Keywords
Decompressive Craniectomy - Infarction, Middle Cerebral Artery - Intracranial Pressure
Palavras-chave
Craniectomia Descompressiva - Infarto de Artéria Cerebral Média - Pressão Intracraniana
INTRODUCTION
Up to 10% of all stroke patients suffer from complete middle cerebral artery (MCA)
infarction[1]; the leading cause of death is brain edema.[2] Swelling of the brain after ischemic insult has been described in numerous studies.[1]
[3]
[4] In an autopsy study, Shaw et al.[5] demonstrated that after an infarction in MCA territory, the midline shift (MLS)
reaches its peak at day four. Silver et al.[4] found that the peak risk for death secondary to transtentorial herniation after
a supratentorial infarction occurred 3 to 5 days (72–120 h) following the ischemic
event. More recently, according to computed tomography (CT) exams, Dohmen et al.[3] showed that in 9 patients with a malignant course of an MCA infarction, MLS peaked
after 54.7 h.
Midline shift has been used as a marker of monohemispheric brain edema in several
studies.[6]
[7] However, it does not measure brain edema with the required volumetric quantitative
precision. Instead, MLS more likely represents subfalcine herniation,[8] which occurs in advanced cases of monohemispheric brain edema. At the onset of brain
edema in the cerebral hemisphere, compensatory mechanisms, including a reduction in
hemispheric cerebrospinal fluid and blood volumes, occur before the establishment
of MLS.[9]
Because of such compensatory mechanisms, intracranial pressure (ICP) monitoring in
patients with malignant MCA infarction before decompressive craniectomy (DC) has limited
use since patients with severe brain stem compression, and even pupillary asymmetry,
can present normal ICP values.[10]
Therefore, in patients with malignant MCA infarction, cerebral edema's time course
determines the therapeutic option and their timing. In this way, comprehensive knowledge
regarding the time course of brain edema after DC could contribute to such patients'
management.
Regarding this matter, cerebral whole-hemisphere volumetry, and its variation according
to time from symptoms onset is an important factor to be considered. There are several
free non-commercial tools used in neuroscience in order to measure brain volumes,
including: FreeSurfer, volBrain and 3D-Slicer. Among these, 3D-Slicer is the unique
option that allows CT analysis (the others are based on magnetic resonance imaging
[MRI]).[11]
Studying patients with DC after malignant MCA infarction offers a unique opportunity
to observe the time course of brain swelling after large ischemic strokes in patients
that do not have a rigid covering of the brain. Therefore, our goal was to evaluate
the course of brain edema in patients with malignant MCA infarction treated with decompressive
craniectomy (DC) using hemispheric volumetric evaluations.
METHODS
Patients' selection
This retrospective study was approved by an ethics committee and was performed following
the declaration of Helsinki. During hospital admission, signed consent was obtained
from patients or relatives, and all data were anonymized at the source. We selected
consecutive patients from a single tertiary hospital between 2013 and 2019. All patients
had been diagnosed with a malignant MCA infarction and were submitted to DC to treat
the ischemic event. Patients with an intracerebral hematoma, either initially or after
hemorrhagic transformation, were excluded. Patients with prolonged time between symptoms
onset and DC (> 96 hours) were excluded as well, since this prolonged ischemic time
in a hemisphere cover by bone could prevent hemisphere to expand.
Diagnosis of malignant MCA infarction was made through a comprehensive analysis of
multiple factors, including infarction size, National Institutes of Health stroke
scale (NIHSS) score, radiological signs of mass effect, age, and past medical history.
Subsequently, a multidisciplinary team that included a neurologist, a neurosurgeon,
and a critical care physician decided whether to apply DC or not. Under this diagnosis
(malignant MCA infarction) we included patients with a large infarction in the vascular
territory of the MCA, including or not impairment of other vascular territories (anterior
cerebral artery or posterior cerebral artery).
Clinical and radiological data were collected from the electronic medical records
and included: age, sex, side of infarction, time from onset of symptoms until DC,
NIHSS score, pupillary status before DC, hemispheric volume time course variation,
infarction vascular territory, past medical history, alteplase use, endovascular procedures
before DC, and modified Rankin scale after 1 year.
Surgical procedure
Neurosurgeons from our hospital department performed all procedures. Two types of
skin flaps were made: a large, inverted question mark (Becker type) and the T-type
skin flap. As a department rule, the craniectomy flap was planned to be as big as
possible. After craniectomy, the dura was incised, and then an expansion duraplasty
using pericranium was made.
Image analysis
All CT exams performed during these patients' clinical care were analyzed and processed
in the software 3D Slicer (v. 4.10 - www.slicer.org), a free, open-source software,[12] including CT exams before and after DC. After loading the CT scan image on 3D Slicer,
we performed whole hemisphere segmentation and calculated the volume of the whole
infarcted hemisphere, excluding the ventricular system ([Figure 1]). To do this, we made the following steps: In Segment Editor Module, we started using the threshold of Hounsfield units between 7 and 70 as a mask, to
exclude the cerebrospinal fluid (CSF) in the intraventricular and subarachnoid spaces.
Then, the whole hemisphere of interest was marked using the automated tool Grow from Seeds in the segment editor module. After that, we inspected the quality of segmentation,
performing some manual corrections when necessary. Lastly, using the Segment Statistics Module, we obtained the volume of the segmented whole hemisphere.
Figure 1 The light green area represents the segmented right hemisphere after decompressive
craniectomy, in a postoperative computed tomography exam.
Statistical analysis
Categorical variables were described as numbers of cases and percentages, and quantitative
variables were characterized as mean ± standard deviations (SDs) or means ± range.
In order to demonstrate the average time between the ischemic event and the peak of
cerebral hemisphere volume, we used a Kaplan-Meyer survival curve, using the highest
measure as the event. If the volumetric variation was seen as a curve (with the highest
measure between two or more lower measures), we considered the event's occurrence.
The event was censored if the volumetric variation did not show as a curve.
All statistical analyses were performed using SPSS Statistics for Windows, version
17.0 (SPSS Inc., Chicago, IL, USA).
RESULTS
We identified 53 consecutive patients who underwent a DC after the diagnosis of a
malignant MCA infarction in our hospital from 2013 to 2019. Six patients were excluded
from the analysis, as we identified hemorrhagic transformation in the ischemic area,
and four patients were excluded because of time from symptom onset to DC (higher than
96 hours).
Therefore, we analyzed data from 43 patients (20 female) and a total of 197 corresponding
CT exams. The mean age at DC was 52.27 (SD, 11.18] years. The mean time between the
ischemic ictus and DC was 41.88 (SD, 29.32) hours ([Table 1]).
Table 1
General characteristics of the patients included in this study
|
Demographics
|
|
Men
|
23 (53.5%)
|
|
Mean age (years)
|
52.27 (SD, 11.18]
|
|
Mean time from symptom onset to DC (hours)
|
41.88 (SD, 29.32)
|
|
Neurological examination
|
|
NIHSS score[12] (admission)
|
18 (IQR 17-20)
|
|
Ipsilateral mydriasis before DC
|
12 (27.9%)
|
|
ASPECTS[13]
|
|
ASPECTS 0
|
12 (27.9%)
|
|
ASPECTS 1
|
7 (16.3%)
|
|
ASPECTS 2
|
14 (32.6%)
|
|
ASPECTS 3-6
|
10 (23.2%)
|
|
Stroke treatment used before DC
|
|
Intravenous alteplase
|
11 (25.6%)
|
|
Intraarterial thrombolysis or thrombectomy
|
7 (16.3%)
|
|
Modified Rankin scale after one year
|
|
mRS 3
|
16 (50%)
|
|
mRS 4
|
9 (15%)
|
|
mRS 5
|
2 (4.7%)
|
|
mRS 6 (death)
|
16 (37.2%)
|
|
Infarction site restricted to ipsilateral MCA territory
|
25 (58.1%)
|
|
Infarction site in MCA territory plus ACA or PCA
|
18 (41.9%)
|
Abbreviations: ACA, anterior cerebral artery; ASPECTS, Alberta stroke program early
CT score; DC, decompressive craniectomy; IQR, interquartile range; MCA, middle cerebral
artery; NIHSS, National Institutes of Health stroke scale; PCA, posterior cerebral
artery; SD, standard deviation.
The time course of the volumetric cerebral hemisphere variation in each patient is
shown in [Figure 2].
Figure 2 Each colored line represents the course of a single patient.
The mean time between the ischemic event and the peak hemisphere volume was 168.84
(95% confidence interval [CI] [142.08, 195.59]) hours. All patients achieved peak
ischemic hemisphere volume after DC.
DISCUSSION
In this study, we evaluated a group of consecutive patients who underwent a DC after
a malignant MCA infarction in our institution. The peak of hemispheric cerebral volume
in our patients occurred 168.84 (7.03 days) hours after the initial presentation of
malignant stroke symptoms. In all patients, peak hemispheric volume occurred after
DC.
Understanding the timing at which ischemic hemisphere volume peaks in these patients
is important in managing ICP after DC. Previous studies demonstrated that the ICP
could be elevated in patients who underwent DC for malignant cerebral infarction in
the postoperative period.[13]
[14]
Dohmen et. al.[3] analyzed a series of 17 patients with MCA infarction and hypoattenuation exceeding
50% of the MCA territory in the early (< 12 h) CT scan. In this series, MLS peaked
62.4 hours after stroke in patients who did not develop malignant edema and 54.7 hours
in patients who developed a malignant MCA syndrome. Only two patients underwent a
DC in this series.
In our study, all patients underwent DC, and after DC, the MLS became an erratic hemispheric
swelling indicator.[15] After DC, many factors could modify the relationship between MLS and hemispheric
cerebral swelling, including the area of bone decompression, the adequacy of duraplasty,
and the elasticity of the skin. These factors could explain the divergence between
the peak of MLS (in previous studies) and the peak of hemispheric cerebral swelling
that we identified in our study.
Previous studies have analyzed the volume of infarcted cerebral areas. Brott et al.[16] found a mean volume of infarcted brain tissue (visualized by CT exams performed
within the first 48 hours after stroke onset, 7 to 10 days after stroke onset, and
3 months after stroke onset) of 14 cm3 (first 48 hours), 55 cm3 (7-10 days), and 41 cm3 (3 months). In a series of patients with a malignant MCA infarction who underwent
DC, Freyschlag et al.[17] observed a mean volume of infarcted brain tissue of 250 cm3 (in the preoperative period), 315 cm3 (postoperative day one), and 349 cm3 (postoperative day three).
In patients with malignant hemispheric cerebral infarction, the whole hemisphere's
volume is the main factor that determines the intracranial pressure. Therefore, therapeutic
measures have to be based on the time course of the entire hemisphere volumetric variation.
Our series demonstrated that the peak volume in the ischemic hemisphere occurred later
than previously suspected. Secondary ischemic lesions that occur at the border of
the DC area secondary to hemisphere ischemic expansion traction in axonal bundles,
among other factors, could explain the delay of the peak hemispheric volume in patients
who undergo DC. Furthermore, delayed involvement of PCA or anterior cerebral artery
(ACA) territory could also contribute to explain our findings.
As previously expected, the peak ischemic hemisphere volume was after DC in all patients
of this series. Preoperative measurements were made to obtain a volumetric baseline
for each hemisphere.
Therefore, according to this study's data, after DC in patients with malignant MCA
infarction, the critical care team should be vigilant in detecting intracranial hypertension
even after the 5th day.
The present retrospective study has several limitations. First, it is based on a single
institution case series. Second, the hemispheric volumetric calculations were made
using a partial manual segmentation method, leading to measurement errors.
In conclusion, the peak of cerebral edema in malignant MCA infarction after DC occurred
on the 7th day (168.84 h) after stroke symptoms onset. Further studies evaluating therapies
for brain edema even after DC should be investigated.
