Life-Threatening Cerebral Hematoma Owing to Aneurysm Rupture

Original Article

Life-Threatening Cerebral Hematoma Owing to Aneurysm Rupture Giuseppe Talamonti1, Michele Nichelatti2, Aysam Adnan Al Mashni1, Giuseppe D’Aliberti1

OBJECTIVE: To refine the surgical indications of surgery for life-threatening cerebral hematomas caused by aneurysm rupture, through the analysis of possible outcome predictors.

-

METHODS: Forty-nine consecutive patients requiring prompt clot evacuation were retrospectively reviewed. In all cases, the hematoma was equal to or greater than 60 mL. The bleeding aneurysm was located on the middle cerebral artery in 26 cases, on the internal carotid artery in 10 cases, and on the anterior cerebral artery in 13 cases; four aneurysms were giant. Six patients underwent aneurysm coiling followed by clot removal, whereas 43 patients were managed by concomitant clot evacuation and aneurysm clipping. The main clinical and radiologic features, the management paths and the treatment modalities were correlated with the outcomes. A statistical analysis was conducted.

favorable outcome are not negligible. Further improvement may be possible through better patient selection and the identification of nonsalvageable subjects.

-

RESULTS: Overall mortality was 32.6%, severe morbidity was 22.4% and 22 patients (44.8%) achieved favorable results. The short-term results were more significantly influenced by the radiological parameters than by the initial clinical conditions. The prognostic weight of the radiologic findings was partially lost for six-month results, whereby management factors gained in importance.

-

CONCLUSIONS: The bleeding severity was strongly determinant for early mortality. However, if patients can survive the initial crucial phase, their chances of a

-

Key words Intracerebral hemorrhage - Intracranial aneurysms - Microsurgical clipping - Predictors - Subarachnoid hemorrhage -

Abbreviations and Acronyms ACoA: Anterior communicating artery CSF: Cerebrospinal fluid CT: Computed tomography CTA: Computed tomographic angiography DA: Digital angiography ED: Emergency department GCS: Glasgow Coma Scale GOS: Glasgow Outcome Score ICA: Internal carotid artery

WORLD NEUROSURGERY - [-]: ---, - 2015

INTRODUCTION

D

evelopment of endovascular techniques has led to a significant general improvement in the outcome of patients with cerebral aneurysms. Nevertheless, when the aneurysms are responsible for intracerebral hematomas (ICHs), the advantages of endovascular treatment become less evident. In the case of life-threatening aneurysmal ICH, emergency clot evacuation may represent the sole chance of survival, but indications and the modalities of aneurysm treatment are still debated.1e10 Emergency clot evacuation with concomitant aneurysm clipping represents the classical treatment option. It has even been attempted without any preoperative angiography to shorten the time for brain decompression,1 but modern computed tomographic angiography (CTA) usually provides prompt and adequate management information. Some authors3,8 think that immediate coiling followed by clot evacuation is time efficient and may be advantageous over immediate clot evacuation and clipping. There are several articles regarding the management of aneurysmal ICHs, but many concern just the middle cerebral artery2,5,7,9,11 or include clots of any size and patients with a wide range of neurologic conditions.4,5,8-12

ICH: Intracerebral hematoma ICP: Intracranial pressure MCA: Middle cerebral artery OR: Operating room SAH: Subarachnoid hemorrhage From the 1Department of Neurosurgery and 2Service of Bio-Statistics, Niguarda Ca’Granda Hospital, Piazza Ospedale Maggiore, Milan, Italy To whom correspondence should be addressed: Giuseppe Talamonti, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2015). http://dx.doi.org/10.1016/j.wneu.2015.08.082 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2015 Elsevier Inc. All rights reserved.

www.WORLDNEUROSURGERY.org

1

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

In this article, we report our personal series of 49 consecutive moribund patients with life-threatening ICHs caused by aneurysm rupture. All patients required emergency ICH evacuation and concomitant aneurysm securing. MATERIAL AND METHODS The charts and the radiologic documentation of all patients undergoing emergency clot evacuation (1998e2013) were retrospectively reviewed. Forty-nine consecutive patients with aneurysmal ICH were found. Patients without life-threatening ICHs, patients with spontaneous subdural hematomas, and those with vital brain destructions or concomitant diseases that could potentially influence the clinical situation or outcome (e.g., arteriovenous malformations, cerebral tumors, coagulation disturbances) are not included in this series. Patients with posterior circulation aneurysms are not included either. Patient Population There were 18 men and 31 women (male:female ratio ¼ 3:5), and age ranged from 39 to 71 years (mean ¼ 53.9 years). Thirty-three patients experienced sudden loss of consciousness with immediate comatose state, and 16 patients were initially described as more or less alert with subsequent early preadmission deterioration. Nevertheless, clear early rebleeding was documented in only 6 cases. All patients but 3 of these patients were referred to our emergency department (ED) within 1 hour of hemorrhage. On admission, all patients but one scored 8 or less on the Glasgow Coma Scale (GCS), and there were 4 Grade IVs and 45 Grade Vs according to the Hunt and Hess Grading System.13 Normal pupils were described in just 9 patients, whereas dilated pupils ipsilateral to the ICH were reported in 32 patients; the remaining 7 patients were admitted with bilaterally fixed and dilated pupils but the presence of the other brain stem reflexes. All patients underwent aggressive resuscitation including intubation, ventilation, fluid administration, correction of any cardiocirculatory or electrolytic anomalies. The main clinical parameters are summarized in Table 1. Radiological Assessments Immediate computed tomography (CT) scans were performed in all cases. All patients harbored life-threatening ICHs, which were classified according to location and size. There were 3 temporal ICHs (Figure 1), 13 frontal ICHs (Figure 2), and 33 frontotemporal or frontotemporoparietal ICHs. The volume of the ICH was grossly calculated considering the clot as an ellipsoid and its three major diameters as the axes, and using the classical equation V ¼ 4/3 pabc (a, b, and c ¼ semiaxes). For example, an ICH with three major diameters of 6, 8, and 4 cm had the following estimated volume: V ¼ 4/3  3.14  3  4  2 ¼ 100.4 mL. With this method, the ICH volume ranged from 60 to 200 mL (mean ¼ 104.3 mL). Midline shift was evident in all patients; it was considered minimal (<5 mm) in 9 patients, intermediate (6e10 mm) in 23 patients, and maximal (>10 mm) in 17 patients. More or less intraventricular blood was reported in 41 patients (83.6%), but true acute hydrocephalus was shown in 19 patients (38.7%).

2

www.SCIENCEDIRECT.com

Afterwards, depending on the clinical conditions, either the patients underwent preoperative neurovascular assessments or they were immediately referred to the neurosurgical operating room (OR) for emergency clot evacuation and cistern exploration: a total of 27 patients (55.1%) were studied by CTA, 8 (16.3%) underwent digital angiography (DA), and 14 (28.6%) underwent direct exploration with clot evacuation and aneurysm clipping. In the 35 patients who underwent preoperative neuroradiologic assessments, the bleeding aneurysm was located on the middle cerebral artery (MCA) in 17 cases, the internal carotid artery (ICA) in 8 cases, the anterior communicating artery (ACoA) in 8 cases, the pericallosal artery in 2 cases. The CTA was able to show the bleeding aneurysm in all patients but one (Figure 2), in whom the aneurysm was hidden by artifacts because of coils that were put in place 2 years earlier when this patient had experienced her first subarachnoid hemorrhage (SAH). The 14 patients who underwent surgery without preoperative angiography harbored bleeding aneurysms on the MCA (9 patients), the ACoA (3 patients), and the ICA (2 patients). Giant aneurysms (larger than 2.5 cm) (Figure 3) were found in 4 patients. Multiple aneurysms were documented in 5 patients. Radiologic features are reported in Table 1. Treatment Among the 8 patients who underwent preoperative DA, 2 (4.1%) were managed by subsequent clot removal and aneurysm clipping, whereas 6 (12.2%) underwent aneurysm coiling at the same time. The decision regarding endovascular treatment mainly depended on the clinical conditions, the aneurysm features, and the availability of neuroradiology facilities. In all the coiled cases, the aneurysm dome could be protected, and more than 80% sac obliteration was achieved. Afterwards, all patients underwent clot removal without any cistern exploration. The remaining 41 patients (83.7%) were surgically managed immediately after computed tomography (CT) or CTA. The mean time from the admission to the ED and the arrival at the OR (ED-OR interval) was 160 minutes (range ¼ 120e180 minutes) for patients whose aneurysms were coiled before surgery, and 115 minutes (range ¼ 30 minutes to 23 hours) for those patients undergoing surgical clipping. The operation consisted of the classical surgical management of cerebral aneurysms. The approach was planned to allow complete ICH evacuation and to access the circle of Willis. Following a wide craniotomy, the ICH was tapped with a blunt Cushing needle to slacken the dural plane. Next, clot removal started distantly from the aneurysm to obtain further brain relaxation and early visualization of the proximal vessels. As soon as possible, the cisterns were dissected and the bleeding aneurysm secured by clipping. In patients undergoing cistern exploration without preoperative angiography, the presumed site of the aneurysm was explored first. Hematoma evacuation was only completed after definitive obliteration of the aneurysm. In 13 patients (26.5%), early intraoperative aneurysm rebleeding occurred before the aneurysm was exposed; in 2 patients, this happened during the craniotomy and before any dural incision. In all these patients, the arrest of the hemorrhage could be obtained by the combination of strong suction, pad compression, brain amputation, temporary trapping, and/or

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2015.08.082

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

Table 1. Clinical, Radiologic, and Treatment Features of 49 Patients with Intracerebral Hematoma Aneurysm

Table 1. Continued Number of Patients

Percentage of Patients

<60

22

44.9

60e120

19

38.8

8

16.3

Features

Number of Patients

Percentage of Patients

Male

18

36.7

Female

31

63.3

>120

<50 years

14

28.6

50e59 years

22

44.9

>59 years

13

26.5

Grade IV

4

8.1

Grade V

45

91.9

2

4.1

Features Sex

Age (mean, 53.9 years)

Hunt-Hess

Glasgow Coma Scale score 9 8e7

3

6.1

6e5

13

26.5

4e3

31

63.3

<100

23

46.9

100e150 mL

21

42.9

5

10.2

Intracerebral hematoma volume (mL)

>150 mL Midline shift (mm) <5

9

18.4

6e10

23

46.9

>10

17

34.7

No

30

61.2

Yes

19

38.8

Acute hydrocephalus

Aneurysm location Internal carotid artery

10

20.4

Middle cerebral artery

26

53.1

Anterior cerebral artery

13

26.5

Surgery without any angiography

14

28.6

CTA or DA with clipping

29

59.2

6

12.2

Treatment

Coiling plus intracerebral hematoma evacuation

*ED-OR interval is the interval between admission to the emergency department and arrival in operating room. Continues

WORLD NEUROSURGERY - [-]: ---, - 2015

ED-OR interval* (minutes)

*ED-OR interval is the interval between admission to the emergency department and arrival in operating room.

blind clipping. As soon as the operatory field was cleared, the aneurysm preparation was completed, the aneurysm was secured, and the temporary clips were removed. In the remaining 36 patients, the exposure of the parent vessel could be obtained without any complication. Minor intraoperative aneurysm leaks were reported in 29 adjunctive patients, which means that intraoperative bleeding occurred in 42 patients (85.7%). However, this number also includes hemorrhages, which occurred in relatively safe conditions (i.e., when the aneurysm was sufficiently exposed). In all cases, adequate hemostasis could be achieved, and the aneurysm neck could be handled definitively. The aneurysms were classically managed by neck clipping, but in four patients (1 MCA, 1 ICA, and 2 ACoA), trapping and parent artery obliteration were needed. The aneurysm was adequately clipped in 43 patients, whereas the ruptured aneurysm dome could be excluded in 2 patients, but there remained significant proximal aneurysm rests, which were subsequently coiled (Figure 4). When multiple aneurysms were found, if the clinical conditions were stable, and the aneurysms were manageable through the same approach, they were clipped at the same time. A ventricular drainage with external cerebrospinal fluid (CSF) diversion was placed in 46 patients (93.8%). All patients were routinely managed with osteodural decompression. The bone flaps were either placed in a subcutaneous abdominal pocket or stored in the tissue bank awaiting subsequent replacement. The main treatment modalities are summarized in Table 1.

Postoperative Course and Management Postoperatively, all patients were managed in the neurosurgical intensive care unit with intracranial pressure (ICP) monitoring, and invasive hemodynamic monitoring. Postoperative CTA or DA was planned as soon as possible. Within a few days, severe vasospasm was documented in 19 patients (38.7%), but vasospasm to a greater or lesser degree was evident in virtually all cases. Accordingly, depending on the severity of vasospasm, aggressive or moderate triple-H therapy (hypervolemia, hypertension, hemodilution) was given. Eight patients also underwent endovascular treatment including transluminal balloon angioplasty and intraarterial vasodilator infusion therapy.

www.WORLDNEUROSURGERY.org

3

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

Figure 1. A: Preoperative CT scan showing a left temporal ICH without SAH but with initial brainstem compression. This 53-year-old man arrived in the ED within one hour of stroke. His GCS score was 8, the Hunt-Hess Grade was IV. B: Concomitant CT-angiography with 3D reconstructions showing a poli-lobulated aneurysm on the bifurcation of the left MCA (arrow). The patient underwent surgery within 90 minutes of his arrival at the ER. Six months later, the patient presented mild dysphasia but was fully independent.

Early postoperative course was challenging in all cases. Thirteen patients required postoperative temporary tracheostomy, which became definitive in just one case. Digiunostomy was needed in two cases. Severe general complications were reported in 21 patients (42.8%): there were 15 cases of sepsis, with 1 case of severe endocarditis that required cardiac surgery to remove endocardial vegetations; 4 patients required abdominal surgery because of pelvic hematoma, mesenteric infarction, hiatal hernia, and intestinal obstruction, respectively; electrocardiographic alterations were relatively frequent; but true coronal ischemia and serious dysrhythmia were reported in just 2 patients. The external CSF diversion was removed as soon as possible. Apart from cases of positive CSF cultures, bacterial CSF infections were also considered when pleocytosis, protein, and glucose levels alterations were documented. Accordingly, 16 patients (32.6%) were found to be infected and were opportunely treated by antibiotics. Eventually, 19 surgical survivors (38.7%) required a ventriculoperitoneal shunt. Thirty patients (61.2%) underwent cranial bone piece replacement: 11 patients during the same hospitalization and 19 patients through a subsequent re-admission. No cranioplasty was performed in the remaining 19 patients (38.8%), because of death (16 cases) or clinical conditions (three cases). Finally, 33 survivors (67.4%) were discharged to neurorehabilitation centers. Statistical Analysis All data were described by the usual statistical methods, based on the distribution of variables. Categorical variables were described by frequency tables, and continuous variables were given as mean  standard deviation or as median and interquartile range, respectively, if they were normally distributed or not. Normality distribution was verified by visual inspection of the graph and the ShapiroeWilk test. Comparisons of categorical variables were

4

www.SCIENCEDIRECT.com

performed with Fisher’s exact test; comparisons of continuousvalued variables were completed with Student’s t test (adjusted for unequal variances with Welch’s method) or the ManneWhitney U test. A logistic regression model was used to assess the effects of continuous and categorical variables on various binary outcomes. The dichotomization of continuous variables in search of possible cut-offs was performed using ROC analysis with Youden’s J statistics; P < 0.05 indicated statistical significance. RESULTS Clinical Results Overall mortality consisted of 16 patients (32.6%). Eight patients died within 1 week, mainly because of the primary damage. Indeed, 1 of these 8 patients died because of massive cerebral infarction attributed to intraoperative left ICA occlusion. These cases were reviewed as “early mortality.” Eight adjunctive patients survived longer but ultimately died because of the primary damage in association with other neurologic or general complications. No case of death was definitely and directly ascribable to vasospasm, although vasospasm was common in severely ill patients who ultimately died. One patient remained in a persistent vegetative state. Morbidity to a greater or lesser degree was reported in all the remaining patients. This was mainly ascribed to the primary damage, but a role was undoubtedly played by vasospasm in 3 patients, by CSF infection and sepsis in 2 patients, and by the surgical obliteration of the MCA in 1 patient. On discharge, the 33 survivors were evaluated according to the Glasgow Outcome Score (GOS): 1 patient was in a persistent comatose state (GOS ¼ 2), 21 patients presented severe neurological deficits (GOS ¼ 3), and 11 patients complained of moderate disability (GOS ¼ 4).

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2015.08.082

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

Figure 2. A: Preoperative CT scan showing a midline frontal ICH with a small amount of intraventricular blood. Two years before, this 62-year-old woman had complained of SAH because of an ACoA aneurysm, which had been completely excluded by coiling. Now, she arrived at the ED in a comatose state within 90 minutes of a new stroke. Her GCS score was 6, the Hunt-Hess Grade was V. B: Concomitant CT-angiography with 3D reconstructions failing to show any aneurysm regrowth. The artifacts caused by the previous coiling impedes the aneurysm visualization. Clot evacuation with exploration of the ACoA region was decided. The ED-OR interval was 80 minutes. As expected, a bleeding aneurysm was present on the right A1-A2 angle. The coils occupied the distal part of the aneurysm, whereas the neck was free and could be easily clipped. C: Postoperative digital angiography showing the aneurysm is completely excluded. A certain degree of vasospasm is evident in some distal vessels. Fifteen days later, this patient was referred to a rehabilitation centre. Six months later, she was heavily disabled and dependent for almost all daily activities.

Six months after surgery, following adequate rehabilitation, the overall results were reevaluated: GOS 1e2 ¼ 17 patients (34.6%); GOS 3 ¼ 10 patients (20.4%); GOS 4e5 ¼ 22 patients (44.8%; Table 2). Statistical Results In this series, there was a female preponderance, and all patients were within a short age range. Neither sex nor age affected the results, and there was no significant gender-related difference in age distribution (Welch’s test, P ¼ 0.253). The ICH volume presented no significant differences between the right and the left side (ManneWhitney U-test, P ¼ 0.236). Furthermore, no significant association was found between the worst GCS and the presence of acute hydrocephalus (Fisher’s exact test, P ¼ 0.569), and between the presence of acute hydrocephalus and the ultimate need for ventriculoperitoneal shunts. The influence of several factors on early mortality and on the negative six-month outcome (GOS 1e3) was analyzed in a logistic

WORLD NEUROSURGERY - [-]: ---, - 2015

regression model with the Wald’s test: a significant effect of age, gender, GCS, Hunt-Hess Grading system, ICH location, side, acute hydrocephalus, spontaneous rebleeding, aneurysm size, vasospasm, and general complications was not evident (Table 3). Conversely, the ICH volume was related to early mortality (P ¼ 0.028); in particular, each unitary volume increase corresponded to a 3% increase in the odds of early death. In this regard, the optimal cutoff for the volume was 100 ml (with Youden’s J statistics). In other words, patients with ICH volume greater than or equal to 100 ml presented a 970% increase in the odds of early death in comparison with patients with ICH volume less than 100 ml (median unbiased test estimates: P ¼ 0.015). On the other hand, the overall significance of ICH volume on the six-month bad outcome (P ¼ 0.107) was lost. However, an optimal cutoff for the six-month GOS 1e3 could be also found at 150 ml (with Youden’s J statistics). Accordingly, although the statistical tests failed to demonstrate that an ICH volume greater than or equal to 150 ml was significant regarding the odds of negative six-month results

www.WORLDNEUROSURGERY.org

5

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

Figure 3. A and B: Preoperative CT scan showing a large left fronto-temporal ICH presumably owed to the rupture of a partially visible giant aneurysm. A marked midline shift is evident. This 58-year-old man suddenly became comatose while he was visiting a friend at the hospital. His GCS was 3; the Hunt-Hess Grade was V. The clinical conditions were rapidly deteriorating and immediate empiric surgical exploration was decided. The ED-OR interval was 30 minutes. A giant aneurysm of the left ICA was found and excluded by multiple clips, through a temporary trapping of 15 minutes. C: Postoperative digital angiography: the aneurysm is completely excluded and the parent artery is patent. In any case, clinical improvement never occurred and the patient died within a few hours.

(median unbiased test estimates: P ¼ 0.499), this volume had a diagnostic value at least. The midline shift on the preoperative CT-scan significantly affected early mortality (Wald’s test: P ¼ 0.020). The optimal cutoff for the midline shift was more than 10 mm, and showed 85.7% sensitivity and 73.8% specificity. However, even a midline

6

www.SCIENCEDIRECT.com

shift between 5 and 10 mm was identified as a possible optimal cutoff for early death, with 100% sensitivity and 21.4% specificity. On the other hand, the value of the midline shift did not determine negative six-month outcome (likelihood ratio test: P ¼ 0.0817), even when this was considered as a continuous value (Wald’s test: P ¼ 0.070).

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2015.08.082

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

Figure 4. A, B and C: Preoperative CT-angiography with 2D reconstructions showing a poli-lobulated ACoA aneurysm pointing upward, downward and posteriorly. A large ICH is also visible. This 61-year-old man arrived at the ED within two hours of stroke. His GCS was 5; the Hunt-Hess Grade was V. The ED-OR interval was 60 minutes. The aneurysm consisted of two different sacs: the one pointing upward was relatively easily clipped, whereas the exposure of the sac pointing toward the hypothalamus was hampered by its relationship with the perforators. Complete exclusion was deemed hazardous and a clip was placed to protect the bleeding dome, deliberately leaving a rest near the neck. D: Postoperative digital angiography confirming the aneurysm remnant. Treatment completion by coiling was postponed awaiting better clinical conditions. Eight days later, an autonomously breathing and alert patient suddenly worsened. A repeated CT scan showed intraventricular rebleeding. The aneurysm rest was immediately coiled. E: Final digital angiography showing the complete aneurysm exclusion and a certain degree of vasospasm. Six months later, the patient was alert but severely obtunded and he was permanently housed in a center for severe disabled people

The duration of the interval between admission to the ED and access to the OR was independent of both the GCS (Wald’s test: P ¼ 0.161) and the ICH volume (Wald’s test: P ¼ 0.596). However, neither early mortality nor the six-month unfavorable GOS was affected by the ED-OR interval, provided that this was less than 2000 minutes (Wald’s test: P ¼ 0.269 and P ¼ 0.509 respectively).

WORLD NEUROSURGERY - [-]: ---, - 2015

Indeed, the patients who died early tended to have nearly significant shorter ED-OR times (Mann-Whitney U test: P ¼ 0.059). This borderline effect was lost when the six-month adverse outcomes were analyzed (Mann-Whitney U test: P ¼ 0.240). In fact, there was a significant association between the worst GCS and empirical exploration without preoperative angiography (Fisher’s

www.WORLDNEUROSURGERY.org

7

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

Table 2. Overall Results Evaluated According to the Glasgow Outcome Score Glasgow Outcome Score Number of Patients

1

2

3

4

5

Hospital discharge

16

1

21

11

—

6-month follow-up

16

1

10

14

8

exact test: P < 0.001), whereas the ICH volume appeared to have no effect on the decision to abbreviate the time for preoperative assessments (Wald’s test: P ¼ 0.156). Early mortality was significantly more frequent in patients who underwent cistern exploration without preoperative angiography (Wald’s test: P ¼ 0.016); in particular, if the patients underwent at least CTA, there was an 89% decrease in the odds of early death; conversely, the absence of preoperative studies did not affect the six-month negative GOS (Wald’s test: P ¼ 0.416). The results of clipping with concomitant ICH evacuation were compared with the results of coiling with subsequent ICH

evacuation. Although clipping provided an 85% reduction in the odds of early death, it did not reach the statistical significance level (P ¼ 0.101). Conversely, the six-month outcome was significantly influenced by the treatment modality (P ¼ 0.027): in comparison with coiling followed by clot evacuation, aneurysm clipping and concomitant ICH evacuation reduced the odds of unfavorable six-month GOS by 89.9%. Finally, we investigated the relationship between the presence of acute hydrocephalus and the development of subsequent chronic hydrocephalus. Three-quarters of the survivors with initial acute hydrocephalus required a permanent VP shunt in comparison with nearly half of the survivors without acute ventricular enlargement. However, these figures were not statistically significant (Fisher’s exact test: P ¼ 0.087). DISCUSSION Epidemiology The incidence of ICH in the case of aneurysm rupture ranges from 4% (in low-grade patients) to 25% (in high-grade patients), and up to 84% in autopsy series.3,8,10-12,14-17 Intracerebral rupture is related to the rapid obstruction of the subarachnoid space by blood, and/or the adhesion of the aneurysm sac to the pia

Table 3. Unfavorable Outcome: Statistical Significance of Potentially Influencing Factors* Potentially Negative Predictors

Early Death (P Value)

GOS 1e3 at 6 months (P Value)

Age (years; per any unitary increase)

0.386

0.395

Sex (male versus female)

0.630

0.585

Initial Glasgow Coma Sscore (score; per any unitary increase)

0.272

0.803

Initial Hunt-Hess (score; per any unitary increase)

0.531

0.418

Intracerebral hematoma location Lobar versus multilobary

0.112

0.705

Frontal versus temporal

0.688

0.531

0.078

0.204

Side (left versus right) Spontaneous rebleeding (yes versus no)

0.981

> 0.99

Intracerebral hematoma volume (ml; per any unitary increase)

0.028 (significant)

0.107

Midline shift (per any unitary increase from minimal [<5mm], to intermediate [6e10 mm], to maximal [>10 mm])

0.020 (significant)

0.070

Acute hydrocephalus (yes or no)

0.99

0.395

ED-OR interval (minutes; per unitary increase)

0.269

0.509

Surgical exploration without preoperative angiography (yes or no)

0.016 (significant)

0.416

Treatments other than clipping (yes or no)

0.101

0.027 (significant)

Aneurysm size (mm; per unitary increase)

0.316

0.237

General complications (yes or no)

0.330

0.096

Vasospasm (yes or no)

0.164

0.99

*All P values were obtained with Wald’s test after running a logistic regression model. For the binary variables (e.g., those designed by “yes versus no”), a significant P value means a reduction of the odds of unfavorable event when the event “yes” is realized. In the case of significant P value for continuous variables, it is intended that any arbitrary unitary increment of the variable be associated with a significant increase of the odds of the unfavorable event. yLobar indicates frontal or temporal; multilobar indicates frontotemporal or frontotemporoparietal.

8

www.SCIENCEDIRECT.com

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2015.08.082

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

mater,9,10 and may even occur without any evidence of SAH.9 Aneurysms located close to the brain parenchyma without much intervening subarachnoid space (such as those on ACA, ACoA and MCA) are more predisposed to ICH.10,11,16 In fact, if we consider ICHs of any size, these may be found in 36% to 56% of hemorrhages from ACA and MCA, and in 1% of bleedings from posterior circulation aneurysms.4,9,10,12 Of course, the low incidence of posterior aneurysms in clinical series may also depend on higher immediate mortality for brain-stem hemorrhages. The MCA aneurysms are responsible for 70% of large ICHs.4 A significant association has been reported between the size of the ICH and the size of the aneurysm4,10: it has been speculated that leaks may be correlated with the aneurysm size.4 Our series included only large ICHs (>60 ml): in fact, half of the cases were MCA aneurysms. In this series, there were four giant aneurysms (>2.5 cm), which accounted for 8.1%. Moreover, the vast majority of the bleeding aneurysms had a diameter ranging between 5 and 15 mm. Accordingly, the overall aneurysm sizes appeared larger in this ICH series than in our global series of SAH.18 These findings paralleled those by Lee and colleagues,5 whose patients with large ICHs had aneurysms with mean diameter of 11 mm. Clinical Grade High clinical grades have been reported in two-thirds of patients with ICH and in less than one-third of patients with pure SAH.12,19 The presence of an ICH has been considered a negative prognostic indicator in both low- and high-grade patients,6,9,14,20,21 and significantly worse results were reported in high-grade patients with associated ICHs.6,15-17,19,20,22,23 However, this has been not confirmed in some recent papers,5,10,12 reporting comparable results among high-grade patients with and without ICH, within the same neurological grade. According to Abbed and Ogilvy,12 poor outcome is related to the initial poor clinical state rather than to the presence of the ICH per se. Conversely, other authors4,9,22,24,25 report that the initial neurological conditions are weaker predictors of the destiny of high-grade patients with ICH than other factors such as the ICP value, the timing of treatment, the severity of the bleeding, and so on. Inagawa and colleagues26 documented that initial GCS was an excellent predictor for the case-fatality rates only for patients with pure SAH, whereas its importance was lost for patients with associated ICHs. We found that the prognostic value of the high grades was timedependent because patients who appeared very critically stricken during the earliest phase could improve a few hours later: Grade V patients at six hours after SAH had more chances of a favorable outcome than Grade V patients at the 48th hour.25 In the series by Le Roux and colleagues,6 the only factor influencing the outcome was neurological improvement within 72 hours. In the present material, we did not find a significant statistical influence of the initial neurological conditions upon the outcome. Perhaps this was because all patients presented with severe neurological conditions and were managed within a few hours of stroke. The vast majority of patients were Grade V,13 and we obtained quite different results even within the same grade. The GCS appeared more precise in defining the clinical severity, but we found no statistical differences in the outcomes between the

WORLD NEUROSURGERY - [-]: ---, - 2015

ICH DUE TO ANEURYSM RUPTURE

higher and the lower scores. Indeed, the GCS was evaluated as soon as the patients were seen. Accordingly, it is possible that the present series included patients who were considered more compromised than they really were simply because their evaluation was made too early. In this series of ICH, however, we obtained even better results than in our past series of highgrade SAH patients.25 We do not think that this related only to improved management and experience. Probably the prompt removal of the mass effect of the ICH played a favorable role, differently from high grades without ICH. In general, the bad prognosis of high grades depends on the combination of the primary damage with the higher incidence of several complications such as brain swelling, edema, hydrocephalus, rebleeding and vasospasm.4,10,11,13,16,17,23,27 However, it has been suggested that there may be patients who benefit from targeting the predominant pathological process (i.e. ICH evacuation), resulting in improved outcome,27 and better management of the possible complications.6,18,22,25,27 We are not denying the importance of neurologically-based grading systems. In a recent series of aneurysmal ICHs, favorable outcome was 25% in low grades and 12.8% in high grades.4 However, we agree with Batjer and Samson1 that these grading scales are less applicable to patients with life-threatening ICHs after aneurysmal rupture. Size of Hematoma Some decades ago a number of authors15-17,20,23 advised against active treatment in the case of ICH larger than 3 cm. It is now recognized, however, that favorable outcomes are possible even in the presence of large ICHs.1,2,4,6,9,10,22,25 In general, larger ICHs are related to higher clinical grades and less favorable outcomes.11 However, Tokuda and colleagues10 reported similar outcomes in ICHs of different sizes but under the same clinical grades. Shimoda and colleagues9 found that the ICH volume would be less important than the midline shift and would be not related to outcome when considered alone.9 By contrast, Lee and colleagues5 and Brott and Mandybur24 considered the ICH volume and the hemorrhage extension as the stronger negative predictors. The cutoff volume for the mortality would be 40e50 ml,4,10,16 but these values have been often derived from series including small clots too,4,8,9,25 and it seems quite obvious that patients with ICHs <25 ml can fare better than patients with larger clots as reported by Bohnstedt and colleagues.11 Our series included only patients with very large and lifethreatening ICHs, since we selected only those for whom the main indication for surgery was clot evacuation rather than aneurysm clipping. Therefore, our analysis did not compare small versus large ICHs but concerned only large clots. This probably accounts for the relatively higher volume cutoffs which we found in our series. The odds of early mortality increased parallel with the increase in the ICH volume, and satisfactory results were virtually impossible beyond a given threshold. Also, the degree of the midline shift was determinant for early mortality, but less important for the six-month outcome. From a practical point of view, the midline shift represented an indicator of acute severity rather than a determinant of the long-term results. In other words, the presence of midline shift indicated an immediate risk of death,

www.WORLDNEUROSURGERY.org

9

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

but it did not preclude the possibility of satisfactory final outcome. Similar findings are also reported by others.6,17 To sum up, in our series of large ICHs, the bleeding severity was far more important than the initial clinical severity. Location of Hematoma It has been reported that the ICH location may influence the prognosis.7,9,16,21,27 The outcome can be different even among the various types of temporal ICHs9: purely temporal clots usually spare the internal capsule, so that their prognosis is correlated neither with the volume nor with the initial neurological status, but only with the presence of brainstem compression. Conversely, the presence of sylvian ICH or the association temporal ICH with diffuse SAH represents a negative prognostic factor.7,9,21,27 Early surgery would ameliorate the results only in the case of large sylvian clots or small temporal ICHs.9,27 We cannot confirm such data, since we did not find a significant correlation between ICH location and outcome. Indeed, in a lot of cases, the clots were so large that they were multicompartmental so that there were really few differences between ICHs originating in different locations but sprouting elsewhere. In any case, it is our impression that, if we compared an aneurysmal ICH with a spontaneous ICH, we would find different behaviors. The former is often associated with SAH and looks more severe during the acute phase; its early mortality is probably higher, but it has more chances to spare the internal capsule, so the permanent morbidity might be lower. The latter may appear initially less severe, but the internal capsule is almost invariably damaged. Rebleeding It is well known that high-grade aneurysms have a greater tendency to rebleeding.4,10,13,17,27 This tendency seems even more accentuated in patients with aneurysmal ICH, in whom the rerupture rates are three times higher.12 The reason for this propensity is unclear. Perhaps, the rupture site is larger in patients with ICH; otherwise, since the aneurysms responsible for ICH tend to be larger than those causing simple SAH, the higher rebleeding rate could be related to the well-known higher risk of re-rupture carried by the larger aneurysms rather than by the presence of the ICH itself.4 In this series, we confirm the tendency to have both larger aneurysms and higher rebleeding rates. Some type of aneurysm rerupture occurred in almost all our patients. However, we must differentiate spontaneous rebleeding from re-ruptures owed to surgical manipulation. Spontaneous rebleeding consisted of eight cases (six preoperative and perhaps two during the craniotomy), which equates to 16.3%. This was not a negligible rate, especially if we consider that all patients were early admitted and early managed. The rate of manipulation rupture is even more impressive: more than 80% of cases. However, again, we must distinguish early from delayed hemorrhages. Early rebleeding accounted for 22% of cases, and was probably related to the abrupt change of the transmural pressure triggered by the simple dural opening or the initial clot evacuation. It was more dangerous because it often required drastic measures, such as arterial trapping or brain amputation. It represents one of the main problems of this type of surgery.4 The delayed hemorrhage was that occurring when the aneurysm was already adequately exposed

10

www.SCIENCEDIRECT.com

and was probably related to large rupture sites, large aneurysm sizes, and excess of surgical confidence. A more meticulous technique could perhaps have prevented this event, but it was always easy to overcome and never resulted in adjunctive morbidity or mortality. Surgical Indications In the past, the classical management of high-grade SAH patients was the postponement of surgery, in the hope of spontaneous improvement.15 Even in the face of life-threatening ICHs, early clipping was not considered, and either the patients were considered entirely hopeless, or the ICH was removed leaving the aneurysm untouched, which carried inacceptable mortality rates.17,23 In the 1980s, a randomized prospective study demonstrated that ICH evacuation with concomitant clipping could reduce mortality by two-thirds.21 Present rates of postoperative mortality are relatively low,8 but the rate of bad functional outcomes of the survivors must be taken into account.11 Although timely removal of mass effect and prevention of the rebleeding now represent the mainstay of treatment,2-4,6,8,9,12,1618,21,25 indications for surgery should take into consideration a number of ethical, social and economic aspects.2,11 In a recent multivariate analysis, Shimoda and colleagues9 tried to refine the surgical indications by the identification of the favorable outcome indicators, but the prediction of which patients could really benefit from surgery remained obscure. In our practice, we are used to managing aneurysmal ICHs as a sort of traumatic hematoma. In general neurosurgical practice, there is relatively little hesitation about performing surgery in large traumatic ICHs in deeply comatose patients. To our mind, there was no reason for a different indication in aneurysmal ICHs of the same size, location, and mass effect. Of course, following the present retrospective review, we realize that our surgical indications were sometimes too aggressive. This series probably also included patients on whom we would not now operate. We have now refined our indications for aneurysm ICHs and we are reflecting on less aggressive indications also for similar cases of traumatic origin. We believe that the ethical, social, and economic perplexities should be the same, but we wonder whether the legal implications of the traumatic ICHs play a role in pushing hospitals towards treatments that are more aggressive for head-injured patients. Preoperative Assessments In these moribund patients, preoperative DA (even a single-vessel DA) could cause a life-threatening delay,6,19 and therefore empiric cistern exploration may be indicated depending on the CT pattern.1,2 However, there are aneurysmal ICHs involving only the basal ganglia, as well as hypertensive ICH extending to the sylvian fissure, mimicking each other. Therefore, the empiric exploration should represent an exception even in specialized neurovascular contexts. We found a close relationship between bad outcomes and empiric exploration, and the odds of early mortality were significantly higher in patients not adequately studied prior to surgery. In fact there was a strict correlation between the worst GCS and empiric exploration. Conversely, the ICH volumes did not significantly influence the indication for empiric exploration. Of course,

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2015.08.082

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

there could be other reasons to skip the preoperative assessments: the temporary lack of neuroradiological facilities, underestimation by the physicians who initially managed the patient, and so on. Nowadays, the availability of CTA allows the prompt diagnosis of 97% of aneurysms without significant delay.4,19 Nevertheless, DA still remains the gold standard for aneurysm treatment, and should be considered whenever possible. CTA may adequately depict the aneurysm neck relationships, but DA still provides superior definition of the circulation times and the collateral circulation. This series included a couple of patients regarding whom, relying only on CTA, we found it very difficult to understand whether ICA was expendable or whether a P-com was sustaining the posterior circulation. To sum up, the avoidance of excessive delay is important, but whether or not there is time for assessments depends mostly on the evaluation and the experience of the attending neurosurgeon.19 Timing In a recent paper,4 the timing of surgery was the most important determinant of the outcome in patients with aneurysmal ICH. Although some authors still consider the mortality rate of acute surgery excessive,10,11 others agree upon the advantages of ultraearly surgery, which allows the rapid prevention/release of herniation, the reestablishment of cerebral perfusion, and the decrease of ICP,12,16,21,23 as well as the prevention of rebleeding.3,4,10 Moreover, even the morbidity would be better thanks to significantly fewer incidences of vasospasm and better functional outcome through the limitation of the secondary damage.7 The time limit on obtaining favorable results has been variously reported as ranging from 3.5 hours,4 to 6 hours7,9 up to 12 hours.11 In our series, we could not find any correlation between the brevity of the ED-OR interval and improved results. On the contrary, although the significance was not reached, there was a clear tendency towards higher early mortality in patients with shorter ED-OR intervals. We think that this is just an apparent contradiction of the importance of ultra-early surgery. All our patients harbored life-threatening ICHs, so the preoperative times had to be the shortest possible. A couple of hours represented the mean duration of the ED-OR interval and this interval had so strict a range that correlations were difficult. In general, the rapidly deteriorating patients were referred to the OR without any delay, but these patients were so severely stricken that in any case they paid a high price in terms of mortality. Notwithstanding that, we found that neither the GCS nor the ICH-volume influenced the duration of the ED-OR interval. Of course, the interval could be longer both in the less compromised patients, who could undergo more accurate preoperative studies, and in more compromised and unstable patients, who require longer preoperative preparations. Although we remain persuaded of the importance of ultra-early surgery, our data failed to show that the shorter the ER-OR interval, the better the results. We must admit that we were sometimes too aggressive: if we excluded the eight patients who died early mainly because of the primary damage, significance was not reached, but a trend towards better outcomes in rapidly managed patients could be seen.

WORLD NEUROSURGERY - [-]: ---, - 2015

ICH DUE TO ANEURYSM RUPTURE

Surgical Technique Aneurysm exposure may represent a major problem because of brain edema and the mass effect of ICH.21 A common technique first consists of entering and partially evacuating the clot to slacken the brain and to gain the aneurysm exposure, which may be relatively easy when the ICH is superficial, as in most MCA lesions, but quite difficult when it is deep as in some ACoA or ICA aneurysms.21 In the present series, in all cases, the aneurysm could be accessed through partial clot evacuation, CSF withdrawal (by ventricular drainage and/or cistern dissection), brain amputation (polectomy or lobectomy), and so on. We did not use any particular personal technique to clip the aneurysm. In contrast with the past, in the present series, we did not feel obliged to pursue the complete aneurysm obliteration at almost any cost, because present endovascular techniques offer a valuable option to complete surgery, if needed. In a couple of patients, we intraoperatively deemed the complete aneurysm exclusion as too hazardous. Accordingly, following aneurysm dome protection by partial clipping, we referred the patients to the interventional neuroradiologists for the completion of aneurysm obliteration. Needless to say, the aneurysm rest must be secured as soon as possible: one of our patients experienced severe rebleeding while he was waiting for the endovascular completion of his clipping (Figure 4). Besides aneurysm securing, surgery obviously aims to obtain the rapid and continued control of elevated ICP,22,25,27 which may be accomplished by clot evacuation, CSF withdrawal, brain amputation, dural grafting, and/or decompressive craniectomy. The last may provide immediate and sustained control of ICP or may be prophylactically performed to prevent delayed brain swelling.7,27 Routine prophylactic decompressive craniectomy has been also advocated because dangerous local ICP increases would be possible before detection by ICP monitoring.27 Should vasospasm develop, it enables a more aggressive hypervolemia, which would be problematic if the persistence of high ICP required mannitol and dehydration.9,27,28 Furthermore, planning a prophylactic craniectomy obviates the need for extended clot removal with excessive manipulation of edematous brain. As correctly outlined by Lawton,28 the operating time is not prolonged if the bone is stored in a tissue bank and the subsequent operation for bone replacement is usually easy and low risk. We used prophylactic osteo-dural decompression in almost all cases which did not reveal particular adjunctive problems. Decompressive craniectomy has been proven to be a lifesaving procedure.9 However, we still wonder whether can really improve the results, or is simply associated with an increased rate of survivors in vegetative or extremely compromised states. Role of Endovascular Treatment Aneurysm coiling prior to ICH evacuation has been suggested to limit the rebleeding rate and to reduce surgical complications by obviating the need for extensive dissection and retraction to expose the aneurysm neck, thus making the subsequent clot removal, easier, safer, and faster.3,8 The obvious delay in brain decompression because of the endovascular treatment would not affect the outcome.3,8 Recently, de los Reyes and colleagues3

www.WORLDNEUROSURGERY.org

11

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

compared patients managed by coiling followed by ICH removal versus those undergoing clipping and concomitant ICH removal: despite the time prior to ICH removal was almost double in coiled patients, the mortality, the complication rates and even the costs were quite similar. Indeed, in that series, coiling represented the first choice in all cases, and the surgically treated aneurysms were only those considered not suitable for coiling. Accordingly, it seems probable that surgery had to deal with the more complex cases. Furthermore, the slight mortality decrease (30% versus 25%) obtained by coiling prior to surgery appeared to be counterbalanced by an increased rate of poor outcomes (70% versus 50%). Theoretically, a possible management strategy would be partial clot removal followed by immediate aneurysm coiling. This would allow prompt brain decompression without excessive brain manipulation. However, to our mind, the aforementioned tendency to early and intraoperative rebleeding does not make this strategy advisable. In our series, the ICH volumes and the clinical conditions did not usually leave other possibilities than immediate brain decompression. However, in a few less compromised patients, coiling prior to surgery was attempted. Despite a slightly longer ED-OR interval, coiled and clipped aneurysms presented no significant differences in the rate of early mortality, although the odds of early death appeared to be reduced in patients undergoing clipping. Conversely, the six-month results were significantly better in clipped aneurysms, with 90% reduction of the odds of unfavorable results. The initial conditions of our patients were probably so severe that their short-term destiny was poorly influenced by the initial treatment. However, if the patients were able to get over the initial crucial phase, they achieved better outcomes only if their brain was promptly decompressed. Postoperative Course A very challenging course was virtually the rule, but we did not find significant differences between high-grade SAH patients with and without ICH.25 We rather received the impression that the prompt removal of the ICH mass effect may sometimes give a smoother postoperative course. In the case of ICH, the mass effect owed to the clot represented a removable cause of the poor conditions, with a chance to limit the secondary damage. Conversely, in the case of pure SAH, the main determinant of the poor conditions was the unmanageable primary damage. Vasospasm has frequently been reported in these patients,4,10,11,13,16,17,23,27 although its clinical relevance may be hard to assess in comatose or deeply sedated patients.7,27 Instrumental findings compatible with more or less severe degrees of vasospasm were present in almost all our patients, but we considered the vasospasm as clinically relevant only in the presence of unequivocal neurological worsening and/or clear CT findings. Although this could lead to underestimation, the incidence of vasospasm in our series was not negligible, above all if we take into account that those patients who died early probably did not have time to develop it. In any case, the vasospasm did not significantly affect the results. Modern effective therapeutic tools undoubtedly played a role, but many of our patients were already so compromised by the primary damage that little remained to be added by the vasospasm in terms of both mortality and morbidity. Similar deductions could be drawn for the ventricular infections, too: CSF

12

www.SCIENCEDIRECT.com

infection occurred in one-third of our patients if we consider the overall series, but in one-half of cases, if we exclude the patients who died early. This relatively high rate probably related to the comatose and very delicate state of these patients, who generally required prolonged external drainage in the contaminated environment of the neurosurgical intensive care unit. Indeed, it is well known that the range of infections may be quite different in different series ranging from 0% to 45% of cases.29 It would depend on the variety of populations evaluated, the use of antibiotic prophylaxis, the technique of catheter placement, the methods of CSF sampling, and the definition of infection adopted (for instance, we considered CSF as infected when either pleocytosis or abnormal CSF protein or glucose levels were evident, even though the CSF cultures failed to show microorganisms). High-grade patients are notably predisposed to a number of general complications, often affecting the outcomes. In spite of several cardiac, abdominal, infectious, and respiratory troubles, we did not find a significant influence on either early mortality or six-month GOS. These complications usually occurred in the more compromised patients who were usually those with the larger ICH. Therefore, the ICH volume likely concealed the influence of other factors on the unfavorable results. Acute hydrocephalus was relatively frequent in this series, but we found no significant association between early ventricular enlargement and the initial GCS, which means that the hydrocephalus played a minor role in sustaining the comatose state of our patients. Moreover, the presence of acute hydrocephalus did not significantly affect the results, probably because of the early admission and the prompt ventricular drainage in all cases. The vast majority of the survivors with initial hydrocephalus ultimately required a permanent shunt, but the presence of acute hydrocephalus did not result in a significant increase of the risk of chronic hydrocephalus. In any case, most of the survivors required a permanent VP shunt. The routine use of ventricular drainage for ICP management and the frequent presence of intraventricular blood could have led to an increased rate of CSF pathways block. In these particular problematic patients, we found it difficult to evaluate the actual influence on the outcome exerted by the chronic hydrocephalus. There was no related mortality, but we were not able to quantify the morbidity at least as long as the shunt kept on working correctly. It is well known that the mid-term outcomes of SAH patients are often better than the short-term results. In our series, all deaths occurred during the hospital stay, and a considerable number of patients improved neurologically following their discharge. Namely, the ICH volume and its mass effect played the leading role as outcome determinants. These two factors were so important that they killed one-third of patients during the earliest phase. However, if the patients could get over this initial crucial phase, then the chances of survival and improvement were not negligible. CONCLUSIONS Severely compromised patients with ICH owed to aneurysm rupture need correct and prompt management by an efficient and skilled neurovascular team.19 Of course, this approach requires a significant commitment of resources, beginning with an efficient paramedic service practicing in-the-field

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2015.08.082

ORIGINAL ARTICLE GIUSEPPE TALAMONTI ET AL.

ICH DUE TO ANEURYSM RUPTURE

resuscitation and ultra-early referral to a center with an interest and expertise in acute aneurysm surgery.6 Unfortunately, the increasing use of endovascular coiling over the last decades has led to the decreased experience of neurosurgeons of more difficult cases. The effort required, however, to manage these apparently hopeless patients may be repaid by the possibility of achieving satisfactory results in a significant number. Even patients with the worst neurological conditions and pupillary anomalies may be saved.3,6,17 Therefore, the identification of the subgroup of salvageable patients is mandatory. In our relatively small series, we identified a cutoff for the ICH volume which could help to refine the surgical indications. When a neurosurgeon is called to take decisions about these patients, a number of unresolved social, economic, and ethical problems

REFERENCES 1. Batjer HH, Samson DS. Emergent aneurysm surgery without cerebral angiography for the comatose patient. Neurosurgery. 1991;28:283-287. 2. Brandt L, Sorenson B, Ljunggren B, Saveland H. Ruptured middle cerebral artery aneurysm with intracerebral hemorrhage in younger patients appearing moribund: Emergency operation? Neurosurgery. 1987;20:925-929. 3. de los Reyes K, Patel A, Bederson JB, Frontera JA. Management of subarachnoid hemorrhage with intracerebral hematoma: clipping and clot evacuation versus coil embolization followed by clot evacuation. J Neuro Intervent Surg. 2013;5:99-103. 4. Guresir E, Beck J, Vatter H, Setzer M, Gerlach R, Seifert V, et al. Subarachnoid hemorrhage and intracerebral hematoma: incidence, prognostic factors, and outcome. Neurosurgery. 2008;63: 1088-1094. 5. Lee CS, Park JU, Kang JG, Lim YC. The clinical characteristics and treatment outcomes of patients with ruptured middle cerebral artery aneurysms associated with intracerebral hematoma. J Cerebrovasc Endovasc Neurosurg. 2012;14:181-185. 6. Le Roux PD, Dailey AT, Newell DW, Grady MS, Winn HR. Emergent aneurysm clipping without angiography in the moribund patient with intracerebral hemorrhage: The use of infusion computed tomography scans. Neurosurgery. 1993;33:189-197. 7. Mutoh T, Ishikawa T, Morol J, Suzuki A, Yasui N. Impact of early surgical evacuation of sylvian hematoma on clinical course and outcome after subarachnoid hemorrhage. Neurol Med Chir. 2010; 50:200-208. 8. Niemann DB, Wills AD, Maartens NF, Kerr RS, Byrne JV, Molyneux AJ. Treatment of intracerebral hematomas caused by aneurysm rupture: coil placement followed by clot evacuation. J Neurosurg. 2003;99:843-847. 9. Shimoda M, Oda S, Mamata Y, Tsugane R, Sato O. Surgical indications in patients with an intracerebral hemorrhage due to ruptured middle cerebral artery aneurysm. J Neurosurg. 1997;87:170-175.

rests on his or her shoulders. However, this is not so different from reaching decisions about treatment in neuro-traumatology or in neuro-oncology. We think that, if one has little hesitation about managing severe head injuries, the same should hold true for aneurysm ICHs of the same size, location, and neurological state. ACKNOWLEDGMENTS The authors thank the women and men of the emergency room and the neuroradiological service: their precious collaboration made possible the management of the patients. Special thanks go to the women and men of the neurosurgical intensive care unit: they were the ones who really looked after the patients.

10. Tokuda Y, Inagawa T, Katoh Y, Kumano K, Ohbayashi N, Yoshioka H. Intracerebral hematoma in patients with ruptured cerebral aneurysms. Surg Neurol. 1995;43:272-277. 11. Bohnstedt BN, Nguyen HS, Kulwin CG, Shoja MM, Helbig GM, Leipzig TJ, et al. Outcomes for clip ligation and hematoma evacuation associated with 102 patients with ruptured middle cerebral artery aneurysms. World Neurosurg. 2013;80: 335-341. 12. Abbed KM, Ogilvy CS. Intracerebral hematoma from aneurysm rupture. Neurosurg Focus. 2003;15: E4. 13. Hunt WE, Hess RM. Surgical risk related to time of intervention in the repair of intracranial aneurysms. J Neurosurg. 1968;24:14-19. 14. Crompton MR. Intracerebral hematoma complicating ruptured cerebral berry aneurysm. J Neurol Neurosurg Psychiatr. 1962;25:378-386. 15. Papo I, Bodosi M, Dorci T. Intracerebral hematomas from aneurysm rupture: Their clinical significance. Acta Neurochir. 1987;89:100-105. 16. Pasqualin A, Bazzan A, Cavazzani P, Scienza R, Licata C. Da Pian R: Intracranial hematomas following aneurysmal rupture: Experience with 309 cases. Surg Neurol. 1986;25:6-17. 17. Tapaninaho A, Hernesniemi J, Vapalahti M. Emergency treatment of cerebral aneurysm with large hematomas. Acta Neurochir. 1988;91:21-24. 18. Collice M, D’Aliberti G, Talamonti G. Current indications for aneurysm surgery. Neuroim Clin North Am. 2006;16:497-512. 19. Bergdal O, Springborg J, Hauerberg J, Eskesen V, Poulsgaard L, Romner B. Outcome after emergency surgery without angiography in patients with intracerebral haemorrhage after aneurysm rupture. Acta Neurochir. 2009;151:911-915. 20. Auer L. Unfavorable outcome following early surgical repair of ruptured cerebral aneurysms: A critical review of 238 patients. Surg Neurol. 1991;35: 152-158. 21. Heiskanen O, Poranen A, Kuurne T, Valtonen S, Kaste M. Acute surgery for intracerebral

WORLD NEUROSURGERY - [-]: ---, - 2015

hematomas caused by rupture of an intracranial arterial aneurysm. A prospective randomized study. Acta Neurochir. 1988;90:81-83. 22. Bailes JE, Spetzler RF, Hadley MN, Baldwin ME. Management morbidity and mortality of poor grade aneurysm patients. J Neurosurg. 1990;72: 559-566. 23. Wheelock B, Weir B, Watts R, Mohr G, Khan M, Hunter M, et al. Timing of surgery for intracerebral hematomas due to aneurysm rupture. J Neurosurg. 1983;58:476-481. 24. Brott T, Mandybur TI. Case-control study of clinical outcome after aneurysmal subarachnoid hemorrhage. Neurosurgery. 1986;19:891-895. 25. Versari PP, Talamonti G, D’Aliberti G, Villa F, Solaini C, Collice M. Surgical treatment of poorgrade aneurysm patients. J Neurosurg Sci. 1998;42: 43-46. 26. Inagawa T, Shibukawa M, Inokuchi F, Tokuda Y, Okada Y, Okada K. Primary intracerebral and aneurysmal subarachnoid hemorrhage in Izumo City, Japan. Part II: Management and surgical outcome. J Neurosurg. 2000;93:967-975. 27. Smith ER, Carter BS, Ogilvy CS. Proposed use of prophylactic decompressive craniectomy in poorgrade aneurysmal subarachnoid hemorrhage patients presenting with associated large sylvian hematomas. Neurosurgery. 2002;51:117-124. 28. Lawton MT. Comment. Neurosurgery. 2002;51:124. 29. Lyke KE, Obasanjo OO, Williams MA, O’Brien M, Chotani R, Perl TM. Ventriculitis complicating use on intraventricular catheter in adult neurosurgical patients. Clin Infect Dis. 2001;33:2028-2033.

Received 22 May 2015; accepted 29 August 2015 Citation: World Neurosurg. (2015). http://dx.doi.org/10.1016/j.wneu.2015.08.082 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2015 Elsevier Inc. All rights reserved.

www.WORLDNEUROSURGERY.org

13