J Neurosurg 106:59–65, 2007

Decompressive hemicraniectomy in malignant middle

cerebral artery infarction: an analysis of long-term
outcome and factors in patient selection

ASHOK PILLAI, M.D.,1 SAJESH K. MENON, M.CH.,1 SATYENDRA KUMAR, M.CH.,2

KARIYATTIL RAJEEV, M.CH.,1 ANAND KUMAR, D.M.,1 AND DILIP PANIKAR, M.CH.1
Departments of 1Neurosurgery and 2Neurology, Amrita Institute of Medical Sciences, Amrita Vishwa
Vidyapeetham University, Kochi, Kerala, India

Object. Middle cerebral artery infarction often occurs at a younger age than other strokes and is associated with sig-
nificant rates of mortality and morbidity. After a period of pessimism regarding decompressive hemicraniectomy in the
management of acute stroke, the method has reemerged in the past decade. The present study was undertaken to assess
the immediate and long-term outcome of this intervention and to help better define the selection criteria for surgery.
Methods. The authors conducted a nonrandomized prospective study using decompressive hemicraniectomy with
duraplasty in patients at various stages of clinical deterioration due to a space-occupying middle cerebral artery infarct.
Patients were assessed at 6 and 12 months postinfarction by using functional scales. Subjective reconsideration was as-
sessed using a questionnaire.
Twenty-six patients were included in the study. The mean age was 48.4 6 11.2 years, and the mean preoperative
Glasgow Coma Scale score was 9.9 6 3.2. The median time from ictus to surgery was 54 hours (range 13–288 hours).
The rate of survival at 1 year postsurgery was 73%. Among survivors, 33.3% were independent (Barthel Index [BI] .
95) and 55.6% were partially dependent (BI 60–95) at 1 year postsurgery, with 72% attaining the ability to walk inde-
pendently by 1 year postsurgery. No patient remained in a vegetative state. The 1-year BI score was inversely related
to patient age (r = 20.47, p = 0.048).
Conclusions. Survival after decompressive hemicraniectomy was better than previously reported using medical man-
agement alone. A vegetative state was avoided and functional independence was possible, especially in younger pa-
tients. Increasing age was a statistically significant predictor of disability and long-term functional dependence.

KEY WORDS • stroke • hypertension • infarction • middle cerebral artery •

decompressive hemicraniectomy • mortality rate • outcome

IDDLE cerebral artery infarction often results in sig- ICP. The outcome in such cases with the best medical man-

M nificant rates of morbidity and mortality. Although

stroke is more common in the elderly, MCA terri-
tory infarction is not uncommon in younger age groups
because of embolic occlusion, and it accounts for nearly
10% of ischemic strokes.8 Approximately 10 to 20% of
these infarctions are massive and cause severe brain edema
resulting in uncal herniation and death.13 These pathological
entities have been referred to as “malignant MCA infarcts.”8
Medical therapy in such cases is limited to osmotic agents
to reduce edema and mechanical ventilation to control the
Abbreviations used in this paper: BI = Barthel Index; CT = com-
puted tomography; FIM = Functional Independence Measure;
GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale;
ICP = intracranial pressure; ICU = intensive care unit; MCA = mid-
dle cerebral artery; NIHSS = National Institutes of Health Stroke
Scale; QOL = quality of life; SD = standard deviation.
agement alone is generally poor—only 20 to 40% survival
at best and a high degree of functional dependence in the
survivors.8,13
Decompressive hemicraniectomy performed for MCA
infarction has been reported to immediately effect a dra-
matic reduction in ICP to normal ranges,21 preventing fatal
uncal herniation and generally leading to a more rapid neu-
rological recovery.2,12,14,16 It helps shorten the ICU stay, thus
reducing medical complications.18 There is adequate evi-
dence to indicate that decompressive hemicraniectomy has
some role in the treatment of select patients with malignant
MCA infarction.1,3,4,7,10–12,14,15,18,20 Several large randomized
controlled trials are ongoing. Questions that remain unan-
swered include the following: 1) which subset of patients
will benefit maximally?; 2) which patients will survive with
an unacceptable degree of functional dependency?; and 3)
what is the optimal timing of surgery?

J. Neurosurg. / Volume 106 / January, 2007 59


Data from several previous studies have already demon-
strated improved outcomes in surgically treated patients
compared with medically treated controls.2,12,16,18 The objec-
tives of the present prospective nonrandomized study were
as follows: 1) to help better define the selection criteria for
performing decompressive hemicraniectomy in cases of
malignant MCA infarcts; 2) to assess the immediate out-
come in terms of time to conscious recovery and survival;
and 3) to assess long-term outcome using standard QOL
and functional assessment scales.
Clinical Material and Methods
Selection Process and Inclusion Criteria
Patients presenting with acute MCA infarction to the
Amrita Institute of Medical Science, a tertiary care univer-
sity teaching hospital, during the period between August
2001 and September 2004 were considered for enrollment
in this study. An institutional protocol was formed with
inclusion criteria (Table 1). Informed consent was obtained
after detailed discussion with the relatives regarding the
option of surgical decompression to improve the chance of
survival given the expected long-term neurological seque-
lae. Patients with both dominant- and nondominant-hemi-
sphere infarcts satisfying the CT criteria (Table 1) but not
yet showing clinical signs of deterioration were admitted to
the stroke unit and underwent surgery only if the clinical
criteria were met. If the initial CT scan failed to meet the se-
lection criteria, the scan was repeated within 12 to 24 hours
or as soon as significant neurological worsening occurred
whenever there was clinical suspicion of a large cortical
infarct.
Medical Management
All patients were admitted to the acute stroke unit and
treated with osmotic therapy (20% mannitol 0.5-g/kg bolus
followed by 0.25–0.5 g/kg every 4–6 hrs, furosemide 10–20
mg every 4–6 hrs). Patients meeting the inclusion criteria
despite maximum medical therapy underwent surgery with-
in 4 to 6 hours. Patients presenting with a poor GCS score
(, 9) or pupil asymmetry immediately underwent intuba-
tion and received ventilation; all other patients received
ventilation just prior to surgery.
TABLE 1
Selection criteria for study inclusion
Variable Description
age #65 yrs
CT findings large hypodensity in .50% of the MCA territory
w/ significant effacement of sulci & ventricles
midline shift
GCS score #14 (nondominant hemisphere) or #9 (domi-
nant hemisphere)
herniation/brainstem no signs
reflexes
operative risk acceptable regarding other major comorbidities
(for example, coronary artery disease, renal
failure, and so forth)
60
A. Pillai et al.
Surgical Procedure
Decompressive hemicraniectomy was performed using
a large frontoparietotemporal curvilinear incision. A large
hemicraniectomy bone flap, including the frontal, parietal,
and temporal squamous bone, was removed. The temporal
squama was removed to the middle cranial fossa floor to re-
duce the chance of subsequent uncal herniation. A curvilin-
ear dural incision was used with radial cuts following major
sulcal vessels to prevent kinking at the dural margin by ede-
matous brain. No brain parenchyma was resected. A lax du-
raplasty was performed using pericranium and temporalis
fascia.
Postoperative Management
Postoperatively, all patients were sedated and paralyzed
and received ventilation for a minimum of 48 hours. After
obtaining a CT scan at 48 hours postsurgery, sedation and
paralysis were withdrawn and ventilation was stopped as
soon as patients were conscious with spontaneous eye open-
ing and a localizing motor score. Tracheostomy was per-
formed to facilitate ventilator weaning if the endotracheal
tube had not been removed by the 7th postoperative day.
Barbiturates were not administered.
Bone Flap Storage and Reimplantation
Bone flaps were stored at –70˚C until reimplantation. At
12 to 20 weeks postsurgery, the bone flap was reimplant-
ed after having been disinfected by boiling in sterile saline
(first six patients) or autoclaving (last 20 patients).17
Outcome Analyses and Long-Term Follow Up
All patients were assessed just before surgery using the
GCS, NIHSS, and standard neurological examination. The
immediate outcome measures included the number of days
to conscious recovery (assessed by spontaneous eye-open-
ing and localizing motor score), the number of days of ven-
tilation, and the duration of the ICU stay. In all except two
of the surviving patients (one lost to follow up and one fol-
lowed up telephonically), long-term follow up was main-
tained through regular outpatient clinic visits. At each visit
the patients were assessed using the NIHSS, BI, GOS, and
the FIM walking score. With regard to assessment with the
GOS scale, a score of 4 (moderate disability) meant that the
patient was independently mobile at home and performing
activities beyond the activities of daily living but had mod-
erate motor aphasia precluding involvement in all previous
activities and/or severe hemiparesis preventing functional
use of the paretic arm. To further assess the QOL, a subjec-
tive retrospective reconsideration questionnaire was sent to
all survivors. On this questionnaire, the patient (if possible)
and the relative involved in the most caregiving (general-
ly a spouse, parent, or child of the patient) were asked the
question, “If you were faced with a similar situation in the
future for yourself or someone close to you, would you
again make the same decision?” The answer was recorded
using a five-point scale (1 = definitely no, 5 = definitely
yes).
Statistical Analysis
Statistical analyses were performed using SPSS software
(standard version 11.01; SPSS, Inc.). Continuous variables
were reported as the means 6 SD. Categorical variables
J. Neurosurg. / Volume 106 / January, 2007
Decompressive hemicraniectomy in MCA infarct

were recorded using numbers and percentages. A compar- TABLE 2

ison of the mean values was performed using the Student Characteristics of 26 patients with MCA infarcts*
t-test. Categorical variables were compared using the Fisher
exact and chi-square tests. Bivariate correlations were per- Variable Value (%)
formed using the Pearson and Spearman rank correlations. sex (no. of patients)
A probability value of 0.05 was considered significant. All M 22 (84.6)
tests were two-tailed. The logistic regression technique was F 4 (15.4)
used for multivariate analysis. Variables that were signifi- mean age in yrs 48.4 6 11.2
cant or had a probability value less than 0.25 on univariate range 28–66
infarct characteristics (no. of patients)
analysis were subjected to multivariate analysis. hemispheric location
dominant 12 (46.2)
nondominant 14 (53.8)
Results MCA 21 (80.8 )
MCA + ACA or MCA + PCA 4 (15.4)
Study Population MCA + ACA + PCA 1 (3.8)
stroke severity before surgery
The characteristics of the study population, which con- mean GCS score 9.9 6 3.2
sisted of 26 patients (22 men and 4 women), are presented dominant hemisphere 7.7 6 1.8
nondominant hemisphere 11.8 6 3.0
in Table 2. The mean patient age was 48.4 6 11.2 years mean NIHSS score 17.7 6 4.1
(range 28–66 years) with a normal distribution. The mean dominant hemisphere 20.6 6 3.1
GCS score was 9.9 6 3.2, and the mean NIHSS score was nondominant hemisphere 15.4 6 3.2
17.7 6 4.1. Nine patients (35%) had signs of uncal hernia- unilateral pupil dilation (no. of patients) 9 (35)
tion as detected by pupil asymmetry. The median time from time from ictus to surgery (hrs)
ictus to surgery was 54 hours (range 13–288 hours). Pre- median 6 SD 54.0 6 59.4
range 13–288
existing cardiac illness was subdivided as follows: cardiac cause of infarct (no. of patients)
arrhythmia, coronary heart disease with or without a major cardioembolic 10 (38.5)
coronary ischemic event, and valvular heart disease. The large-vessel atherosclerosis 5 (19.2)
cause of the infarct was classified as cardioembolic if the ICA dissection 1 (3.8)
patient had a known arrhythmia or heart valve lesion pre- unknown 10 (38.5)
disposing to embolic infarction or if there was echocardio- CT findings
mean midline shift (mm) 9.8 6 5.7
graphic evidence of a left ventricular or atrial clot. hemorrhagic change (no. of patients) 7 (27)
medical comorbidities (no. of patients)
Immediate Postoperative Outcome diabetes 10 (38.5)
hypertension 10 (38.5)
The mortality rate in the 1st postoperative month was cardiac abnormalities
28% (seven of 25 patients, one was lost to follow up). There atherosclerotic coronary disease 10 (38.5)
were no deaths after this period. The median ICU stay valvular heart disease 4 (15.4)
was 5 6 4.1 days, and the mean period of mechanical ven- cardiac rhythm abnormality 4 (15.4)
tilation was 4.5 6 1.9 days. Four patients (15%) required COPD 1 (3.8)
renal insufficiency 1 (3.8)
temporary tracheostomy to facilitate weaning of ventilation. cause of death (no. of patients/total)
transtentorial herniation 3/7 (43)
Long-Term Outcome Measures cardiac/unknown (sudden arrest) 4/7 (57)
Among the long-term survivors (19 patients), 17 (89%) * Mean values include SDs. Abbreviations: ACA = anterior cerebral ar-
were evaluated at 6 months and 18 (95%) were reviewed tery; COPD = chronic obstructive pulmonary disease; ICA = internal ca-
rotid artery; PCA = posterior cerebral artery.
at 1 year (Table 3). At 6 months postsurgery, 23.5% were
functionally independent in activities of daily living (BI .
95), 64.7% were partially dependent (BI 60–95), and 52.9%
were walking independently (FIM Scores 6 and 7). At 6 Six of the 12 patients with infarcts in the speech-domi-
months postsurgery, 11.7% of patients were functionally de- nant hemisphere survived; among these, five (83%) had
pendent. No patient was in a vegetative state. At 1 year post- moderate to severe motor aphasia at 6 and 12 months and
treatment, 92.9% of survivors had mild deficits (NIHSS one had fluent speech with only deficits in naming and rep-
Score 0–7), 7.14% had moderate deficits (NIHSS Score etition. Among the patients with severe motor aphasia, two
8–14), and no patients had severe deficits (NIHSS Score $ were able to communicate in a limited form in writing. The
15). According to the GOS at 1 year postsurgery, 60% of incidence of major depressive symptoms was higher in
survivors had a good outcome (GOS Scores 4 and 5) and patients with infarcts in the speech-dominant hemisphere,
12% were severely disabled (GOS Score 3). Thirteen pa- with two patients requiring repeated admission and treat-
tients (72%) were independently ambulant, three patients ment for major depression.
(17%) required minimal assistance in walking (FIM Walk-
ing Score 4–5), and two patients (11%) remained immobile Statistical Correlations
at 1 year postsurgery. Among the functionally independent,
four patients (22%) were eventually able to resume their There was no statistically significant association between
previous employment or to participate in previous house- the time from infarct to surgery and the outcome measures
hold duties in a limited way. (GOS, NIHSS, and BI) at 6 months or 1 year postsurgery on

J. Neurosurg. / Volume 106 / January, 2007 61

 A. Pillai et al.

TABLE 3
Summary of long-term outcome characteristics
No. (%)

Variable 6 Mos 1 Yr

no. of survivors available for follow up 17 18

GOS
no. of patients 25 25
score
1 7 (28) 7 (28)
2 0 0
3 2 (8) 3 (12)
4 12 (48) 10 (40)
5 4 (16) 5 (20)
BI
no. of patients 17 18
mean 6 SD score 80.0 6 21.9 80.6 6 24.4
no. independent (BI $95) 4 (23.5) 6 (33.3)
no. partially dependent (BI 60–95) 11 (64.7) 10 (55.6)
no. dependent (BI ,60) 2 (11.7) 2 (11.1)
no. w/ favorable outcome 15 (88.2) 16 (88.9)
NIHSS
total no. of patients 16 14
median score 6.0 6 3.5 5.0 6 2.6
no. w/ mild deficit (Score 1–7) 11 (68.8) 13 (92.9)
no. w/ moderate deficit (Score 8–14) 5 (31.3) 1 (7.14)
no. w/ severe deficit (Score $15) 0 0
FIM
no. of patients 17 18
mean walking score 5.5 6 1.74 6.1 6 1.70
no. independently walking 9 (52.9) 13 (72.2)
(FIM walking Scores 6 & 7)

either bivariate analysis (Spearman correlation) or a com-

parison of mean outcome values when divided into cate-
gories based on early (# 48 hours from ictus) compared
with late surgery (. 48 hours from ictus).
A statistically significant negative correlation (r = 20.47)
existed between patient age and the BI (p = 0.048, Pearson
test) at 1 year postsurgery (Fig. 1A), although this correla-
tion was not present at 6 months posttreatment (p = 0.071).
There was a similar negative correlation between patient
age and the FIM walking score at 1 year posttreatment
(r = 20.54, p = 0.020, Pearson coefficient). Regarding ra-
diological criteria, outcome scores were the same regardless
of whether hemorrhagic transformation was present.
Among the seven deaths, three were due to transtentorial
herniation (one patient with all three vascular territories in-
volved) and four were sudden deaths of unknown cause. Fig. 1. Scatterplots showing statistical correlations. A: Func-
Death was not related to patient age when comparing the tional score (BI) compared with patient age (r = 20.47, p = 0.048,
mean age of survivors with that of the dead (p = 0.31). Pearson test). B: A similar negative correlation between walking
score and patient age at 1 year postsurgery (r = 20.54, p = 0.020,
There was also no association with pupil asymmetry (p = Pearson test). C: A downward trend in the subjective reconsider-
0.20). The variables subjected to univariate analysis were ation score over time (r = 20.61, p = 0.028, Spearman correlation).
age, all medical comorbidities, preoperative stroke severity
based on the GCS and NIHSS, time from infarct to surgery,
and pupil asymmetry. Only the laterality of the stroke and
the presence of preexisting hypertension were significant tivariate analysis, however, only hypertension was found as
predictors of death on univariate analysis. Six of 13 patients a statistically significant predictor of death (p = 0.01, confi-
with dominant-hemisphere infarcts died compared with one dence interval 2.56–348.52), and stroke laterality was no
of 14 with nondominant-hemisphere infarct (p = 0.03, Fish- longer significant (p = 0.08). Note, however, that the study
er exact test). Among the 18 survivors, only three had pre- was limited in this respect given its small sample size.
existing hypertension, whereas six of seven deaths occurred Subjective Reconsideration
in known hypertensive patients. On univariate analysis,
preexisting hypertension emerged as a predictor of a high Retrospective reconsideration data were available for 14
mortality rate, with an odds ratio of 30 (p = 0.003). On mul- (74%) of the 19 survivors. The mean score was 4.4 6 1.2

62 J. Neurosurg. / Volume 106 / January, 2007

Decompressive hemicraniectomy in MCA infarct
(4 = probably yes, 5 = definitely yes). When an outlying
point (that from a 65-year-old patient who died after 21
months) was removed, we observed a significant downward
trend (r = 20.61, p = 0.028, Spearman correlation) in the re-
consideration score over time (Fig. 1B).
Surgical Complications and Other Long-Term Sequelae
Two patients suffered postoperative extradural hemato-
mas requiring reexploration and evacuation. Three patients
(one receiving anticoagulation therapy for atrial fibrillation)
had hematomas after reimplantation of the bone flap—one
thin chronic subdural hematoma and two extradural hema-
tomas, both of which resolved with conservative manage-
ment. There were three cases of delayed osteomyelitis of
the reimplanted bone flap. One of these cases subsided with
antibiotic treatment and two involved extradural abscesses
requiring drainage and delayed acrylic cranioplasty. There
were no cases of meningitis. The incidence of delayed
(poststroke) generalized seizures by the 6th to 12th month
postsurgery (seven [37%] of 19 patients) was significantly
greater than the reported incidence for poststroke epilepsy
(4.1–13%).19 The control of seizures often required anticon-
vulsion polytherapy. Two patients (8%) suffered major os-
teoporotic fractures of the paretic lower limb requiring sur-
gical internal fixation at 2 to 2.5 years postinfarction.
Discussion
Selection Criteria
As decompressive hemicraniectomy emerges as a viable
treatment option for a select group of patients with space-
occupying MCA infarction, an institutional protocol involv-
ing the stroke neurologist, intensivist, and neurosurgeon is
required for timely identification of patients who will bene-
fit from this intervention. Several other investigators report-
ing poor results seemed either to utilize decompressive sur-
gery as a salvage therapy in patients at various stages of
herniation or operate on an older population. A comparison
of initial study data published by Rieke et al.16 with that by
Schwab et al.18 supports the notion that outcome is worse
when surgery is performed during the stages of herniation.
Authors of several previously reported studies have shown
an improved outcome when surgery is undertaken earlier
in the course of neurological deterioration.2,15,18 A higher
chance of a vegetative outcome when surgery is performed
in the late stages of herniation has been reported.15 Our
study data failed to demonstrate a direct correlation between
clinical outcome and timing of surgery from the acute onset
of symptoms due to ischemia. We believe that the stage of
clinical deterioration rather than the time since infarction is
probably more important in determining outcome. Based on
TABLE 4
Proposed selection of patients for decompressive hemicraniectomy
Candidacy
good, likely to benefit
intermediate, likely to benefit, but significant chance of functional dependence
poor, high chance of death & poor functional outcome*
* May also include hypertension, but this factor not confirmed by other data.
J. Neurosurg. / Volume 106 / January, 2007
our data and that from the available literature, we propose
that younger patients with infarcts in the nondominant
hemisphere are likely to benefit significantly and thus
should undergo surgery. Although patients with dominant-
hemisphere infarcts are likely to survive with more dis-
abling deficits, surgery can be undertaken with the hope of
reducing the hospital stay and rates of morbidity and mor-
tality. We are reluctant to operate on elderly patients (. 65
years age) given that they tend to recover poorly (Table 4).
Data from animal studies certainly support the proposal
for earlier surgery. A reversal of the diffusion/perfusion
mismatch on magnetic resonance imaging studies was dem-
onstrated in rats undergoing experimental MCA occlusion
followed by early decompressive hemicraniectomy, thus in-
dicating salvage of ischemic penumbra.5,6 A statistically sig-
nificant, time-to-surgery–dependent improved functional
outcome and a decrease in the infarct volume of killed ani-
mals were also demonstrated. It was proposed that decom-
pressive hemicraniectomy helps in the reestablishment of
leptomeningeal collateral vessels, which are lost when com-
pressed by edematous brain, ultimately leading to a reduced
infarct volume. Severe brain edema causes a regional in-
crease in ICP, further reducing the regional cerebral perfu-
sion pressure and cerebral blood flow, which may potenti-
ate further infarction and thus create a vicious cycle (Fig. 2).
Decompressive hemicraniectomy probably helps to break
this cycle.
Radiological Selection Criteria
Computed tomography inclusion criteria described to
date have generally involved MCA infarction greater than
50% of the MCA territory together with findings of mass
effect. Specifically, we used midline shift in this study.
Several authors have included hemorrhagic transformation
among the exclusion criteria. In the present study, we did
not observe any difference in outcome in this subgroup of
patients. Authors of one large-scale retrospective study de-
scribed volumetric analysis to predict malignant behavior,
that is, an infarct volume of 200 cm3 having a 91% accura-
cy for predicting malignant behavior.12 As there is a trend
toward earlier surgery, there may be a more important role
for diffusion-weighted magnetic resonance imaging as de-
scribed in several recent reports.2
Technical Considerations in Decompressive
Hemicraniectomy
Authors of other case series have reported on the removal
of edematous temporal lobe to achieve further reduction in
ICP and prevent herniation.2 With the possible exception of
hemispheric infarcts involving all three arterial territories,
we found that this step was generally not necessary. A gen-
Age (yrs) Infarct Location
,55–60 nondominant hemisphere
55–65 dominant hemisphere
.65 multiple arterial territories
63

FIG. 2. Schematic demonstrating the possible mechanism of de-
compressive craniectomy in breaking the cycle of malignant infarc-
tion. CPP = cerebral perfusion pressure; MAP = mean arterial pres-
sure.
erous temporal osseous decompression should probably al-
low for the same effect.
Comparison With Other Studies
The functional outcome in our patient population was
generally better than has been previously reported. The
mean BI of 45 to 65 and GOS score of 3 to 3.9 at 6 months
postsurgery have been reported by others whose studies in-
cluded patients with a mean age ranging from 48.8 to 65
years.2,12,16,18,20 Given that increasing age was a statistically
significant predictor of a worse outcome in the present
series of patients and other similar studies,7,20 the improved
outcome in the present study may reflect the younger age of
the patients (mean age 48 6 11.2 compared with mean ages
in the 50- to 65-year range in other reported series). The
average time to surgery from the onset of ischemia was sim-
ilar to that in other previously published study data. A sim-
ilar statistically significant, inverse relation between age
and functional outcome score has been demonstrated by
other groups.7,20 Recovery from poststroke aphasia has also
been shown to be age-dependent.9 Regarding factors con-
tributing to higher mortality rates, preexisting hypertension
appeared to be a significant risk factor for death in our
series—a finding not previously reported to our knowledge.
In our experience, outcome scores tended to improve sig-
nificantly on long-term follow up (that is, at 6 months or 1
year postsurgery). Among the several prospective and retro-
spective studies reported on to date, some authors have
described poor functional outcomes based on 3-month out-
come analyses. Data from such studies will be less infor-
mative in assessing the efficacy of this intervention.
Quality of Life Issues
The unresolved controversy of whether to perform de-
compressive hemicraniectomy in patients with malignant
MCA infarction centers on the issue of QOL among the sur-
vivors. Data from this study and several others have demon-
strated that an acceptable functional outcome following de-
compressive hemicraniectomy is possible. Specifically,
there were no survivors in a vegetative state in our series of
patients. Other reported series have also demonstrated that
a vegetative outcome generally occurs much less with sur-
gery.20 This issue raises an important concern: based on
64
A. Pillai et al.
existing data of conservatively managed malignant MCA
infarction,8 if surgery is avoided while accepting a higher
mortality rate, the survivors (~ 20%) will be subjected to an
increased chance of a vegetative outcome and a probably
worse functional outcome.
Note that it was eventually possible to achieve a near-
normal QOL in 22% of the survivors at 6 months postsur-
gery. The majority of patients with dominant-hemisphere
infarcts or their caretakers had given favorable retrospective
reconsideration scores. However, the psychological impact
and practical difficulties created by major deficits such as
motor aphasia and limb paresis preventing employment and
resumption of previous activities were significant. The
downward trend in the retrospective reconsideration score
over time could indicate that these long-term difficulties
make patients and their caretakers eventually doubt whether
their choice of a life-saving intervention was for the best. In
another recent prospective study on decompressive hemi-
craniectomy for ischemic stroke, the use of accepted sub-
jective measurement scales of QOL showed that the true
burden of the disease was much greater than that reflected
in the objective functional outcome scores (BI, GOS, and so
forth).7 This outcome probably reflects the often overlooked
aspects of cognitive impairment and the psychological im-
pact resulting from loss of fine motor skills and language
function.
Conclusions
Our data agree with those from other similar case series
demonstrating that survival after decompressive hemicrani-
ectomy for MCA infarction is better than that reported after
medical management alone. A vegetative state is avoided,
and there is a significant chance of reduced functional de-
pendence, especially in the younger population. Increasing
age was a statistically significant predictor of disability and
long-term functional dependence. Centers at which this pro-
cedure is performed will need to formulate an institutional
protocol based on clinical and radiological selection criteria
for optimizing the outcome.
The fact remains that malignant MCA infarction implies
some amount of long-term disability despite the best man-
agement. Perhaps new developments in restorative therapy
can be combined in the future to reduce the burden of this
disabling condition. We hope that our findings will add to
existing information on decompressive hemicraniectomy to
serve as guidelines until further data are available from the
ongoing randomized control trials.
Acknowledgments
We thank Ms. S. Sumitra for her sincere work in the statistical an-
alysis of the data. We also acknowledge the Mata Amritanandamayi
Math for the charitable subsidy received by several patients enrolled
in this study. We sincerely thank all staff and faculty members of our
operative, anesthesia/critical care, and rehabilitation teams who con-
tributed significantly to the treatment of patients enrolled in this
study.
References
1. Carter BS, Ogilvy CS, Candia GJ, Rosas HD, Buonanno F: One-
year outcome after decompressive surgery for massive nondomi-
nant hemispheric infarction. Neurosurgery 40:1168–1176, 1997
J. Neurosurg. / Volume 106 / January, 2007
Decompressive hemicraniectomy in MCA infarct
2. Cho DY, Chen TC, Lee HC: Ultra-early decompressive craniecto-
my for malignant middle cerebral artery infarction. Surg Neurol
60:227–233, 2003
3. Csokay A, Egyud L, Nagy L, Pataki G: Vascular tunnel creation
to improve the efficacy of decompressive craniotomy in post-trau-
matic cerebral edema and ischemic stroke. Surg Neurol 57:
126–129, 2002
4. Delashaw JB, Broaddus WC, Kassell NF, Haley EC, Pendleton
GA, Vollmer DG, et al: Treatment of right hemispheric cerebral
infarction by hemicraniectomy. Stroke 21:874–881, 1990
5. Doerfler A, Engelhorn T, Heiland S, Benner T, Forsting M: Per-
fusion- and diffusion-weighted magnetic resonance imaging for
monitoring decompressive craniectomy in animals with experi-
mental hemispheric stroke. J Neurosurg 96:933–940, 2002
6. Engelhorn T, von Kummer R, Reith W, Forsting M, Doerfler A:
What is effective in malignant middle cerebral artery infarction:
reperfusion, craniectomy, or both? An experimental study in rats.
Stroke 33:617–622, 2002
7. Foerch C, Lang JM, Krause J, Raabe A, Sitzer M, Seifert V, et al:
Functional impairment, disability, and quality of life outcome
after decompressive hemicraniectomy in malignant middle cere-
bral artery infarction. J Neurosurg 101:248–254, 2004
8. Hacke W, Schwab S, Horn M, Spranger M, De Georgia M, von
Kummer R: Malignant middle cerebral artery territory infarction:
clinical course and prognostic signs. Arch Neurol 53:309–315,
1996
9. Kastrau F, Wolter M, Huber W, Block F: Recovery from aphasia
after hemicraniectomy for infarction of the speech-dominant he-
misphere. Stroke 36:825–829, 2005
10. Koh MS, Goh KY, Tung MY, Chan C: Is decompressive craniec-
tomy for acute cerebral infarction of any benefit? Surg Neurol
53:225–230, 2000
11. Kondziolka D, Fazl M: Functional recovery after decompressive
craniectomy for cerebral infarction. Neurosurgery 23:143–147,
1988
12. Mori K, Nakao Y, Yamamoto T, Maeda M: Early external decom-
pressive craniectomy with duroplasty improves functional recov-
ery in patients with massive hemispheric embolic infarction: tim-
J. Neurosurg. / Volume 106 / January, 2007
ing and indication of decompressive surgery for malignant cere-
bral infarction. Surg Neurol 62:420–430, 2004
13. Moulin DE, Lo R, Chiang J, Barnett HJM: Prognosis in middle
cerebral artery occlusion. Stroke 16:282–284, 1985
14. Pranesh MB, Dinesh Nayak S, Mathew V, Prakash B, Natarajan
M, Rajmohan V, et al: Hemicraniectomy for large middle cerebral
artery territory infarction: outcome in 19 patients. J Neurol Neu-
rosurg Psychiatry 74:800–802, 2003
15. Reddy AK, Saradhi V, Panigrahi M, Rao TN, Tripathi P, Meena
AK: Decompressive craniectomy for stroke: indications and re-
sults. Neurol India 50 (Suppl 1):S66–S69, 2002
16. Rieke K, Schwab S, Krieger D, von Kummer R, Aschoff A,
Schuchardt V, et al: Decompressive surgery in space-occupying
hemispheric infarction: results of an open, prospective trial. Crit
Care Med 23:1576–1587, 1995
17. Schultke E, Hampl JA, Jatzwauk L, Krex D, Schackert G: An easy
and safe method to store and disinfect explanted skull bone. Acta
Neurochir (Wien) 141:525–528, 1999
18. Schwab S, Steiner T, Aschoff A, Schwarz S, Steiner H, Jansen O,
et al: Early hemicraniectomy in patients with complete middle ce-
rebral artery infarction. Stroke 29:1888–1893, 1998
19. Sitajayalakshmi S, Mani J, Borgohain R, Mohandas S: Post stroke
epilepsy. Neurol India 50 (1 Suppl):S78–S84, 2002
20. Uhl E, Kreth FW, Elias B, Goldammer A, Hempelmann RG,
Liefner M, et al: Outcome and prognostic factors of hemicraniec-
tomy for space occupying cerebral infarction. J Neurol Neuro-
surg Psychiatry 75:270–274, 2004
21. Yoo DS, Kim DS, Cho KS, Huh PW, Park CK, Kang JK: Ventric-
ular pressure monitoring during bilateral decompression with du-
ral expansion. J Neurosurg 91:953–959, 1999
Manuscript received November 4, 2005.
Address reprint requests to: Dilip Panikar, M.Ch., Department of
Neurosurgery, Amrita Institute of Medical Sciences, Amrita Vishwa
Vidyapeetham University, Amrita Lane, Elamakkara P.O., Kochi,
Kerala, India 682026. email: panikar@aims.amrita.edu.
65