Clinical and radiographic predictors of neurological outcome following posterior fossa decompression for spontaneous cerebellar hemorrhage

Journal of Clinical Neuroscience 19 (2012) 1236–1241

Contents lists available at SciVerse ScienceDirect

Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn

Clinical Study

Clinical and radiographic predictors of neurological outcome following posterior fossa decompression for spontaneous cerebellar hemorrhage Nader S. Dahdaleh a,⇑, Brian J. Dlouhy a, Stephanus V. Viljoen a, Ana W. Capuano b, David K. Kung a, James C. Torner b, David M. Hasan a, Matthew A. Howard 3rd a a b

Department of Neurosurgery, College of Medicine, University of Iowa, 200 Hawkins Drive, Iowa City, IA 52262, USA Department of Epidemiology and Biostatistics, College of Medicine, University of Iowa, Iowa City, IA, USA

a r t i c l e

i n f o

Article history: Received 3 October 2011 Accepted 11 November 2011

Keywords: Cerebellar hemorrhage Posterior fossa decompression Suboccipital decompression Outcome

a b s t r a c t Spontaneous cerebellar hemorrhage often requires surgical suboccipital decompression and clot evacuation. Predictors of postoperative neurological deficits and outcome are not widely addressed in the literature. A retrospective review was conducted on 37 consecutive patients with the diagnosis of cerebellar hemorrhage requiring suboccipital decompression and clot evacuation. Clinical and radiographic variables were analyzed. Outcome measures were postoperative Glasgow Coma Scale (GCS) score, and long-term outcome measured by Rankin score and Glasgow Outcome Scale (GOS) score. A multivariate statistical analysis was conducted. The average age of patients was 71.1 years. There was significant improvement of neurological exam from a mean preoperative GCS score of 8.8 to a mean postoperative GCS score of 13.0. The mortality rate was 37.9%. According to the Rankin scale, 58.6% were functionally independent, 3.4% had a moderate disability, and none had a major disability or was in a vegetative state. Using GOS score, 62.1% had a favorable outcome. The presence of multiple comorbidities was associated with worse postoperative GCS and long-term outcome. A worse preoperative neurological exam, age older than 70 years, and the presence of intraventricular hemorrhage correlated only with a worse postoperative exam but not with the long-term outcome. Patients improve neurologically after posterior fossa decompression for cerebellar hemorrhage and a high percentage attain long-term functional outcome. Only the presence of multiple clinical comorbidities was associated with a worse outcome. Since there are no other preoperative predictors of long-term outcome, we recommend suboccipital decompression, when indicated, for patients with cerebellar hemorrhage regardless of age, hematoma size, or preoperative neurological exam. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Spontaneous cerebellar hemorrhage constitutes 5% to 10% of hemorrhagic strokes.1–3 It is often due to hypertension,4,5 and is associated with high mortality and morbidity.3,6 In contrast to the management of hypertensive supratentorial hemorrhages which often is expectant and medical,7,8 the treatment of spontaneous cerebellar hemorrhage often requires surgical suboccipital decompression.5 Spontaneous cerebellar hemorrhage is a disease that affects the elderly population.9 Clinical and radiographic predictors of postoperative neurological exam and outcome are not widely addressed in the current literature. These predictors might prove valuable not only in tailoring management and decision making, but also in predicting prognosis after suboccipital decompression and clot evacuation. We conducted a retrospective study

⇑ Corresponding author. Tel.: +1 319 356 2774; fax: +1 319 353 6605. E-mail address: [email protected] (N.S. Dahdaleh). 0967-5868/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jocn.2011.11.025

of 37 consecutive patients who presented with spontaneous cerebellar hemorrhage requiring suboccipital decompression through a craniectomy or a craniotomy. Correlation between clinical and radiographic parameters and postoperative GCS and long-term outcome was performed.

2. Methods 2.1. Strategy At our institution we pursue the following measures in the treatment of spontaneous cerebellar hemorrhage: Patients with small hematomas (67.5 cm3) and normal neurological exams are treated expectantly. Those with small hematomas and hydrocephalus are treated with external ventricular drainage (EVD). Patients with altered neurological examination, with or without large hematomas (>7.5 cm3, or 3 cm in largest diameter) are treated with suboccipital decompression and hematoma evacuation, with

N.S. Dahdaleh et al. / Journal of Clinical Neuroscience 19 (2012) 1236–1241

or without EVD, depending on the presence of hydrocephalus. Neither age nor associated medical comorbidities are factored into our decision in pursuing surgical decompression. Patients with cerebellar hemorrhage extending into the brainstem were excluded from the study. 2.2. Surgical technique A midline incision spanning the inion to the level of the upper subaxial spine is proposed. Suboccipital craniectomy is performed either through multiple burr holes, using a perforator, that are then connected with a rongeur or with the use of a drill and a dural guard attachment. The size of the craniectomy spans the inferior nuchal line to the inferior foramen magnum craniocaudally. Transversely, the size of the craniectomy depends on the size and location of the hematoma. After the hematoma is evacuated, a dural substitute is placed above the dural defect and the fascia is closed tightly. The same procedure is followed in a craniotomy; however, the bone flap is replaced and is secured to surrounding bone with titanium plates and screws. 2.3. Study procedure From January 1999 to September 2009, 37 consecutive patients were evaluated by the Neurosurgical Service at our institution with the diagnosis of cerebellar hemorrhage requiring suboccipital decompression and evacuation of hematoma. Clinical variables were retrospectively chart reviewed. These included: age, gender, type of surgery (suboccipital craniectomy or craniotomy), coumadin or warfarin use, hypertension, associated medical comorbidities (including coronary heart disease, congestive heart failure, atrial fibrillation, history of transient ischemic attack or stroke, diabetes mellitus, chronic renal failure or insufficiency, chronic obstructive pulmonary disease, peripheral vascular disease), and preoperative Glasgow Coma Scale (GCS) score. The preoperative head CT scans were also analyzed. The radiographic variables analyzed were: hematoma volume (cm3) measured by the ABC/2 formula,10 presence of obstructive hydrocephalus, presence of 4th ventricular deformity or obliteration, and the presence of intraventricular hemorrhage (IVH). The outcome measures analyzed were postoperative GCS score, the degree of improvement measured by subtracting preoperative GCS score from postoperative GCS score, Rankin scale score and Glasgow Outcome Score (GOS) on the last follow up. Clinical variables were available on 37 patients; long-term outcome measures using GOS score and Rankin score were available on 29 patients, whereas radiographic variables were available on 23 patients. Initial statistical analyses were performed using the t-test (paired for preoperative versus postoperative), chi-squared and Fisher exact methods depending on the variable. Statistical modeling was performed by using cumulative ordinal logistic regression (Proportional Odds model and extensions for detection of trend in odds ratios). Multicollinearity was examined. The final multivariate model was determined by manual backward selection of multiple saturated models. Clinical saturated models included variables such as age (continuous or dichotomized on the median), gender, hypertension, number of comorbidities (with and without hypertension), preoperative GCS score, anticoagulation, and craniectomy. Radiographic saturated models included the clinical variables, as well as hydrocephalus, IVH, and clot volume. Due to multicollinearity, 4th ventricular obliteration was not entered in the saturated model. Odds ratios for higher outcome scale were ascertained. In the analyses, a significance level of 0.05 was used. Statistical Analysis Software version 9.2 (SAS Institute, Cary, NC, USA) was used for all analyses.

1237

3. Results 3.1. Sociodemographics The average age of the cohort was 71.1 years ± 10.8. Twenty (54.1%) were female. Twenty-nine (78.4%) were known to be hypertensive. Fifteen patients (40.5%) had no other associated comorbidities, seven (18.9%) had one, eight (21.8%) had two, three (8.1%) had three, and four (10.8%) had four comorbidities. Thirteen patients (35.1%) were on coumadin. Twenty-nine (78.4%) underwent craniectomy. The average preoperative GCS score was 8.8 ± 4.2. Eleven (29.7%) patients died postoperatively or at follow-up. Eight patients were lost to follow-up. The average follow-up was 11.2 ± 14.8 months. 3.2. Long-term neurological outcome Rankin and GOS scores Among patients who were followed up, 17 out of 29 patients (58.6%) lived functionally independently (Rankin 0–2), one (3.4%) patient had a moderate disability (Rankin 3), none had a major disability or was in vegetative state (Rankin 4–5), and 11 (37.9%) had died (Rankin 6). Using the GOS score, 18 (62.1%) patients had a favorable outcome (lived independently or with minimal disability (GOS score 4–5), and none had a poor outcome (GOS score 2–3). 3.3. Clinical predictors of postoperative GCS score and of the difference between preoperative and postoperative GCS score There was statistically significant evidence of improvement from preoperative to postoperative GCS score (p < 0.01). The mean postoperative GCS score was 13.0 (standard deviation [SD] 3.3, 95% confidence interval [CI] = 11.9–14.1) in contrast to the mean preoperative GCS score of 8.8 (SD 4.2, 95% CI = 7.4–10.1). Considering clinical variables, patients with a better preoperative GCS score had higher odds (OR = 1.19, p = 0.03) of having a better postoperative GCS score. The other clinical variables did not predict the postoperative GCS score. Pertaining to the degree of improvement (difference between preoperative and postoperative GCS scores), patients who were 70 years of age or older had lower odds of having a higher GCS score (OR = 0.27, p = 0.032). (Table 1) 3.4. Clinical predictors of long-term outcome No clinical variable studied was associated with GOS score or with Rankin scale. There was weak statistical evidence that patients with more medical comorbidities had higher odds of having a higher (worse) Rankin score (OR = 1.91, p = 0.055) (Table 2). 3.5. Radiographic predictors of postoperative GCS score and the difference between preoperative and postoperative GCS scores Clinical, radiographic, and GCS score information was available for 23 patients. The final model for postoperative GCS score included the number of comorbidities, hypertension, and presence of IVH at presentation. Patients with IVH and hypertension at presentation had lower odds of having a higher (better) postoperative GCS score (respectively adjusted OR = 0.23, p = 0.009 and adjusted OR = 0.40, p = 0.051). Each increment in the number of comorbidities decreased the odds of higher (better) postoperative GCS score in about 90% (adjusted OR = 0.109, p = 0.023). Patients with hydrocephalus on presentation and females had higher odds of having a higher degree of improvement in GCS score (respectively adjusted OR = 7.006, p = 0.013, and adjusted OR = 2.828, p = 0.018). All patients presenting with hydrocephalus were treated with a ventriculostomy. (Table 3)

1238

N.S. Dahdaleh et al. / Journal of Clinical Neuroscience 19 (2012) 1236–1241

Table 1 Bivariate and multivariate Proportional Odds model for calculation of the odds of higher preoperative/postoperative Glasgow Coma Scale (GCS) scores considering clinical variables Variables

n

Postoperative GCS

Difference between Pre and Post-Operative GCS

Bivariate

Bivariate

OR

Test statistic

p-value

OR

Test statistic

p-value

37

1.01

0.18

0.667

0.95

2.91

0.088

<70 P70

19 18

Reference 0.49

Reference 1.34

Reference 0.248

Reference 0.27⁄⁄

Reference 4.59

Reference 0.032

Female Male

20 17

1.16 Reference

0.23 Reference

0.634 Reference

1.48 Reference

1.78 Reference

0.183 Reference

No Yes

8 29 37 37

Reference 1.08 0.85 1.19⁄⁄

Reference 0.04 0.51 4.63

Reference 0.833 0.475 0.031

Reference 0.58 1.18

Reference 2.26 0.58

Reference 0.133 0.445

No Yes

24 13

Reference 1.07

Reference 0.04

Reference 0.835

Reference 1.37

Reference 1.06

Reference 0.304

No Yes

8 29

Reference 1.57

Reference 1.51

Reference 0.219

Reference 1.23

Reference 0.34

Reference 0.561

Age Age group (years)

Gender

Hypertension

No. of comorbidities* Preoperative GCS score Anticoagulation

Craniectomy

OR = odds ratio. * Hypertension not included. ** Only variable selected for the final model (p < 0.05).

Table 2 Bivariate and multivariate Proportional Odds model for calculation of the odds of a higher Rankin and Glasgow Outcome Scale (GOS) score at last follow-up considering clinical variables Variables

n

Rankin score

GOS score

Bivariate

Bivariate

Estimate

Test statistic

p-value

Estimate

Test statistic

p-value

29

1.04

1.38

0.240

0.97

1.11

0.292

<70 P70

16 13

Reference 3.09

Reference 2.55

Reference 0.111

Reference 0.28

Reference 3.11

Reference 0.078

Female Male

18 11

0.64 Reference

1.55 Reference

0.213 Reference

1.63 Reference

1.78 Reference

0.182 Reference

No Yes

5 24 29 29

Reference 1.06 1.91 0.97

Reference 0.02 3.67 0.16

Reference 0.897 0.055 0.688

Reference 1.00 0.57 1.02

Reference 0.00 2.88 0.04

Reference 1.000 0.090 0.846

No Yes

20 9

Reference 1.96

Reference 2.96

Reference 0.085

Reference 0.50

Reference 3.02

Reference 0.082

No Yes

6 15

Reference 1.48

Reference 0.89

Reference 0.346

Reference 1.00

Reference 0.00

Reference 1.000

Age Age group (years)

Gender

Hypertension

No. of comorbidities* Preoperative GCS score Anticoagulation

Craniectomy

*

Hypertension not included.

3.6. Radiographic predictors of long-term outcome Clinical, radiographic, and long-term outcome information was available for 17 patients. No radiographic variable predicted longterm outcome measured by the GOS score or the Rankin score. In this analysis the presence of more medical comorbidities was associated with a worse Rankin score (OR = 3.86, p = 0.009), and GOS score (OR = 0.31, p = 0.023). Interestingly, craniectomies were associated with worse Rankin scores (OR = 3.62, p = 0.04) (Table 4). 4. Discussion The first surgical evacuation of a cerebellar hematoma was reported in 1906 by Ballance.1 To date this has been common practice

in the treatment of cerebellar hemorrhages. The determinants of operative management are usually hematoma size and neurological examination findings.11,12 Hematoma sizes of 3 cm in diameter or larger and a neurological exam worse than a GCS score of 13 have been used as determinants for operative intervention in many reports and studies.2–4,13 Wijdicks et al.9 studied 94 patients diagnosed with cerebellar hemorrhage. The median age of their cohort was 74. The overall surgical probability was 33%. They demonstrated that at their institution, a hematoma size of <3 cm does not influence their decision to operate on patients younger than 70 years old if clinical deterioration occurs, but surgery is withheld from patients older than 70 with large hematomas. Our study demonstrated that even though the degree of postoperative improvement was less in patients older than 70, age did not influence long term outcome reflected by Rankin score, and GOS.

1239

N.S. Dahdaleh et al. / Journal of Clinical Neuroscience 19 (2012) 1236–1241

Table 3 Bivariate and multivariate Proportional Odds model for calculation of the odds of a higher Preoperative/Postoperative Glasgow Coma Scale (GCS) score considering clinical and radiographic variables Variables

n

Post-Operative GCS

Difference between Pre and Post-Operative GCS

Bivariate

Multivariate

Bivariate

Multivariate

OR

Test statistic

pvalue

OR

Test statistic

pvalue

OR

Test statistic

pvalue

OR

Test statistic

pvalue

23

1.05

1.74

0.188

–

–

–

0.95

2.23

0.135

–

–

–

<70 P70

14 9

Ref. 1.34

Ref. 0.12

Ref. 0.726

– –

– –

– –

Ref. 0.19

Ref. 4.04

Ref. 0.045

Female Male

11 12

0.57 Ref.

0.80 Ref.

0.371 Ref.

– –

– –

– –

0.56 Ref.

1.33 Ref.

0.249 Ref.

– – Ref. 2.828

– – Ref. 5.613

– – Ref. 0.018

No Yes

4 19 23

Ref. 0.57 0.80

Ref. 0.80 0.45

Ref. 0.371 0.504

Ref. 0.403 0.109

Ref. 3.806 5.167

Ref. 0.051 0.023

Ref. 0.56 0.79

Ref. 1.33 0.62

Ref. 0.249 0.433

– – –

– – –

– – –

23

1.08

0.72

0.396

–

–

–

No Yes

17 6

Ref. 507.40

Ref. 0.01

Ref. 0.943

– –

– –

– –

Ref. 1.33

Ref. 0.47

Ref. 0.493

– –

– –

– –

No Yes

4 19

Ref. 1.39

Ref. 0.41

Ref. 0.523

– –

– –

– –

Ref. 2.29

Ref. 2.50

Ref. 0.114

– –

– –

– –

No Yes

3 20

Ref. 0.71

Ref. 0.27

Ref. 0.603

– –

– –

– –

Ref. 3.30

Ref. 3.76

Ref. 0.052

Ref. 7.006

Ref. 6.223

Ref. 0.013

No Yes

12 11

Ref. 0.43

Ref. 3.64

Ref. 0.056

Ref. 0.225

Ref. 6.837

Ref. 0.009

Ref. 1.28

Ref. 0.46

Ref. 0.497

– –

– –

– –

No Yes

5 18 23

Ref. 0.79 0.99

Ref. 0.22 0.21

Ref. 0.643 0.650

– – –

– – –

– – –

Ref. 2.12 1.06

Ref. 2.57 2.78

Ref. 0.109 0.095

– – –

– – –

– – –

Age Age group (years)

Gender

Hypertension

No. comorbidities* Pre-operative GCS score Anticoagulation

Craniectomy

Hydrocephalus

IVH

Obliteration of the 4th ventricle

Clot volume (cm3)

Bold text indicates a statistically significant result. IVH = intraventricular hemorrhage, Ref. = reference. Hypertension not included.

*

Luparello et al.14 reported the results of 22 patients suffering from cerebellar hemorrhages with a mean age of 56 years. The study evaluated admission GCS, maximal hematoma diameter on CT, location of the hematoma, IVH, hydrocephalus, and the degree of quadrigeminal cistern involvement (compression, obliteration, or the presence of blood). The outcome measure was GOS at the time of discharge. Their study showed that patients with a hematoma size 3 cm or larger and a GCS worse than 9 had an unfavorable outcome despite surgery. Patients with hematoma sizes smaller than 3 cm and a GCS of 9 or better had a favorable outcome. For patients with hematomas larger than 3 cm and a GCS score of 9 or better, outcome depended on the location of the hematoma, and the concurrent presence of hydrocephalus, quadrigeminal cistern involvement, and IVH. The worse results were observed if hydrocephalus, quadrigeminal cistern obliteration, and IVH were present. The authors were aggressive in the treatment of patients with poor presenting GCS, however were not specific in terms of the mode of surgical therapy or cerebrospinal fluid diversion. Moreover, GOS was obtained only at the time of discharge, and major clinical criteria were not factored in their analysis including the presence of other medical comorbidities. Cohen et al.15 studied 37 patients with cerebellar hemorrhages, 30 of whom underwent suboccipital craniectomy and the rest were treated conservatively. The mean age of their cohort was 64.8 years. Comparisons were made between the two groups with the outcome measure being GOS score at discharge and three months later. They evaluated the following clinical and radiographic parameters: size of hematoma, GCS, presence of long tract signs, and hydrocephalus. The prediction of outcome by a logistic

regression model included both the size of the hematoma and the treatment modality. In their study the size of the hematoma was the only significant predictor of the outcome, with larger hematomas predicting worse outcome. They suggested that surgical evacuation should be pursued in all patients presenting a with large hematoma (>3 cm). Their study did not consider other important clinical variables including age and the presence of other comorbidities, which are potential confounding variables. Our study demonstrated that in patients with cerebellar hemorrhage requiring hematoma evacuation, the preoperative size does not correlate with either postoperative GCS score, Rankin score, or with GOS score. Taneda et al.16 demonstrated that obliteration of the quadrigeminal cistern is a more important determinant than the size of the hematoma for surgical evacuation. In our study this did not influence post operative GCS, or long-term outcome. Moreover, Da Pian et al.17 analyzed 155 patients with cerebellar hemorrhage. Their study suggested that surgery is indicated for patients with obliteration of the fourth ventricle as well as hydrocephalus. Our study is in agreement with this finding as all patients generally improve neurologically. Their study, however, lacked statistical analysis of potential confounders. Kirollos et al.6 studied 50 patients with spontaneous cerebellar hemorrhage with a mean age of 62 years. Their study also emphasized the importance of fourth ventricular appearance on CT scan. The appearance was graded as: I, normal; II, deformed; and III, obliterated. Along with the fourth ventricular grade, the GCS score factored in their decision for conservative compared with operative management. In their cohort, 28 patients underwent surgical

1240

N.S. Dahdaleh et al. / Journal of Clinical Neuroscience 19 (2012) 1236–1241

Table 4 Bivariate and multivariate Proportional Odds model for odds of higher Rankin and Glasgow Outcome Scale (GOS) scores, considering clinical and radiographic variables Variables

n

Rankin

GOS

Bivariate

Bivariate

Estimate

Test statistic

p-value

Estimate

Test statistic

p-value

17

1.01

0.10

0.755

0.99

0.05

0.822

<70 P70

11 6

Reference 1.53

Reference 0.22

Reference 0.641

Reference 0.57

Reference 0.35

Reference 0.553

Female Male

9 8

1.81 Reference

0.72 Reference

0.397 Reference

0.69 Reference

0.25 Reference

0.614 Reference

No Yes

2 15 17 17

Reference 1.81 3.86 0.99

Reference 0.72 6.74 0.02

Reference 0.397 0.009 0.887

Reference 0.69 0.31 1.00

Reference 0.25 5.14 0.00

Reference 0.614 0.023 0.992

No Yes

13 4

Reference 1.35

Reference 0.35

Reference 0.557

Reference 0.76

Reference 0.28

Reference 0.598

No Yes

4 13

Reference 3.62

Reference 4.20

Reference 0.040

Reference 0.48

Reference 1.47

Reference 0.226

No Yes

3 14

Reference 0.46

Reference 1.63

Reference 0.202

Reference 2.39

Reference 1.88

Reference 0.171

No Yes

10 7

Reference 1.94

Reference 2.01

Reference 0.156

Reference 0.49

Reference 2.11

Reference 0.147

No Yes

4 13 17

Reference 1.50 1.02

Reference 0.61 0.37

Reference 0.436 0.544

Reference 0.87 0.98

Reference 0.07 0.38

Reference 0.794 0.538

Age Age group (years)

Gender

Hypertension

No. of comorbidities* Preoperative GCS score Anticoagulation

Craniectomy

Hydrocephalus**

IVH

**

Fourth ventricular obliteration**

Clot volume (cm3)**

Bold text indicates a statistically significant result. IVH = intraventricular hemorrhage, Ref. = reference. * Hypertension not included. ** No radiographic variable selected for the final model.

evacuation of the cerebellar hemorrhage. Correlation between the fourth ventricular grade and hematoma volume, hydrocephalus, and preoperative GCS score were statistically significant. All patients with grade III (18 patients) and patients with grade II and a GCS score of less than 13 who did not improve with EVD (10 patients) underwent surgical evacuation. The overall mortality in their study was 54% among patients who underwent surgical evacuation. Good outcome was defined as GOS score of 4 or 5 at three months and was achieved in 16.7% of patients with grade III and GCS score <8 who underwent surgical evacuation, and 60% of the patients with grade II fourth ventricular appearance who underwent surgical evacuation. In our study, among patients for whom radiographic data were available, none had a normal appearing fourth ventricle on CT scan, 20.8% had deformed ventricles, and most (79.2%) had complete obliteration of their fourth ventricle. Using multivariate analysis and controlling for potential confounders, the presence of a deformed or obliterated fourth ventricle did not, however, affect postoperative neurological examination or outcome following suboccipital decompression. Salvati et al.11 examined 50 patients with cerebellar hemorrhage with an average age of 63.7 years. Their criteria for operative management was the presenting GCS score and hematoma size. Patients with GCS scores of <13 and hematoma sizes of >25 mm underwent surgical evacuation. Karnofsky performance scale score was evaluated at three, six and twelve months. In their study, analysis included both patients treated surgically and conservatively. Age was not associated with outcome. The presence of two or more risk factors significantly influenced mortality and quality of life. These findings are in agreement with our study, although their study did not analyze other potential confounding variables.

In a recent study, Dammann et al.18 reported their experience in treating 57 patients with cerebellar hemorrhage with suboccipital surgical evacuation. In their study, age, preoperative GCS score, hematoma size, acute hydrocephalus, tight posterior fossa, brain stem compression, fourth ventricular compression, and the presence of IVH were examined. The GOS score was the outcome measure used. The average age of their group was 64.4 years. The mortality rate was 25%. At the last follow-up 47% of patients had a good outcome with GOS score of 4 and 5, and 28% of patients had a poor outcome with GOS score of 2 or 3. Preoperative GCS score was found to be a significant determinant for clinical outcome, with a lower presenting GCS score predicting a poorer outcome. Among the radiographic criteria examined, brain stem compression and a tight posterior fossa predicted clinical outcome while hematoma size did not correlate with outcome. The findings pertaining to hematoma size are in agreement with our study; however, our study did not show that preoperative GCS score predicted outcome and this is possibly due to the small number in our cohort. Alternatively the average age of our population was older and we examined the effect of comorbidities, which were found to be a potential confounding factor in predicting poor outcome Thus, our study showed that there is a significant improvement in postoperative GCS following suboccipital craniotomy for cerebellar hemorrhage. While age, hematoma size, and preoperative neurological examination are reportedly historical determinants of operative or conservative management, hematoma size affected neither the postoperative neurological exam nor long-term outcome following posterior fossa decompression. Operating on patients 70 years of age or older resulted in a lesser degree of neurological improvement compared to patients younger than

N.S. Dahdaleh et al. / Journal of Clinical Neuroscience 19 (2012) 1236–1241

70 years of age but this did not affect long-term outcome. A worse preoperative GCS was associated with a worse postoperative GCS, but this did not affect long-term outcome. Treating hydrocephalus, when present, improves the postoperative neurological examination significantly. Moreover, the presence of IVH was associated with worse postoperative neurological examination but this did not affect long-term outcome. Finally, a higher number of medical comorbidities was associated with a worse postoperative GCS as well as a poorer long-term neurological outcome. Limitations to this study include its retrospective nature as well as its small sample size. 5. Conclusion Overall, patients improve neurologically after posterior fossa decompression for cerebellar hemorrhage and a high percentage of patients attain long-term functional improvement. The only predictor of a worse neurological outcome was a greater burden of medical comorbidities. Since there were no other preoperative clinical or radiographic predictors of long-term outcome, we recommend suboccipital decompression, when indicated, for patients with cerebellar hemorrhage regardless of age, hematoma size, or preoperative neurological exam. References 1. Ballance H. Case of traumatic hemorrhage into the left lateral lobe of the cerebellum, treated by operation, with recovery. Surg Gynecol Obstet 1906;3:223–5. 2. Heros RC. Cerebellar hemorrhage and infarction. Stroke 1982;13:106–9. 3. Lui TN, Fairholm DJ, Shu TF, et al. Surgical treatment of spontaneous cerebellar hemorrhage. Surg Neurol 1985;23:555–8. 4. Little JR, Tubman DE, Ethier R. Cerebellar hemorrhage in adults. Diagnosis by computerized tomography. J Neurosurg 1978;48:575–9.

1241

5. Tamaki T, Kitamura T, Node Y, et al. Paramedian suboccipital mini-craniectomy for evacuation of spontaneous cerebellar hemorrhage. Neurol Med Chir (Tokyo) 2004;44:578–82 [discussion 583]. 6. Kirollos RW, Tyagi AK, Ross SA, et al. Management of spontaneous cerebellar hematomas: a prospective treatment protocol. Neurosurgery 2001;49:1378–86 discussion 1386–7. 7. Broderick J, Connolly S, Feldmann E, et al. Guidelines for the management of spontaneous intracerebral hemorrhage in adults: 2007 update: a guideline from the American Heart Association/American Stroke Association Stroke Council, High Blood Pressure Research Council, and the Quality of Care and Outcomes in Research Interdisciplinary Working Group. Stroke 2007;38: 2001–23. 8. Mendelow AD, Gregson BA, Fernandes HM, et al. Early surgery versus initial conservative treatment in patients with spontaneous supratentorial intracerebral haematomas in the International Surgical Trial in Intracerebral Haemorrhage (STICH): a randomised trial. Lancet 2005;365:387–97. 9. Wijdicks EF, St Louis EK, Atkinson JD, et al. Clinician’s biases toward surgery in cerebellar hematomas: an analysis of decision-making in 94 patients. Cerebrovasc Dis 2000;10:93–6. 10. Kothari RU, Brott T, Broderick JP, et al. The ABCs of measuring intracerebral hemorrhage volumes. Stroke 1996;27:1304–5. 11. Salvati M, Cervoni L, Raco A, et al. Spontaneous cerebellar hemorrhage: clinical remarks on 50 cases. Surg Neurol 2001;55:156–61 [discussion 161]. 12. van der Hoop RG, Vermeulen M, van Gijn J. Cerebellar hemorrhage: diagnosis and treatment. Surg Neurol 1988;29:6–10. 13. Freeman RE, Onofrio BM, Okazaki H, et al. Spontaneous intracerebellar hemorrhage. diagnosis and surgical treatment. Neurology 1973;23:84–90. 14. Luparello V, Canavero S. Treatment of hypertensive cerebellar hemorrhage– surgical or conservative management? Neurosurgery 1995;37:552–3. 15. Cohen ZR, Ram Z, Knoller N, et al. Management and outcome of non-traumatic cerebellar haemorrhage. Cerebrovasc Dis 2002;14:207–13. 16. Taneda M, Hayakawa T, Mogami H. Primary cerebellar hemorrhage. Quadrigeminal cistern obliteration on CT scans as a predictor of outcome. J Neurosurg 1987;67:545–52. 17. Da Pian R, Bazzan A, Pasqualin A. Surgical versus medical treatment of spontaneous posterior fossa haematomas: a cooperative study on 205 cases. Neurol Res 1984;6:145–51. 18. Dammann P, Asgari S, Bassiouni H, et al. Spontaneous cerebellar hemorrhage– experience with 57 surgically treated patients and review of the literature. Neurosurg Rev 2011;34:77–86.