total
10 25.7 2.3:1
13 38.53 5.5:1 13
23 32.95 3.6:1 13 7
7
14-15 Mechanism of injury RTA Fall from height Assault CT scan findings Cerebellar clot 3cm size 3cm Supratentorial lesion Fourth ventricle Normal Compressed Surgical treatment Sub-occipital craniectomy Supratentorial craniotomy V-P Shunt Outcome at hospital discharge Favorable (based on GOS-HD) Poor
3 4 2 4 8 2 5 8 2 2 2 2 5 5
10 3 0 6 7 10 2 11 7 4 1 2 11
Mortality
1
6
3 14 5 4 14 9 15 10 13 9 6 3 7 16 7
T, Computed Tomography; GOS-HD, Glasgow Outcome Scale at the time of hospital discharge; GCS, Glasgow Coma Scale C score; RTA, road traffic accident; V-P, ventriculo peritoneal
4 V. Harsh et al. Table II. Age distribution of patients with intracerebellar haematoma. Age group (years) Number 1–10 11–20 21–30 31–40 41–50 51–60 60
2 3 7 7 1 2 1
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Discussion The unusually low incidence of cerebellar haematomas at our institution, 0.12% of all head injuries as compared to the normal incidence of 3.3–6% is notable. This deviation from the normal may have several causes. The most readily apparent is the long patient delivery time associated with a developing nation like India. This prolonged time to care would result in the death of patients who might have otherwise been picked up if admitted earlier. The propensity for males to experience this type of injury can also be explained by India’s developing nation status. In accordance with social customs, males are more commonly involved in outdoor activities than females and this reflects as higher incidence of intracerebellar trauma in males in our series. Though trauma is the obvious cause, specific mechanism for this type of injury remains unclear. Takeuchi et al.31 created a classification system for the types of trauma causing cerebellar haematomas. Their three classes included coup injuries, countercoup injuries and acceleration-deceleration injuries. This classification system proved useful in predicting the site of haematoma, with coup injuries causing only superficial bleeds while countercoup injuries resulted in deep cerebellar bleeds. In our study, RTA was the most common and most damaging mechanism of injury. This is likely attributable to other associated injuries caused by RTAs. Both high-speed and low-speed accidents produced intracerebellar haematoma with equal frequency but low speed accidents resulted in less severe associated injuries which resulted in more favorable outcome. Fall from height was more common amongst children while assault was more commonly associated with women and the elderly. All patients having history of assault fell under Group A. Interestingly, it seems that direct blows to the occiput are less likely to cause significant haemorrhage than acceleration/deceleration injury involved with RTA or a fall from height. In accordance with other studies on this topic, CT scan was extremely valuable in predicting patient outcomes. Clots greater than or equal to 3 cm in diameter, associated Table III. CT scan findings of patients with intracerebellar haematoma. Number of patients CT scan finding DAI EDH Acute SDH Skull fracture (other than occipital) Occipital bone fracture
Group A 3 1 0 7 10
Group B 7 2 4 0 13
Total 10 3 4 7 23
AI, Diffuse Axonal Injury; EDH, Extradural Haemorrhage; SDH, Subdural D Haemorrhage
supratentorial injuries and fourth ventricle compression were all shown by our study to predict a poor outcome. It makes sense that patients having associated lesion on CT scan sustained more extensive injury of brain and hence their chances of recovery depended in part on severity of other associated injuries as well. Outcomes were worse for patients in Group B. Such patients were critically injured and their brain function was already compromised at the time of admission. In accordance with previous studies on this topic our data showed GCS at initial presentation to be the most predictive clinical tool.11 In all patients undergoing surgical intracerebellar clot evacuation sub-occipital craniectomy was done and preferred over sub-occipital craniotomy as posterior fossa has less space for accommodation of any post-operative bleeding or post-operative edema. Some may note our use of VP shunts over external ventricular drains (EVDs) for certain patients experiencing hydrocephalus resulting from cerebellar haematoma. VP shunt was chosen over EVD placement because in our clinical setting EVDs have allegedly been found to more frequently result in infections as ventriculitis and meningitis three to four days post-EVD placement. However, patients developing acute hydrocephalus need drainage of CSF usually for a period of more than two weeks during which the clot resolves and perilesional edema subsides. There were several notable limitations of our study. First and most significant was the small sample size of 23 patients which reduces the statistical power of our study. This is however a result of the relative rarity of intracerebellar haematoma cases, as prior multi-institutional study11 could gather data for only 81 patients. Second, due to India’s developing status and overcrowded health system there were poor follow-up records making it difficult to draw any conclusions on the long-term outcome. Third, no attempt has been made to quantify the severity of associated non-neurological injuries in this group of patients. Considering many of these patients were involved in polytrauma situations it is reasonable to consider non-neurological trauma as an important component in their outcomes. Additionally, time from trauma to intervention is not included in our data set rendering us unable to comment how this critical factor affected outcomes. Finally, the retrospective character of this study creates the potential for observation and assessment bias.
Conclusion As previously indicated, traumatic intracerebellar haematomas are a rare and incompletely understood entity. In an attempt to study the factors which may be associated with poor outcome in intracerebellar haematoma cases we found that GCS score at the time of admission, related supratentorial lesion(s), advanced age, status of fourth ventricle and development of chest infection which interfered in tissue oxygenation, play an important role. Owing to the small number of patients and poor follow-up records we are unable to statistically associate these factors with poor outcome; however surgery in patients with the aforementioned negatively associated factors should be cautiously approached and a management protocol as depicted in
Traumatic intracerebellar haematoma 5 Fig. 1 must be followed. More multi-institutional studies should be considered in future to enhance pool size and more clearly elucidate the potential prognostic factors, both positive and negative, associated with traumatic intracerebellar haematoma.
Declaration of interest: The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
References 1. Anegawa S, Yoshimura K, Kawai K, Kuramoto S, Ichinomiya C. [Traumatic intracerebellar hematoma—a case report and the review of the literature (author’s transl)]. No Shinkei Geka 1979;7:911–5. 2. St John JN, French BN. Traumatic hematomas of the posterior fossa. A clinicopathological spectrum. Surg Neurol 1986;25: 457–66. 3. Tsai FY, Teal JS, Itabashi HH, Huprich JE, Hieshima GB, Segall HD. Computed tomography of posterior fossa trauma. J Comput Assist Tomogr 1980;4:291–305. 4. Zuccarello M, Andrioli GC, Fiore DL,et al. Traumatic posterior fossa haemorrhage in children. Acta Neurochir (Wien) 1982;62: 79–85. 5. Jamieson KG, Yelland JD. Extradural hematoma. Report of 167 cases. J Neurosurg 1968;29:13–23. 6. Pevehouse BC, Bloom WH, Mc KW. Ophthalmologic aspects of diagnosis and localization of subdural hematoma. An analysis of 389 cases and review of the literature. Neurology 1960;10:1037–41. 7. Nagata K, Ishikawa T, Shigeno T, et al. Delayed traumatic intracerebellar hematoma: correlation between the location of the hematoma and the pre-existing cerebellar contusion—case report. Neurol Med Chir (Tokyo) 1991;31:792–6. 8. Fisher RG, Kim JK, Sachs E Jr. Complication in posterior fossa due to occipital trauma; their operability. J Am Med Assoc 1958;167:176–82. 9. Arseni C, Maretsis M. Traumatic cerebellar haematoma associated with posterior cerebral fossa subdural haematoma. Psychiatr Neurol Neurochir 1972;75:113–5. 10. Karasawa H, Furuya H, Naito H, et al. Acute hydrocephalus in posterior fossa injury. J Neurosurg 1997;86):629–32. 11. d’Avella D, Servadei F, Scerrati M, et al. Traumatic intracerebellar hemorrhage: clinicoradiological analysis of 81 patients. Neurosurgery 2002;50:16–25; discussion 25–17.
12. Martin AJ, Thomas NW. Evolving traumatic cerebellar hematoma. Neurology 2001;57:1565. 13. Seida M, Ito U, Kito K, Tomida S, Inaba Y. [Traumatic cerebellar hematoma in children. Report of two cases]. Neurol Med Chir (Tokyo) 1986;26:318–22. 14. Papadakis N, Safran A, Ramirez L, Sujatanond M, Mark VH. Traumatic cerebellar hematoma without subdural hematoma. JAMA 1976;235:530–1. 15. Rothballer AB. Traumatic cerebellar hematoma in the newborn. Case report of operative removal with survival. J Neurosurg 1962;19:913–5. 16. Wright RL. Traumatic hematomas of the posterior cranial fossa. J Neurosurg 1966;25:402–9. 17. Cassinari V, Dorizzi A, Pauli P, Monolo L. [Expansive traumatic lesions of the posterior cranial fossa. On 4 cases of cerebellar contusion and laceration]. Minerva Neurochir 1967;11:230–9. 18. Hamasaki T, Yamaki T, Yoshino E,et al. [Traumatic posterior fossa hematoma]. No to shinkei = Brain Nerve. 1987;39:1083–90. 19. Koziarski A, Frankiewicz E. Medical and surgical treatment of intracerebellar haematomas. Acta Neurochir (Wien) 1991;110:24–8. 20. Pozzati E, Grossi C, Padovani R. Traumatic intracerebellar hematomas. J Neurosurg 1982;56:691–4. 21. Pozzati E, Piazza G, Padovani R, Gaist G. Benign traumatic intracerebellar hematoma. Neurosurgery 1981;8:102–3. 22. Sato K, Hinokuma K, Matsuzawa Y, et al. [Clinical study of traumatic cerebellar contusion]. No Shinkei Geka 1987;15:1285–9. 23. Sokol JH, Rowed DW. Traumatic intracerebellar haematoma. Surg Neurol 1978;10:340–1. 24. Tibbs PA, Goldstein SJ, Smithson JR Jr. Delayed traumatic intracerebellar hematoma. Surg Neurol 1981;16:309–11. 25. Tsubokawa T, Yamada J, Tomizawa N, et al. [Classification of traumatic intracerebral hematoma by repeated CT-scan and clinical course (author’s transl)]. Neurol Med Chir (Tokyo) 1979;19:1127–37. 26. Sokol JH, Rowed DW. Traumatic intracerebellar haematoma. Surgl Neurol 1978;10:340–1. 27. Schneider RC. Craniocerebral trauma. In: Kahn EA, Bassett RC, Schneider RC, Crosby EC, eds. Correlative Neurosurgery. Springfield III: Thomas, 1955:275–326. 28. Nagata K, Ishikawa T, Ishikawa T, et al. Delayed traumatic intracerebellar hematoma: correlation between the location of the hematoma and the pre-existing cerebellar contusion–case report. Neuro Med Chir 1991;31:792–6. 29. Yokota H, Mizunari T, Kuzuhara M, et al. [Traumatic delayed intracerebellar hematoma]. Neuro Medico Chir 1988;28):886–90. 30. Buczek M, Jagodzinski Z, Kopytek M, Dabrowska E. [Conservative treatment of post-traumatic intracerebellar hematoma]. Wiad Lek (Warsaw, Poland : 1960). 1989;42:550–5. 31. Takeuchi S, Takasato Y, Masaoka H, Hayakawa T. Traumatic intra-cerebellar haematoma: study of 17 cases. Br J Neurosurg 2011;25:62–7.
ORIGINAL ARTICLE
Traumatic intracerebellar haematoma: To operate or not to operate? Viraat Harsh1*, Anand Prakash1*, James Marcellus Barry2 & Anil Kumar1 1Department of Neurosurgery, Rajendra Institute of Medical Sciences, Ranchi, Jharkhand, India, and 2Department of
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Neurosurgery, Baylor College of Medicine, Houston, TX, USA origin in the absence of posterior fossa subdural or epidural haematoma is even rarer, accounting for only 0.6%– 0.82%.7–9 About one-fourth of all posterior fossa traumatic lesions can be accounted to intra-parenchymal cerebellar haemorrhage.3,10 Only few case reports and series have been documented.10–25 Wright16 recorded 17 posterior fossa haematomas and reported only 6 to have intracerebellar haematomas over a period of 10 years. Though trauma is the obvious cause, specific mechanism for this type of injury remains unclear. The clinical presentation of cerebellar haematoma may be readily apparent with classic signs including ataxia, nystagmus and signs of increased ICP like headache, lethargy and nausea and vomiting. It is critical to note that patients may be totally asymptomatic before rapid deterioration of neurologic status.26 Owing to the rarity of pure intracerebellar haemorrhage cases and the risk of rapid decompensation, Schneider27 emphasized that all patients with occipital trauma should be ruled out for blood collection in both supratentorial and infratentorial spaces. Some patients present with a delayed traumatic intracerebellar haematoma several hours after the trauma and have been reported to be associated with an unfavorable course and poor prognosis in the past.28,29 However, with the increasing use of computed tomography (CT) for screening of all head trauma patients, increasing number of cases have been reported to be diagnosed even before the appearance of symptoms and have thus been treated in a timely manner.7 Treatment options for this pathology are still evolving. The current standard is conservative management for fully conscious patients with a superficial bleed of small diameter ( 3cm) and surgical evacuation for patients with large bleeds ( 3cm), 4th ventricle compression, associated paradural bleeding, or hydrocephalus.30 Outcomes for this type of injury have previously been studied, with initial GCS score ( 7 vs. 7) found to be the most predictive factor.11 Other previously proven predictive factors include haematoma location (superficial vs. deep), haematoma volume, degree of fourth ventricle and cistern deformation, and associated SAH.31 Since very few patients present with this pathology, documentation of every case is important. In this study we
Abstract Background. Pure cerebellar haematoma of traumatic etiology, without associated posterior fossa sub- or epidural haematomas is a rare entity and has been reported to have a poor outcome. We report 23 patients with traumatic intracerebellar haematoma. We sought to study the pattern of such presentations and assess the factors which could be associated with their outcome. Methods. A retrospective review of prospectively collected data for all patients who were admitted for the management of traumatic intracerebellar haematoma at Rajendra Institute of Medical Sciences, Ranchi, India provided data for the 23 consecutive patients admitted for aforesaid over a seven-year study period. Medical records, diagnostic imaging and operative notes were reviewed for all patients. We divided the patient pool in to two groups based on their GCS score at the time of presentation – Group A (GCS 7) and Group B (GCS 7). The association of different allied factors was studied and statistically analyzed. The relevant medical literature was also reviewed. Results. Most Group B patients were found to be associated with poor outcome at hospital discharge. The overall incidence of poor outcome in our study was 69.56%. GCS score at time of admission, allied supratentorial lesions, advanced age, condition of fourth ventricle and chest infection were found to be important factors which could be associated with poor outcome. Conclusion. Surgery in patients with the mentioned risk factors remains debatable and should be approached cautiously. Larger multiinstitutional and meta-analytic studies are required to study and statistically establish the factors which might be associated with poor outcome in these patients. An algorithm which may be used in the management of traumatic intracerebellar haematoma patients is proposed. Keywords: haemorrhage; intracerebellar haematoma; traumatic brain injury
Introduction Haematomas of the posterior fossa are by themselves uncommon and account for only 3.3–6% of all head injuries.1–6 Pure cerebellar haematoma of traumatic
*These authors contributed equally to this work. Correspondence: Anil Kumar MBBS, MS, MCh, Professor and Head, Department of Neurosurgery, Rajendra Institute of Medical Sciences, Bariatu, Ranchi, Jharkhand 834009, India. Tel: 91 943 111 5510. E-mail: [email protected] Received for publication 18 October 2012; accepted 2 November 2014
1
2 V. Harsh et al. endeavor to review the pattern of such cases that we encountered. We have also reviewed the relevant medical literature.
Objectives We conducted this retrospective study to analyze the pattern of traumatic intracerebellar haematoma in our institute and study the factors which could be associated with the outcome.
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Patients and methods We retrospectively screened and (where required) reviewed records of more than 19300 head injury patients who were admitted in our Department of Neurosurgery at Rajendra Institute of Medical Sciences, Ranchi, India between January 2005 and January 2012 and all those who were diagnosed with an intracerebellar haemorrhage were selected for the study. Data on patient age, gender, GCS score, mechanism of injury, CT scan findings, associated supratentorial lesions, management strategy and outcome were gathered and analyzed. Patients were divided in to two groups based on their level of consciousness assessed by the GCS score at the time of presentation – Group A with GCS score greater than 7 and Group B with GCS lesser than or equal to 7. This was done to study the possible factors which could be associated with the outcome of either group. The mechanisms of injury were broadly categorized as road traffic accident (RTA), fall from height and assault. CT scan findings included clot location, size (largest transverse diameter of clot, categorized as 3 cm or 3 cm, or clot volume categorized as 15 ml or 15 ml), other associated findings and status of the fourth ventricle (normal vs. compressed). Surgical treatment (sub-occipital craniectomy or supratentorial craniotomy) and insertion of ventriculoperitoneal (VP) shunt was documented. A pre-decided general management protocol was followed. An algorithm based on the management protocol used by us is proposed and is depicted in Fig. 1. In the absence of an operable supratentorial lesion, conservative management was considered safe and acceptable in patients having GCS 14 or 15 and with clot size having largest diameter less than 3 cm or volume less than 15 ml. Surgical decompression of posterior fossa via a sub-occipital craniectomy was undertaken for patient with deteriorating level of consciousness with clot size measuring greater than or equal to 3 cm in diameter or volume greater than or equal to 15 ml or compression of fourth ventricle causing hydrocephalus. VP shunt was inserted on an emergency basis in patients having cerebellar clots which were near the fourth ventricle and caused compression, thus resulting in acute hydrocephalus but clot size was not big enough ( 15 ml) to merit surgical clot evacuation. Supratentorial craniotomy was undertaken for patients having injury in the supratentorial compartment. Outcome was documented as favorable or poor based on the Glasgow Outcome Scale at the time of hospital discharge (GOS-HD). Good recovery was counted as favorable whereas
all other states (death, vegetative state, severe disability and moderate disability) were counted as poor outcome. Sixmonth follow-up reports of all patients who turned in for the follow-up were also analyzed, however most patients could not be followed up and hence we are unable to discuss data related to six-month follow-up.
Results Retrospective screening and review of medical records, imaging data and operative notes yielded a total of 24 patients who were admitted in our institute for the management of intracerebellar haematoma over a period of seven years. There was 1 patient with firearm injury in occipital region who died prior to investigations and was therefore not included in our study, leaving us with 23 patients. The data related to 23 cases is shown in Table I. Male to female ratio in Group A (GCS score 7) was 2.3:1 and in Group B was 5.5:1. The age of the patients ranged from 6 to 72 years, mean age being 25.7 years in Group A and 38.5 years in group B. Overall mean age was 32.9 years. Most of the patients in our series (13/23, 56.5%) were between the age of 21 and 40. The rest of the patients were almost equally distributed between all age groups (Table II). More patients had poor GCS, i.e. lesser than or equal to 7 (13/23, 56.5%) and fell in Group B. The most common mode of injury was found to be RTA (60.86%) in either group – with 10 out of 14 patients in Group B (71.42%). CT scan records of the patients revealed clot size greater than or equal to 3 cm in 39% (9/23) of patients, most of whom (7/9 77%) were Group B patients. Also six patients out of 14 (42.85%) who had clot size lesser than 3 cm were Group B patients and had poor GCS score because of associated supratentorial lesions. Supratentorial lesions were present in 65.21% (15/23) and most of them (10/15, 66.6%) fell under Group B. Compression of 4th ventricle due to clot or surrounding edema was seen in 56.52% of patients. Two of the 13 cases having compression of 4th ventricle were from Group A. Two of the 10 having no compression of 4th ventricle on initial CT scan were from Group B at the time of admission. All patients had occipital bone fracture and 10 of them also had diffuse axonal injury (DAI). Seven of the patients with DAI were from Group B. Other associated pathology as seen on the CT scan is shown in Table III. Sub-occipital craniectomy with lax duroplasty was done to evacuate clot and brain debris in all the 7 Group B patients having clot size greater than or equal to 3 cm. In 40% cases (6/15) that had supratentorial lesion, extradural or subdural clot was surgically evacuated. VP shunt was inserted in lateral ventricle in three cases to tide over acute rise in intracranial pressure (ICP) due to dilation of ventricles. The overall incidence of poor outcome in our series at the time of hospital discharge was 69.56%. This incidence was reported to be 44.4% by d’Avella et al.11 Poor outcome was higher, 84.6% (11/13) in Group B as against 50% (5/10) in Group A. Mortality was 46.1% (6/13) in Group B and only 10% (1/10) in Group A.
Traumatic intracerebellar haematoma 3 TICH on CT scan Group A GCS > 7
Group B GCS ≤ 7
Clot size ≥ 3 cm or Clot vol. ≥ 15 ml
Clot size < 3 cm or Clot vol. < 15 ml
GCS ≤ 13
GCS> 13
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Clot size < 3 cm or Clot vol. < 15 ml
Clinical and radiological monitoring
No
Clinicoradiological stability and GCS Maintained
Conservative management
Also consider other pathologies as cause of low GCS and treat as per pathology eg. Supratentorial craniotomy for supratentorial hematomas
Extradural haematoma, Subdural haematoma or Brainstem compression present
Clinicoradiological deterioration and GCS decreasing
Yes
Further clinicoradiological or GCS score deterioration
Sub-occipital craniectomy + T/t of associated pathologies
If hydrocephalus is present, perform a VPS or place EVD
Fig. 1. Algorithm for management of traumatic intracerebellar haematoma. CT Computed Tomography; EDH Extradural Haemorrhage; EVD External ventricular drain; GCS Glasgow Coma Scale score; SDH Subdural Haemorrhage; T/t treatment; vol volume; VPS Ventriculo peritoneal shunting. Table I. Demographics and pattern of intracerebellar haematoma. Sl. Factors studied 1 2 3 4
5 6
7 8
9
Number of patients Mean age Male: Female ratio GCS score at admission
3–7 8–13
Group A
Group B
Total
10 25.7 2.3:1
13 38.53 5.5:1 13
23 32.95 3.6:1 13 7
7
14-15 Mechanism of injury RTA Fall from height Assault CT scan findings Cerebellar clot 3cm size 3cm Supratentorial lesion Fourth ventricle Normal Compressed Surgical treatment Sub-occipital craniectomy Supratentorial craniotomy V-P Shunt Outcome at hospital discharge Favorable (based on GOS-HD) Poor
3 4 2 4 8 2 5 8 2 2 2 2 5 5
10 3 0 6 7 10 2 11 7 4 1 2 11
Mortality
1
6
3 14 5 4 14 9 15 10 13 9 6 3 7 16 7
T, Computed Tomography; GOS-HD, Glasgow Outcome Scale at the time of hospital discharge; GCS, Glasgow Coma Scale C score; RTA, road traffic accident; V-P, ventriculo peritoneal
4 V. Harsh et al. Table II. Age distribution of patients with intracerebellar haematoma. Age group (years) Number 1–10 11–20 21–30 31–40 41–50 51–60 60
2 3 7 7 1 2 1
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
Discussion The unusually low incidence of cerebellar haematomas at our institution, 0.12% of all head injuries as compared to the normal incidence of 3.3–6% is notable. This deviation from the normal may have several causes. The most readily apparent is the long patient delivery time associated with a developing nation like India. This prolonged time to care would result in the death of patients who might have otherwise been picked up if admitted earlier. The propensity for males to experience this type of injury can also be explained by India’s developing nation status. In accordance with social customs, males are more commonly involved in outdoor activities than females and this reflects as higher incidence of intracerebellar trauma in males in our series. Though trauma is the obvious cause, specific mechanism for this type of injury remains unclear. Takeuchi et al.31 created a classification system for the types of trauma causing cerebellar haematomas. Their three classes included coup injuries, countercoup injuries and acceleration-deceleration injuries. This classification system proved useful in predicting the site of haematoma, with coup injuries causing only superficial bleeds while countercoup injuries resulted in deep cerebellar bleeds. In our study, RTA was the most common and most damaging mechanism of injury. This is likely attributable to other associated injuries caused by RTAs. Both high-speed and low-speed accidents produced intracerebellar haematoma with equal frequency but low speed accidents resulted in less severe associated injuries which resulted in more favorable outcome. Fall from height was more common amongst children while assault was more commonly associated with women and the elderly. All patients having history of assault fell under Group A. Interestingly, it seems that direct blows to the occiput are less likely to cause significant haemorrhage than acceleration/deceleration injury involved with RTA or a fall from height. In accordance with other studies on this topic, CT scan was extremely valuable in predicting patient outcomes. Clots greater than or equal to 3 cm in diameter, associated Table III. CT scan findings of patients with intracerebellar haematoma. Number of patients CT scan finding DAI EDH Acute SDH Skull fracture (other than occipital) Occipital bone fracture
Group A 3 1 0 7 10
Group B 7 2 4 0 13
Total 10 3 4 7 23
AI, Diffuse Axonal Injury; EDH, Extradural Haemorrhage; SDH, Subdural D Haemorrhage
supratentorial injuries and fourth ventricle compression were all shown by our study to predict a poor outcome. It makes sense that patients having associated lesion on CT scan sustained more extensive injury of brain and hence their chances of recovery depended in part on severity of other associated injuries as well. Outcomes were worse for patients in Group B. Such patients were critically injured and their brain function was already compromised at the time of admission. In accordance with previous studies on this topic our data showed GCS at initial presentation to be the most predictive clinical tool.11 In all patients undergoing surgical intracerebellar clot evacuation sub-occipital craniectomy was done and preferred over sub-occipital craniotomy as posterior fossa has less space for accommodation of any post-operative bleeding or post-operative edema. Some may note our use of VP shunts over external ventricular drains (EVDs) for certain patients experiencing hydrocephalus resulting from cerebellar haematoma. VP shunt was chosen over EVD placement because in our clinical setting EVDs have allegedly been found to more frequently result in infections as ventriculitis and meningitis three to four days post-EVD placement. However, patients developing acute hydrocephalus need drainage of CSF usually for a period of more than two weeks during which the clot resolves and perilesional edema subsides. There were several notable limitations of our study. First and most significant was the small sample size of 23 patients which reduces the statistical power of our study. This is however a result of the relative rarity of intracerebellar haematoma cases, as prior multi-institutional study11 could gather data for only 81 patients. Second, due to India’s developing status and overcrowded health system there were poor follow-up records making it difficult to draw any conclusions on the long-term outcome. Third, no attempt has been made to quantify the severity of associated non-neurological injuries in this group of patients. Considering many of these patients were involved in polytrauma situations it is reasonable to consider non-neurological trauma as an important component in their outcomes. Additionally, time from trauma to intervention is not included in our data set rendering us unable to comment how this critical factor affected outcomes. Finally, the retrospective character of this study creates the potential for observation and assessment bias.
Conclusion As previously indicated, traumatic intracerebellar haematomas are a rare and incompletely understood entity. In an attempt to study the factors which may be associated with poor outcome in intracerebellar haematoma cases we found that GCS score at the time of admission, related supratentorial lesion(s), advanced age, status of fourth ventricle and development of chest infection which interfered in tissue oxygenation, play an important role. Owing to the small number of patients and poor follow-up records we are unable to statistically associate these factors with poor outcome; however surgery in patients with the aforementioned negatively associated factors should be cautiously approached and a management protocol as depicted in
Traumatic intracerebellar haematoma 5 Fig. 1 must be followed. More multi-institutional studies should be considered in future to enhance pool size and more clearly elucidate the potential prognostic factors, both positive and negative, associated with traumatic intracerebellar haematoma.
Declaration of interest: The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.
Br J Neurosurg Downloaded from informahealthcare.com by Nyu Medical Center on 06/05/15 For personal use only.
References 1. Anegawa S, Yoshimura K, Kawai K, Kuramoto S, Ichinomiya C. [Traumatic intracerebellar hematoma—a case report and the review of the literature (author’s transl)]. No Shinkei Geka 1979;7:911–5. 2. St John JN, French BN. Traumatic hematomas of the posterior fossa. A clinicopathological spectrum. Surg Neurol 1986;25: 457–66. 3. Tsai FY, Teal JS, Itabashi HH, Huprich JE, Hieshima GB, Segall HD. Computed tomography of posterior fossa trauma. J Comput Assist Tomogr 1980;4:291–305. 4. Zuccarello M, Andrioli GC, Fiore DL,et al. Traumatic posterior fossa haemorrhage in children. Acta Neurochir (Wien) 1982;62: 79–85. 5. Jamieson KG, Yelland JD. Extradural hematoma. Report of 167 cases. J Neurosurg 1968;29:13–23. 6. Pevehouse BC, Bloom WH, Mc KW. Ophthalmologic aspects of diagnosis and localization of subdural hematoma. An analysis of 389 cases and review of the literature. Neurology 1960;10:1037–41. 7. Nagata K, Ishikawa T, Shigeno T, et al. Delayed traumatic intracerebellar hematoma: correlation between the location of the hematoma and the pre-existing cerebellar contusion—case report. Neurol Med Chir (Tokyo) 1991;31:792–6. 8. Fisher RG, Kim JK, Sachs E Jr. Complication in posterior fossa due to occipital trauma; their operability. J Am Med Assoc 1958;167:176–82. 9. Arseni C, Maretsis M. Traumatic cerebellar haematoma associated with posterior cerebral fossa subdural haematoma. Psychiatr Neurol Neurochir 1972;75:113–5. 10. Karasawa H, Furuya H, Naito H, et al. Acute hydrocephalus in posterior fossa injury. J Neurosurg 1997;86):629–32. 11. d’Avella D, Servadei F, Scerrati M, et al. Traumatic intracerebellar hemorrhage: clinicoradiological analysis of 81 patients. Neurosurgery 2002;50:16–25; discussion 25–17.
12. Martin AJ, Thomas NW. Evolving traumatic cerebellar hematoma. Neurology 2001;57:1565. 13. Seida M, Ito U, Kito K, Tomida S, Inaba Y. [Traumatic cerebellar hematoma in children. Report of two cases]. Neurol Med Chir (Tokyo) 1986;26:318–22. 14. Papadakis N, Safran A, Ramirez L, Sujatanond M, Mark VH. Traumatic cerebellar hematoma without subdural hematoma. JAMA 1976;235:530–1. 15. Rothballer AB. Traumatic cerebellar hematoma in the newborn. Case report of operative removal with survival. J Neurosurg 1962;19:913–5. 16. Wright RL. Traumatic hematomas of the posterior cranial fossa. J Neurosurg 1966;25:402–9. 17. Cassinari V, Dorizzi A, Pauli P, Monolo L. [Expansive traumatic lesions of the posterior cranial fossa. On 4 cases of cerebellar contusion and laceration]. Minerva Neurochir 1967;11:230–9. 18. Hamasaki T, Yamaki T, Yoshino E,et al. [Traumatic posterior fossa hematoma]. No to shinkei = Brain Nerve. 1987;39:1083–90. 19. Koziarski A, Frankiewicz E. Medical and surgical treatment of intracerebellar haematomas. Acta Neurochir (Wien) 1991;110:24–8. 20. Pozzati E, Grossi C, Padovani R. Traumatic intracerebellar hematomas. J Neurosurg 1982;56:691–4. 21. Pozzati E, Piazza G, Padovani R, Gaist G. Benign traumatic intracerebellar hematoma. Neurosurgery 1981;8:102–3. 22. Sato K, Hinokuma K, Matsuzawa Y, et al. [Clinical study of traumatic cerebellar contusion]. No Shinkei Geka 1987;15:1285–9. 23. Sokol JH, Rowed DW. Traumatic intracerebellar haematoma. Surg Neurol 1978;10:340–1. 24. Tibbs PA, Goldstein SJ, Smithson JR Jr. Delayed traumatic intracerebellar hematoma. Surg Neurol 1981;16:309–11. 25. Tsubokawa T, Yamada J, Tomizawa N, et al. [Classification of traumatic intracerebral hematoma by repeated CT-scan and clinical course (author’s transl)]. Neurol Med Chir (Tokyo) 1979;19:1127–37. 26. Sokol JH, Rowed DW. Traumatic intracerebellar haematoma. Surgl Neurol 1978;10:340–1. 27. Schneider RC. Craniocerebral trauma. In: Kahn EA, Bassett RC, Schneider RC, Crosby EC, eds. Correlative Neurosurgery. Springfield III: Thomas, 1955:275–326. 28. Nagata K, Ishikawa T, Ishikawa T, et al. Delayed traumatic intracerebellar hematoma: correlation between the location of the hematoma and the pre-existing cerebellar contusion–case report. Neuro Med Chir 1991;31:792–6. 29. Yokota H, Mizunari T, Kuzuhara M, et al. [Traumatic delayed intracerebellar hematoma]. Neuro Medico Chir 1988;28):886–90. 30. Buczek M, Jagodzinski Z, Kopytek M, Dabrowska E. [Conservative treatment of post-traumatic intracerebellar hematoma]. Wiad Lek (Warsaw, Poland : 1960). 1989;42:550–5. 31. Takeuchi S, Takasato Y, Masaoka H, Hayakawa T. Traumatic intra-cerebellar haematoma: study of 17 cases. Br J Neurosurg 2011;25:62–7.
