Keywords
decompressive craniectomy - spontaneous intracerebral hemorrhage - stroke - mRs -
surgical outcome
Introduction
Spontaneous intracerebral hemorrhage (SICH) is one of the most devastating forms of
stroke with a mortality of 30 to 40%.[1] Following the STICH trials[2] interest in surgery for SICH had declined, and even to date there is no class 1
evidence on the ideal management of SICH with refractory raised intracranial pressure
(ICP). However, of late there is mounting evidence in literature that minimally invasive
surgery is associated with significantly improved outcomes when compared with conservative
treatment and conventional surgical evacuation strategy. There has been a revived
interest in decompressive craniectomy (DC) also over the past few years with several
large studies and meta-analysis highlighting the superiority of DC in intracerebral
hemorrhage (ICH) for reducing mortality. In fact the recent American Heart Association
guidelines on SICH recommend that DC with or without hematoma evacuation might reduce
mortality for patients with supratentorial ICH who are in a coma, have large hematomas
with significant midline shift (MLS), or have elevated ICP refractory to medical management
(class 2b; level of evidence C).[2] The rationale for DC is based upon the Monro-Kellie doctrine and the procedure was
first described by Kocher and Cushing in the early 1900s.[3] DC aims to enlarge the intracranial space and expand intracranial volume with the
restoration of perimesencephalic cisterns, decrease the MLS, and thereby improve cerebral
compliance[4] and reduce ICP.[5] The optimal size of the craniectomy, which results in maximal decompression without
increasing the risk of complications, is still an enigma. The size and volume of DC
have been studied in detail for ischemic stroke[6] and traumatic brain injury (TBI).[7] However, no attempt has been made to study the effect of craniectomy size in SICH.
No level 1 evidence exists regarding the ideal size of DC[8] with various studies suggesting that anteroposterior (AP) diameter of at least 8
cm[9] is adequate and others suggesting that at least 15 cm is mandatory.[10] Unlike in patients with trauma and ischemic stroke, where the primary insult results
in significant cerebral edema, in SICH hematoma evacuation causes significant brain
tissue collapse, which is noted intraoperatively. However, perihematomal edema, rebleed,
or residual bleed may lead to delayed deterioration and DC tends to have a therapeutic
role in preventing such secondary deteriorations.
Our study was based on the hypothesis that in patients with large SICH, as there tends
to be a collapse of brain tissue post hematoma evacuation, a large DC might not be
required, like in cases of trauma or ischemic stroke. Moreover, large DC may be harmful
as intraoperative brain collapse might lead to complications like shearing of bridging
veins. We aimed to evaluate the effect of craniotomy size and volume of decompression
achieved on surgical outcome, complications, mortality, and morbidity in patients
with supratentorial capsuloganglionic bleeds who underwent DC at our institute.
Methods and Methodology
We retrospectively reviewed records of all patients who underwent primary standard
DC with hematoma evacuation for unilateral spontaneous capsuloganglionic hemorrhage
at our center between January 2015 and December 2019. As per our institutional protocol,
patients who had two of the following features were taken up for DC: hematoma volume
greater than 60 mL, MLS > 10 mm, Glasgow Coma Score (GCS) < 8, and pupillary asymmetry.
Patients with lobar hemorrhage were excluded in this study.
Patient Information and Initial Management
Patient demographic data, that is, age, sex, comorbidities, GCS at admission, were
collected retrospectively from our prospectively maintained institutional ICH register.
All the patients underwent nonenhanced computed tomography (CT) of the brain on admission
and the images were analyzed after retrieving from our picture archiving and recovery
system. On admission, all the patients were intubated given the poor GCS. Blood pressure
reduction was performed in all cases with a target of < 140 mmHg as per INTERACT II.[11] Brain-specific antiedema measures were initiated in all patients. Hematoma volume
was calculated from the CT brain done before surgery using the formula ABC/2, where
A and B are the perpendicular maximal diameters of the lesion and C is the total length
in the vertical plane.
Surgery and Measurement of Craniectomy
Following the confirmation of the decision to operate—a reverse question mark—Dandy’s
flap was taken and the fronto-temporoparietal bone was exposed. Burr holes were placed
at the key, midline, temporal floor, and behind the parietal eminence and craniotomy
was performed using a high-speed cutter. The craniectomy flap was made to include
subtemporal decompression, medially till 1 to 2 cm from the midline and posteriorly
up to the posterior ear line. There was no standardization in the size of craniectomy.
Complete evacuation of hematoma was attempted in all cases. Lax duroplasty was performed
in all cases using a free pericranial patch. Six surgeons with various levels of experience
performed the cases. The bone flap was placed in the abdomen in a subcutaneous pouch.
Postoperative CT scans were performed routinely at 12 to 24 hours following surgery.
CT scanning was obtained in 3-mm-slice thickness without gap or overlap. Scans were
assessed for completion of hematoma removal, MLS, and size of DC.
The distance between the posterior and anterior margins of the bone defect was measured
on each slice using the inner table of the skull bone. AP diameter and height of the
craniectomy were measured in axial and coronal CT sections, respectively. The measurements
were confirmed after three-dimensional (3D) reconstruction of the images using InVasalius
version 3.1. The distance between the lower border of the craniectomy and the middle
cranial fossa base was measured at the level of the uncus. The distance of the margin
of the craniotomy from the midline in coronal sections was measured and the mean was
used in the analysis ([Fig. 1]). The MLS was assessed in the patients pre and postoperative scans at the level
of foramen of Monro in axial cuts.
Fig. 1 Representative computed tomography scan of a brain: (A) (a) diameter of the craniectomy; (b) longest perpendicular line to the dura. Margin of the area gained by the craniectomy
was hand-traced. Image B- (c) Distance from midline and base of craniectomy.
The volume gained by craniectomy—decompressive volume (DV)—was measured using the
Eq. ½ABC, which has a strong correlation with computed 3D volumetric assessment, which
remains the gold standard.[12] The CT slice showing the largest diameter in AP measurement was measured as A. A
perpendicular line (B) was measured from A to the dural flap. The number of CT slices
of the craniectomy, multiplied by the slice thickness, was the height (C) of DV ([Fig. 1]). Area that was gained by DC was calculated as suggested by Nagatani et al.[12]
Outcome
The outcome was measured using the Modified Rankin Scale (mRS) at discharge and 90
days. A favorable outcome was defined as mRS 0 to 3, and a poor outcome was defined
as mRS 4 to 6.
In all patients who survived, a CT brain was performed at 6 weeks post surgery to
assess hydrocephalus, subdural hygroma, and to plan replacement of bone flap. Hydrocephalus
was assessed using Evan’s index [24]. A value of > 0.3 was considered significant. The following variables were analyzed
for their influence on mRS, MLS, and hydrocephalus: AP length of the flap, height
of the flap, and distance from the midline.
Statistical Analysis
Descriptive statistics were reported for continuous variables. Continuous variables
were analyzed using independent Student’s t-test and one-way analysis of variance for normally distributed data while Welch–Satterthwaite
t-test was used for variables with unequal variances. Categorical variables were analyzed
with the chi-squared test. Univariate Spearman’s correlations were performed between
all the independent variables and mRS at follow-up. Mann–Whitney U test was used to analyze nonparametric data. A 5% α error and 80% β error along with
a corrected p-value of < 0.05 were considered as statistically significant. The analysis was performed
using IBM SPSS version 24 for Windows.
Results
A total of 63 patients underwent DC during the study period, and 55 patients were
included in the study group. The mean age was 50.73 ± 11.51 years with a male preponderance
(47:8). Mean systolic blood pressure on admission was 186.47 ± 29.53 mmHg and diastolic
was 104.69 ± 17.7 mmHg. Median GCS on presentation was 8 and preoperative GCS was
7 with the volume of hematoma of 49.39 ± 23.33 mL. Twenty-three patients had an intraventricular
extension of hematoma and 10 had hydrocephalus in the presenting scan. Mean preoperative
MLS was 1.05 ± 0.32, mean time to surgery was 18.47 hours, and hospital stay was 15
± 8.12 days ([Table 1]).
Table 1
Demographic and radiological data, and outcome of our cohort
|
Parameter
|
n = 55
|
|
Abbreviation: mRS, Modified Rankin Scale.
|
|
Age (y)
|
50.73 ± 11.51
|
|
Male (n)
|
47
|
|
Female (n)
|
8
|
|
Side (n):
Right
|
33
|
|
Left
|
22
|
|
Volume of hematoma
|
49.39 ± 23.33 mL
|
|
Preoperative midline shift
|
1.05 ± 28.262 cm
|
|
Postoperative midline shift
|
0.39 ± 0.27 cm
|
|
Anteroposterior diameter
|
12.42 ± 0.90 cm
|
|
Height
|
9.88 ± 0.81 cm
|
|
½ABC (cm3)
|
126.67 ± 34.78
|
|
Patients with craniectomy to base (n)
|
28
|
|
Complications (n)
|
10
|
|
mRS at 3 months (median)
|
4
|
|
Mortality at 3 months (n)
|
14
|
Mean AP diameter of the DC bone flap in our study was 12.42 cm (10.9–14.7 cm). The
AP diameter had no statistically significant impact on morbidity or mortality of the
patient (p = 0.4 and 0.8, respectively). The mean AP diameter of the patients who developed
postoperative hydrocephalus was higher than the patients who did not (12.8 and 12.1
cm, respectively; p = 0.036). No similar correlation could be made with the development of postoperative
hygroma (p = 0.49).
The mean height of the DC flap was 9.88 ± 0.8 cm. The patients who had a greater height
of the flap had a better outcome. Mean height of DC in the patients who had a good
outcome was: 10 vs. 9.2 cm (p = 0.021). Of the 55 patients, 28 had craniotomy flush with the temporal base and
27 patients did not. In these patients the residual temporal bone from the base ranged
from 4 to 20 mm. The patients who had a craniotomy up to the temporal base had a 3-month
mortality rate of 21% in comparison with 33% among the residual temporal bone group
(p = 0.227). The mean distance from the midline to the craniotomy was 2.3 cm in patients
who survived at 3 months compared with 2 cm in patients who did not. But no statistical
significance was noted for the observation on overall mortality or outcome.
The volume gained by decompression was estimated by the formula ½ ABC. The mean volume
was 126.67 ± 34.78 mm3. The mean volume of craniotomy was 123.87 ± 33.90 mm3 in patients who survived and 132.75 ± 37.30 mm3 in patients who died (p = 0.317). The volume of decompression did not influence mortality or outcome ([Tables 2] and [3]) in our cohort.
Table 2
Relationship between surgical parameters and mortality
|
Parameter
|
Alive (41)
|
Dead (14)
|
p-Value
|
|
Abbreviations: AP, anteroposterior; MLS, midline shift.
|
|
AP (cm)
|
12.47 ± 0.85
|
12.27 ± 1.05
|
0.529
|
|
Diameter (cm)
|
1.99 ± 0.46
|
2.25 ± 0.52
|
0.116
|
|
Height (cm)
|
9.93 ± 0.82
|
9.71 ± 0.80
|
0.378
|
|
ABC/2 (mm3)
|
123.87 ± 33.90
|
132.75 ± 37.30
|
0.346
|
|
Distance from midline (cm)
|
2.14 ± 0.49
|
2.08 ± 0.69
|
0.76
|
|
Postoperative MLS (cm)
|
0.34 ± 0.24
|
0.55 ± 0.27
|
0.011
|
Table 3
Relationship between surgical parameters and outcome
|
Parameter
|
mRS 0–3
|
mRS 4–6
|
p-Value
|
|
Abbreviations: AP, anteroposterior; MLS, midline shift; mRS, Modified Rankin Scale.
|
|
AP (cm)
|
12.47 ± 0.85
|
12.47 ± 0.86
|
0.981
|
|
Diameter (cm)
|
2.17 ± 0.45
|
1.93 ± 0.45
|
0.16
|
|
Height (cm)
|
9.35 ± 0.70
|
10.12 ± 0.77
|
0.009
|
|
ABC/2 (mm3)
|
127.10 ± 30.21
|
122.83 ± 35.42
|
0.714
|
|
Distance from midline (cm)
|
2.09 ± 0.50
|
2.16 ± 0.49
|
0.701
|
|
Postoperative MLS (cm)
|
0.23 ± 0.22
|
0.37 ± 0.24
|
0.075
|
Persisting postoperative MLS after hematoma evacuation and DC was assessed in the
two groups and the patients with good outcome had an MLS of 0.34 ± 0.24 cm in comparison
to 0.55 ± 0.27 cm in those with poor outcome (p = 0.001). Of the patients with persistent MLS of greater than 5 mm, 53% died in the
first 3 months (p = 0.001). However, patients with an AP diameter of < 12 cm (14/55) had a significantly
lesser reduction in MLS postoperatively (p-value:–0.037).
Overall mortality in the study group at 3 months was 14. Mean GCS and mRS on discharge
were 10.34 and 4.85, respectively. Mean mRS at 90 days was 4.85. Ten patients had
a good outcome and 45 had a poor outcome at 3 months.
Follow-up scans were performed for all the patients at 6 weeks. Twenty-five patients
had hydrocephalus (Evans index of > 0.3). Also, 68% patients with an AP diameter of
> 13 cm were observed to have hydrocephalus compared with 35.6% patients with AP diameter
of < 13 cm (p = 0.025; [Table 4]). Of them, eight patients who had improvement with lumbar cerebrospinal fluid (CSF)
drainage were taken up for permanent CSF diversion procedure. A subdural hygroma along
the ipsilateral convexities was observed in six of these patients. The volume of craniotomy
and distance from the midline did not correlate with hydrocephalus or hygroma (p > 0.05).
Table 4
Anteroposterior diameter and mortality, and postoperative outcome
|
Parameter
|
AP diameter
|
p-Value
|
|
>13 cm
|
<13 cm
|
|
Abbreviations: AP, anteroposterior; MLS, midline shift; mRS, Modified Rankin Scale.
|
|
Mortality
|
10 (25.6%)
|
4 (25%)
|
0.96
|
|
Outcome (mRS < 3)
|
17.9%
|
18.8%
|
0.94
|
|
Mean reduction in MLS
|
0.52 cm
|
0.7 cm
|
0.53
|
|
Hydrocephalus
|
35.9%
|
68%
|
0.025
|
Sixteen patients had tracheostomy due to prolonged ventilatory requirement. Of the
total patients, five developed postoperative meningitis, one patient had wound complication
with CSF leak, and four developed ventilator-assisted pneumonia. There was no correlation
between the size of craniectomy and complications.
Discussion
Surgery for ICH has been a much-debated topic in the past few decades, with multiple
studies comparing conservative and surgical methods. STICH 1 and 2 trials have concluded
that early surgery might have a small but clinically relevant survival advantage for
patients with spontaneous superficial ICH without intraventricular hemorrhage.[13] Some recent studies have opined that surgery improved survival in comparison with
medical treatment with lasting benefits.[14] Many centers are henceforth reverting to surgery for SICH, but the ideal type of
surgery for the best outcome has been the dilemma faced by neurosurgeons. Although
minimally invasive procedures are at the forefront of surgery for SICH, DC has been
found to improve mortality and morbidity outcome in several series.[14]
[15] A recent meta-analysis of seven high-quality studies to compare DC versus craniotomy
for spontaneous intracerebral bleeds has concluded that DC effectively reduced mortality
in patients with SICH.[14] The ideal size of the decompression required and the complications of DC in relation
to the volume, dimensions, need for craniectomy up to the middle cranial fossa base,
and the ideal distance from the midline have never been analyzed.
A study by Tanrikulu et al, with a cohort comprising of both TBI and SICH, concluded
that a craniectomy of 12 to 18 cm was effective as complete hemispheric exposure (i.e.,
AP diameter > 18 cm) in relieving intracranial hypertension. Also, a less extensive
approach does not increase the risk for secondary complications such as parenchymal
shear stress, hemorrhage, and swelling.[16] In a similar study with 526 patients undergoing DC with 21% SICH patients, the size
of the bone flap did not relate either to survival or outcome[7] disbanding the notion that larger the craniotomy better the outcome.
In our study, no direct correlation could be proved between AP diameter of the bone
flap and outcome or mortality. The persistence of MLS, which is an indicator of raised
ICP[17] in the postoperative scans after DC, and incomplete hematoma evacuation were observed
to have a significant impact on mortality and outcome. We observed that a diameter
of 12 to 13 cm is adequate to achieve the required reduction in MLS. Any further increase
in size is associated with complications.
Extension of DC up to the temporal base is known to improve outcome in patients with
TBI and large cerebral infarcts. Few authors believe this surgical step to be more
important than the size of the craniectomy flap.[18] In our study, patients who had DC up to the temporal base had better survival (mortality
21 and 33%, respectively; p = 0.157), suggesting that extension up to the temporal base is a valuable addition
in all cases even if the brain is lax post hematoma evacuation.
Volume Gained by Decompressive Craniectomy
The size of therapeutic DC is important in reducing refractory raised ICP.[19] Several methods have been described for the analysis of the volume of decompression,
with computer-assisted volumetric analysis as the gold standard.[12] The mean volume gained through DC in our study was 126.64 ± 34.78 cm, which was
concurrent with other previously done studies that had used 3D volumetric analysis.[3] It has been a common belief in clinical practice that “go big or go home” is useful
in DC, that is, larger the volume of decompression better the outcome.[16] This was not found to be true in our study as no significant inference could be
achieved to prove the benefit of a larger volume of craniectomy.
Size of Craniotomy and Its Influence on Hydrocephalus
De Bonis et al postulated that if the DC margin is too close to the midline, it reduces
the external force to the bridging vein and causes an increased venous outflow to
the sinus. The increased extracellular fluid absorption results in the decreased volume
of the brain parenchyma and induces ventricular enlargement. Craniectomy with a superior
limit too close to the midline can predispose patients to develop hydrocephalus. Hence
it is recommended to perform wide DCs with the superior limit of > 25 mm from the
midline.[20] This correlation between hydrocephalus and distance from midline could not be proved
in our study.
An interesting finding in our study was that patients with an AP diameter of > 13
cm had a significantly higher risk of developing hydrocephalus (p = 0.03). This was probably due to large DC playing a “flattening” role in the normally
dicrotic CSF pulse wave in patients because of transmission of a pressure pulse out
through the cranial defect. Arachnoid granulation function is dependent on the pressure
difference between the subarachnoid space and draining venous supply. So, it is possible
that disruption of pulsatile ICP dynamics secondary to opening the cranial defect
results in decreased CSF outflow and absorption, thereby leading to hydrocephalus.[21]
Limitations
As it was a retrospective study, no randomization of the size of the craniotomy was
performed preoperatively and it was left to the decision of the operating surgeon.
As the cases were operated by six different surgeons with various levels of expertise,
it could have played a role in the outcome.
Due to the small sample size, no specific recommendations could be made and further
analysis with larger sample size is required.
Conclusion
The common belief that larger the craniectomy the better the outcome has not proven
to be true in case of SICH. The recommended AP diameter of the DC is 12 to 13 cm to
achieve best outcome with minimal complications. Larger AP diameter resulted in a
higher incidence of hydrocephalus. Extension to the temporal base has better survival
outcomes. Patients with persistent MLS of > 5 mm in the immediate postoperative scan
after complete hematoma evacuation had a significantly poorer outcome and higher mortality.
