Surgical decompression for space-occupying
hemispheric infarction
Jeannette Hofmeijer
ISBN:
Cover:
Lay out:
Printed by:
978-90-393-4487-3
Roy Sanders (based on Wieslaw Rosocha: Shakespeare’s HAMLET,
Polish theatre, 1996)
Roy Sanders
Gildeprint drukkerijen BV
Surgical decompression for space-occupying
hemispheric infarction
Chirurgische decompressie voor het ruimte-innemende
supra-tentoriële herseninfarct
(met een samenvatting in het Nederlands)
Proefschrift
ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de
rector magniﬁcus, prof.dr. W.H. Gispen, ingevolge het besluit van het college voor
promoties in het openbaar te verdedigen op donderdag 19 april 2007 des middags te
12.45 uur
door
Jeannette Hofmeijer
geboren op 20 december 1971
te Ermelo
Promotor:
Prof. dr. L.J. Kappelle
Co-promotor:
Dr. H.B. van der Worp
Research described in this thesis was supported by a grant of the Netherlands
Heart Foundation (grant number 2002B138). Financial support by the
Netherlands Heart Foundation for the publication of this thesis is gratefully
acknowledged. Financial support for the publication of this thesis by Boehringer
Ingelheim is gratefully acknowledged.
Everything is what it is, because it got that way.
Alles is wat het is, omdat het zo geworden is.
(‘d Arcey W. Thompson. On growth and form. 1917.)
Aan mijn vader
Contents
Chapter 1
General introduction
9
Part I. Experimental studies
Chapter 2
Treatment of space-occupying middle cerebral artery
infarction
17
Chapter 3
The time course of ischemic damage and perfusion in
a rat model of space-occupying cerebral infarction
41
Chapter 4
Perfusion MRI by ﬂow-sensitive alternating inversion
recovery and dynamic susceptibility contrast in rats
with permanent middle cerebral artery occlusion
57
Chapter 5
Delayed decompressive surgery improves peri-infarct
perfusion in rats with space-occupying cerebral
infarction
67
Part II. Clinical studies
Chapter 6
Predictors of life-threatening brain edema in middle
cerebral artery infarction
83
Chapter 7
A new algorithm for sequential allocation of two
treatments in small clinical trials
99
Chapter 8
Hemicraniectomy After Middle cerebral artery
infarctionwithLife-threateningEdemaTrial(HAMLET).
Protocol for a randomized controlled trial of
decompressive
surgery
in
space-occupying
hemispheric infarction
111
Chapter 9
Appreciation of the informed consent procedure
in a randomized trial of decompressive surgery for
space-occupying hemispheric infarction
123
Chapter 10
Long-term cognitive outcome after decompressive
surgery for space-occupying hemispheric infarction
133
Chapter 11
Early decompressive surgery in space-occupying
hemispheric infarction: a pooled analysis of three
randomized controlled trials
143
Chapter 12
General discussion
159
Appendix
Summary
Samenvatting
Dankwoord
Curriculum vitae
Publications
169
171
175
181
185
187
Introduction
Chapter 1
General introduction
9
Chapter 1
I
n 1999 a 32 year old woman was admitted because of weakness of the left arm
and left leg. She had no medical history, worked as a school teacher, and had two
young children. On neurological examination she had a normal consciousness but
a severe left-sided hemiparesis. CT showed infarction of the right anterior and middle
cerebral artery (MCA) territory and a dissection of the right internal carotid artery
(Figure 1). Twenty-four hours after the onset of symptoms she deteriorated with a decrease in consciousness and dilatation of the right pupil. A CT scan revealed massive
space-occupying edema with tissue shift across the midline (Figure 2). Decompressive
surgery was considered, but not performed, because of uncertainty about the beneﬁt. Her neurological condition deteriorated, despite medical therapy on the stroke
unit and she died as a result of transtentorial herniation 48 hours after the onset of
symptoms.
Preface
Large cerebral infarcts may be associated with space-occupying edema, that may lead
to transtentorial or uncal herniation.1 Fatal space-occupying brain edema occurs in 1
to 5% of all patients with a supratentorial infarct.2,3 Case fatality rates of up to 80% have
been reported, despite maximal medical therapy on an intensive care unit.1,4
Diﬀerent treatment modalities have been suggested for patients with supratentorial
infarcts who deteriorate as a result of edema formation. None of these therapies
has been proven to improve clinical outcome.5 According to the guidelines of the
American Heart Association, these patients may be treated with osmotic agents
and hyperventilation.6 However, several reports suggest that these measures are
ineﬀective1,4,7 or even detrimental.8
Because of the limitations of medical therapies, there have been proposals for
decompressive surgery in patients with neurological deterioration due to large
hemispheric infarction and edema (Figure 3). The rationale of this therapy is to revert
brain tissue shifts and to normalize intracranial pressure, thereby preserving cerebral
blood ﬂow and preventing secondary damage.9 Decompressive surgery is not a new
modality for the treatment of patients with an intracranial mass.10 The technique is
relatively simple and consists of a large hemicraniectomy and a duraplasty. It can be
performed in every neurosurgical center. Case reports and non-randomized patient
series have suggested a substantial reduction of mortality after decompressive surgery
without an increase in the number of severely disabled survivors.11-14 However,
information on functional outcome and quality of life of the surviving patients is
insuﬃcient.13,14
The work presented in this thesis consists of the preparation and preliminary results
of a randomized clinical trial to study the eﬀect of decompressive surgery on functional
outcome in patients with hemispheric infarction, who deteriorate as a result of brain
edema formation.
10
Introduction
Figure 1 CT scan of a 32 year old patient with a large infarct in
the territory of the right anterior and middle cerebral arteries
one hour after the onset of symptoms.
Figure 2 CT scan of the same patient 24 hours after the onset of
symptoms, showing space-occupying edema and midline shift.
Figure 3 CT scan of a 44 year old patient with a large infarct
in the territory of the left anterior and middle cerebral arteries
after decompressive surgery.
11
Chapter 1
Outline of the thesis
In chapter 2 the existing evidence of eﬃcacy of treatment modalities for spaceoccupying hemispheric infarction is summarized. In chapter 3 a rat model of spaceoccupying cerebral infarction is presented. By means of this model the time course
of ischemic damage and perfusion as measured by magnetic resonance imaging
(MRI) are evaluated (chapter 3), brain perfusion is studied and non-invasive and
invasive MR perfusion measurements are compared (chapter 4), and the eﬀects of
decompressive surgery on tissue damage and perfusion are investigated (chapter 5).
In chapter 6 a meta-analysis on predictors of fatal edema formation is presented. In
chapter 7 an alternative way of randomizing patients is described, useful for relatively
small clinical trials. Chapter 8 contains the study protocol of the ongoing multi-center
randomized Hemicraniectomy After MCA infarction with Life-threatening Edema
Trial (HAMLET). The results of studies on recall and appreciation of the informed
consent procedure in HAMLET (chapter 9) and of cognitive outcome of a cohort of
patients after space-occupying MCA infarction and hemicraniectomy (chapter 10) are
presented. Chapter 11 brings the results of a pooled analysis of individual patient data
of three European trials on decompressive surgery after space-occupying hemispheric
infarction. In chapter 12 the implications of the studies described in this thesis on
patient care and future research are discussed.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
12
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 1996;53:309-315.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
Heinsius T, Bogousslavsky J, Van Melle G. Large infarcts in the middle cerebral artery territory. Etiology
and outcome patterns. Neurology 1998;50:341-350.
Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction and early brain swelling:
progression and eﬀect of age on outcome. Mayo Clin Proc 1998;73:829-836.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral infarction. Crit
Care Med 2003;31:617-625.
Adams Jr HP., Adams RJ, Brott T. Guidelines for the early management of patients with ischemic stroke:
A scientiﬁc statement from the Stroke Council of the American Stroke Association. Stroke 2003;34:10561083.
Schwab S, Spranger M, Schwarz S, Hacke W. Barbiturate coma in severe hemispheric stroke: useful or
obsolete? Neurology 1997;48:1608-1613.
Muizelaar JP, Marmarou A, Ward JD. Adverse eﬀects of prolonged hyperventilation in patients with
severe head injury: a randomized clinical trial. J Neurosurg 1991;75:731-739.
Schwab S, Rieke K, Aschoﬀ A, Albert F, von Kummer R, Hacke W. Hemicraniotomy in space-occupying
hemispheric infarction: useful early intervention or desparate activism? 1996;6:325-329.
Introduction
10.
11.
12.
13.
14.
Kerr FW. Radical decompression and dural grafting in severe cerebral edema. Mayo Clin Proc
1968;43:852-864.
Koh MS, Goh KY, Tung MY, Chan C. Is decompressive craniectomy for acute cerebral infarction of any
beneﬁt? Surg Neurol 2000;53:225-230.
Carter BS, Ogilvy CS, Candia GJ, Rosas HD, Buonanno F. One-year outcome after decompressive
surgery for massive nondominant hemispheric infarction. Neurosurgery 1997;40:1168-1175.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
13
Chapter 1
14
Treatment of space-occupying MCA infarction
Part I
Experimental studies
15
Treatment of space-occupying MCA infarction
Chapter 2
Treatment of space-occupying middle
cerebral artery infarction
Jeannette Hofmeijer, H. Bart van der Worp, and L. Jaap Kappelle
Based on Critical Care Medicine 2003;31:617-625
17
Chapter 2
Summary
Patients with a hemispheric infarct and massive edema have a poor prognosis. The
case fatality rate is approximately 80% and most survivors are left severely disabled.
Various treatment strategies have been proposed to limit brain tissue shifts and to
reduce intracranial pressure, but their use is controversial. In this chapter the results
of a systematic review of the literature addressing the evidence of eﬃcacy of these
treatments is presented.
Literature searches were performed in Medline and PubMed. Studies were included
if they were published in English between Januari 1966 and August 2006 and addressed
the eﬀect of osmotherapy, hyperventilation, barbiturates, steroids, hypothermia,
or decompressive surgery in supratentorial infarction with edema in animals or
humans.
Animal studies of medical treatment strategies in focal cerebral ischemia produced
conﬂicting results. If any, experimental support for these strategies is derived from
studies with animal models of moderately severe focal ischemia instead of severe
space-occupying infarction. None of the treatment options has improved outcome
in randomized clinical trials. Two large non-randomized studies of decompressive
surgery yielded promising results in terms of reduction of mortality and improvement
of functional outcome.
There is no treatment modality of proven eﬃcacy for patients with space-occupying
hemispheric infarction. Decompressive surgery might be the most eﬀective therapeutic
option. For decisive answers randomized controlled clinical trials are needed.
18
Treatment of space-occupying MCA infarction
L
arge cerebral infarcts are commonly associated with variable degrees of brain
edema. In severe cases this may lead to transtentorial or uncal herniation. Fatal
space-occupying brain edema occurs in 1-5% of all patients with a supratentorial infarct.1 Transtentorial herniation accounts for up to 78% of deaths during the
ﬁrst week after supratentorial infarction and up to 27% of deaths during the ﬁrst 30
days.1,2 In younger patients with ischemic stroke, herniation is the cause of about half
of the deaths in the ﬁrst month.3 In recent prospective intensive care-based series the
case fatality rate of space-occupying cerebral infarcts was about 80%, despite maximal
conservative therapy.4,5
Various treatment strategies have been proposed to limit brain tissue shifts and
to reduce intracranial pressure, such as osmotic therapy, hyperventilation, and
sedation with barbiturates.6-8 In the guidelines of the American Heart Association
osmotherapy and hyperventilation are recommended for patients whose condition
is deteriorating secondary to increased intracranial pressure or brain herniation.6
These treatment options are considered standard care by experts in various stroke
centers worldwide. However, no trials have addressed the eﬃcacy of these therapies
to improve clinical outcome,9 and several reports suggest that these are ineﬀective4,5,10
or even detrimental.11,12 In this chapter the evidence of eﬃcacy of treatments that have
been proposed to improve outcome after space-occupying hemispheric infarction will
be reviewed.
Methods
Literature searches were carried out in Medline and PubMed, using a combination
of keywords covering brain infarction, brain edema, and the diﬀerent options of
intervention. Keywords related to brain infarction were stroke, cerebral infarction,
and cerebral ischemia, and the alternatives presented in the Thesaurus of Medline.
Keywords related to edema were brain edema, brain swelling, and their alternatives.
Furthermore, relevant papers were checked for references. Studies were included if
they were published in English between Januari 1966 and August 2006 and addressed
treatment of edema in supratentorial focal cerebral ischemia in animals (Table 1) or
humans (Table 2). Treatment modalities that had only been used in animals, but not
in humans, were excluded.
19
20
Hoﬀ, 1982
Albright, 1984
Millson, 1981
Selman, 1982b
Kobayashi, 1995
Kotwica, 1991a
Selman, 1981b
Kobayashi, 1994
Selman, 1982b
Selman, 1981b
Corkill, 1978
Karibe, 1995
Harbaugh, 1979
Tuor, 1999
Slivka, 2001c
Moseley, 1975
Albright, 1984
Tuor, 1993
de Courten, 1994
Hoﬀ, 1975
Little, 1980
Chumas, 1993
Barks, 1991
Braughler, 1986
Little, 1979
Tosaki, 1985
Steroids
Smith, 1974
Barbiturates
Hoﬀ, 1973
Paczynski, 2000
Albright, 1984
Furosemide
Little, 1978
Suzuki, 1980
Suzuki, 1980
Hypertonic
Saline
Little, 1978
Popovic, 1978
Dietis, 1986
Dodson, 1975
Karibe, 1995
Paczynski, 2000
Meyer, 1972
Ehteshami, 1988
Paczynski, 1997
Glycerol
Mannitol
Table 1 Summary of experimental studies on treatment of hemispheric infarction.
ICP lowering
Reduced ischemic damage
Reduced edema
Treatment result
Chapter 2
a
Steroids
Donley, 1973
Barbiturates
Hoﬀ, 1982
Increased volume ratios/shift
Increased ICP
indicates rebound phenomenon; b, with reperfusion and if therapy was initiated within 30 minutes after ischemia; c, only after reperfusion
Paczynski, 2000
Paczynski, 1997
Koide, 1986
Increased edema
Increased infarct volume
Kaufmann, 1992
Sapolsky, 1985
Accumulation in infarcted tissue
No eﬀect on histopathology
Treatment result
Kaufmann, 1992
Oktem, 2000
Bhardwaj, 2000
de la Torre, 1976
Furosemide
Gueniau, 1997
Bhardwaj, 2000
Hypertonic
Saline
Lee, 1974
Glycerol
Ogilvy, 1996
Koc, 1994
Mannitol
Table 1 – Continued –
Treatment of space-occupying MCA infarction
21
22
a
Patten, 1972
Steroids
a
indicates temporary eﬀect with reduction of cerebral perfusion pressure
Norris, 1976
a’Rogvi, 2000
Ogun, 2001
Norris, 1986
Santambrogio, 1978
Mulley, 1978
Kaste, 1976
Yu, 1993
Santambrogio, 1978
Bauer, 1973
Larsson, 1976
Candalise, 1975
Bayer, 1987
Fawer, 1978
Rockoﬀ, 1979
Schwab, 1997
Barbiturates
ICP lowering
Treatment result
No eﬀect on outcome
Improved outcome
Increased volume ratios
Furosemide
Videen, 2001
Schwarz, 1998
Hypertonic Saline
No eﬀect tissue shift
Mathew, 1972
Glycerol
Manno, 1999
Schwarz, 1998
Mannitol
Table 2 Summary of clinical studies on treatment of hemispheric infarction.
Chapter 2
Treatment of space-occupying MCA infarction
Results
Osmotherapy
Mannitol
Osmotic agents, such as mannitol, are presumed to draw water from interstitial
and intracellular spaces into the intravascular compartment by creating an osmotic
pressure gradient over the semi-permeable blood-brain barrier.13 In addition to its
osmotic capacity, reported eﬀects of mannitol include reduction of blood viscosity
and improvement of microvascular cerebral blood ﬂow,14-17 vasoconstriction with a
reduction in cerebral blood volume,18,19 and scavenging of free radicals.20
In various animal studies, mannitol, administered before or within 24 hours after the
onset of transient or permanent focal cerebral ischemia reduced edema formation or
ischemic damage.21-31 In a recent study, a trend towards a dose-dependent decrease in
the water content of the ischemic middle cerebral artery (MCA) cortex and infarcted
hemisphere was found. However, relatively more dehydration of the non-infarcted
hemisphere than of the hemisphere with the infarction was seen in rats receiving a
high dose (2.5mg / kg) than in rats receiving a low dose (0.5mg / kg) of mannitol,
suggesting that more severe hyperosmolar dehydration may increase pressure
gradients and aggravate tissue shifts.30 In a subsequent study, high doses of mannitol
(1.5g / kg every ﬁve hours) caused paradoxical increases in the water content of the
infarcted hemisphere, in the infarcted / non-infarcted hemisphere volume ratios, and
in midline shift. In contrast, low doses of mannitol had signiﬁcant positive eﬀects on
these variables.31
Other experimental studies failed to show a statistically signiﬁcant positive eﬀect
of mannitol therapy on infarct volume32-35 or cerebral edema.36 The lack of eﬃcacy
might be the result of continuous infusion instead of bolus administration35 or of
administration before rather than after the ischemic insult, so that the osmotic
gradient had already disappeared before the occurrence of edema.32 Other proposed
reasons for the lack of eﬀect include relatively late administration36 or administration
of too low doses of mannitol (0.2g / kg).33
Several authors have suggested that administration of hypertonic agents in the
presence of cerebral ischemic injury may be detrimental.12,35 To create an osmotic
gradient, an intact blood-brain barrier is needed. As the blood-brain barrier is
compromised during cerebral ischemia, it is presumed that water will be extracted
from healthy but not from ischemic brain tissue,37 resulting in a worsening of tissue
shifts. In cats with cortical cold injury, accumulation of mannitol in damaged brain
tissue has been reported after 5 doses of 0.33 g/kg. In the edematous white matter the
mannitol concentration exceeded the plasma concentration by a ratio of 2.69:1. This
caused a reversal of the osmotic gradient and aggravation of cerebral edema.12 In other
23
Chapter 2
animal studies, rebound phenomena have been observed.37 These have been attributed
to the longer elimination half-life of mannitol from cerebro-spinal ﬂuid (CSF) than
from blood, with a consequent temporary reversal of the serum / CSF concentration
gradient during elimination.38,39
In a clinical study in 9 patients with recent ischemic stroke, single doses of 40g
mannitol were eﬀective in temporarily reducing elevated ICP to more than 10% below
the baseline value in 10 of 14 episodes. The maximum eﬀect occurred at the end of
infusion and the eﬀect was visible over 4 hours.8 In a recent series of seven patients with
edema and midline shift due to hemispheric infarction, successive magnetic resonance
imaging (MRI) before, during, and after the administration of a bolus of mannitol (1,5g
/ kg) did not reveal any change in midline shifts, nor did the neurological status of the
patients change. Although these ﬁndings dispel some of the concerns of increases in
mass eﬀect after administration of mannitol, the clinical implications concerning either
beneﬁcial or harmfull eﬀects are limited, since the study did not address the eﬀects of
multiple dosing.40 Furthermore, subsequent analyses of hemispheric volumes of these
patients revealed that there was a slight reduction in brain volume after mannitol
treatment that was restricted to the non-infarcted hemisphere.41
No randomized clinical trial has addressed the eﬀect of mannitol on outcome in
patients with space-occupying hemispheric infarction. One early prospective42 and
one retrospective43 clinical study failed to show a signiﬁcant beneﬁt on outcome in
patients with acute stroke. However, these studies were not CT based and cases of
cerebral hemorrhage could therefore have been included inadvertently. In addition,
clinically diﬀerent infarct subtypes were included, and treatment was not primarily
aimed at reducing edema formation in large infarcts. In a systematic (Cochrane)
review outcome analyses could not be performed due to lack of appropriate trials.44
Glycerol
Glycerol has been reported to improve cerebral blood ﬂow,45,46 and to have edemareducing as well as neuroprotective properties.47 Only few experimental studies have
addressed the eﬀect of glycerol in cerebral infarction. In rat models of focal cerebral
ischemia the compound reduced edema.48-51
A single bolus decreased ICP52 or T2 hyperintensity on MRI53 rapidly in patients
with MCA infarction, but accumulated in infarcted tissue.52 Glycerol has been tested
in several randomized and non-randomized clinical trials of acute stroke, but none
of these speciﬁcally addressed its eﬀect on space-occupying hemispheric infarction.
A systematic (Cochrane) review of these trials suggests a favorable eﬀect of glycerol
treatment on short term survival, but no long term eﬃcacy. The lack of proven beneﬁt
on long term survival does not support the routine use of glycerol in patients with
acute ischemic stroke.54
24
Treatment of space-occupying MCA infarction
Hypertonic saline
Sodium chloride is actively excluded from an intact blood-brain barrier, which
theoretically makes it a more potent osmotic agent than mannitol.55 In recent studies
of transient focal cerebral ischemia in rats, edema in both the aﬀected and the
unaﬀected hemisphere decreased after continuous hypertonic (7.5%) saline infusion
started 6 or 24 hours after induction of ischemia respectively.56,57 In a comparable study
hypertonic saline increased rather than decreased infarct volume. Water content in the
contralateral (non-injured) hemisphere was signiﬁcantly less in the hypertonic salinetreated group than in the control group at 22 hours of reperfusion, whereas there was
no diﬀerence in water content of the injured hemisphere between the two groups.35
In patients with space-occupying hemispheric infarction or putaminal hemorrhage
with perifocal edema, single doses of hypertonic saline temporarily reduced elevated
ICP.8 A temporary reduction of elevated ICP, leading to an increased cerebral perfusion
pressure, was also seen in eight patients treated with 75ml 10% saline after conventional
therapy with mannitol had failed. However, the group of patients described in this
study was heterogeneous; six had space-occupying hemispheric infarction and two
had supra-tentorial hemorrhage with edema.58 In addition, patients received diﬀerent
combinations of other ICP lowering therapies (two with decompressive surgery, four
with hypothermia, and two with CSF drainage via an intraventricular catheter), which
makes the results diﬃcult to interpret. No clinical trials have addressed the eﬀect of
hypertonic saline on functional outcome.
Furosemide
Loop diuretics such as furosemide may decrease ICP by decreasing total body water
and increasing blood osmolality, and thereby removing water from the edematous
brain.59,60
In two studies of experimental brain edema induced by cortical freezing in rabbits,
furosemide signiﬁcantly decreased ICP.61,62 In rats with transient focal cerebral ischemia,
furosemide (0.5mg / kg 5 hourly) reduced body weight, but had no signiﬁcant eﬀect on
the water content of the aﬀected hemisphere.31
There have been no controlled clinical studies testing the eﬀect of loop diuretics on
outcome after ischemic stroke.
Barbiturates
Barbiturates may have neuroprotective properties by reducing the cerebral metabolic
rate63-65 and by acting as free radical scavengers.66,67 Reduction of the cerebral metabolic
rate and the subsequent lowering of cerebral blood ﬂow and volume could theoretically
reduce edema formation and lower ICP.
Barbiturate treatment ameliorated the clinical course and reduced lesion size in
25
Chapter 2
some animal studies of focal cerebral ischemia68-74 but had no eﬀect on edema and ICP
in experimental space-occupying cerebral infarction.74-76 In baboons with permanent
MCA occlusion fatal ICP elevation occurred even more often after barbiturate
treatment than in controls. On the other hand, barbiturates reduced mortality if
ischemia was transient and if treatment was initiated within 30 minutes after the onset
of ischemia, before edema formation had occurred.73,77
Case studies suggested that barbiturate treatment may be eﬀective to reduce
intracranial hypertension caused by traumatic brain injury,78 cerebral ischemia during
aneurysm surgery,79 or severe pre-eclampsia.80 However, in most clinical reports the
eﬀect of barbiturates on brain swelling secondary to infarction was disappointing.81,82
In the only prospective, but uncontrolled, clinical study on this subject the usefulness
of barbiturate coma in reducing elevated ICP after MCA infarction was limited.10
Barbiturate coma was induced with thiopental infusion in 60 patients with critically
increased ICP secondary to large hemispheric infarction, after failure of osmotherapy
and mild hyperventilation. Although doses were high enough to achieve a burstsuppression pattern on the EEG, and ICP was signiﬁcantly lowered in the early phases
after initiation of therapy, long term control of ICP could not be achieved. Moreover,
cerebral perfusion pressure decreased during the course of treatment. Only ﬁve of the
60 patients treated with thiopental survived; all other patients died from transtentorial
herniation. Randomized trials are lacking.
Steroids
In very high doses, steroids have been claimed to have neuroprotective properties
in ischemic stroke.83 In addition, corticosteroids reduce vasogenic cerebral edema in
patients with brain tumors.84
There is no evidence from experimental studies that steroids reduce edema in
cerebral infarction.85-88 This may be explained by the fact that edema in ischemic stroke
consists of both a vasogenic and a cytotoxic component.89
Dexamethasone improved outcome after acute stroke in a single placebocontrolled clinical trial,90 whereas in other clinical studies no favorable eﬀects of
dexamethasone42,91-96 or prednisolone97 treatment on clinical outcome were found.
Dexamethason increased short-term mortality in older patients with ischemic
stroke.98 A meta-analysis of randomized trials comparing corticosteroid treatment
with placebo in patients with acute ischemic stroke did not show a positive treatment
eﬀect on functional outcome.99 There are no trials addressing the eﬃcacy of steroids to
reduce edema formation in space-occupying cerebral infarction.
Hyperventilation
Hyperventilation lowers ICP by inducing serum and CSF alkalosis and vasoconstriction.100 However, the eﬀect of hyperventilation may diminish within hours.101
26
Treatment of space-occupying MCA infarction
Moreover, rebound vasodilatation with increases of ICP may occur if normoventilation
is resumed.59 This may even induce a steal phenomenon if vasodilatation is more
profound in healthy than in ischemic brain tissue.102 Several clinical studies suggested
that prolonged hyperventilation induces cerebral ischemia103,104 and worsens clinical
outcome in patients with traumatic brain injury.11
In primate models of focal brain ischemia, hypocapnia initiated after induction
of ischemia did not alter mortality, degree of neurological deﬁcit, or volume of the
infarct.105,106
Clinical studies in the early 1970’s addressing the eﬀect of normocapnic (40 mmHg)
and hypocapnic (20-25 mmHg) hyperventilation in stroke found no beneﬁcial treatment
eﬀect on patient outcome.107,108 In these studies hyperventilation was continued for 7274 hours. More recent clinical trials are lacking.
Hypothermia
Hypothermia is presumed to reduce cerebral ischemic damage by means of reducing
brain metabolism,109,110 preservation of the blood brain barrier,111 a reduction in the
inﬂammatory respons,112 and a reduced neurotransmitter release.113-115
In a variety of animal studies hypothermia reduced infarct size after focal cerebral
ischemia as mono-therapy112,116-129 or additionally, next to neuroprotective agents.130132
In experimental studies of hypothermia in transient focal133,134 or global135 cerebral
ischemia, a reduction of edema development during reperfusion was found. In only
one early study of acute ischemic stroke in primates, hypothermia (29°C) had a
detrimental eﬀect: all treated animals had infarction with massive edema and died.106
Two non-randomized studies in patients with severe space-occupying edema
after MCA infarction suggested that moderate hypothermia (32°C -34°C) on an ICU
could help to control critically elevated ICP values and improve clinical outcome.
Hypothermia was associated with several side eﬀects, of which thrombocytopenia,
bradycardia and pneumonia were most frequently encountered. Most deaths occurred
during rewarming as a result of excessive ICP rise and herniation.136,137 A shorter
rewarming period was associated with a more pronounced rise of ICP.137 Rebound
ICP rise could possibly be prevented by slow and controlled, instead of passive
rewarming.138 In a recent non-randomized open pilot study of hypothermia in severe
MCA infarction, cooling to 32 ± 1°C appeared to be safe. No signiﬁcant improvement
of functional outcome in the hypothermia treated group was seen, but sample sizes
were small and outcome trends favored hypothermia.139 Randomized trials have not
been performed.
Surgical decompression
Because of the limitations of medical therapies, there have been proposals for
decompressive surgery in patients with hemispheric infarction and neurological
27
Chapter 2
deterioration caused by space-occupying edema. This therapy is presumed to revert
brain tissue shifts, to normalize intracranial pressure, and to preserve cerebral blood
ﬂow, thus preventing secondary brain damage. The technique of decompressive surgery
is relatively simple and consists of a large hemicraniectomy and a duraplasty.140
Animal studies have shown that this intervention reduces mortality and improves
functional and histological outcome.141-144
Case reports and small retrospective or non-controlled studies suggested that
hemicraniectomy lowers mortality without increasing the number of severely
disabled survivors.145-153 This ﬁnding has been conﬁrmed in two prospective series, in
which patients younger than 70 years with clinical and CT evidence of acute severe
MCA infarction were included. In the ﬁrst series decompression was performed
in 32 patients after clinical deterioration consisting of a progressive decrease in
consciousness. Mortality was reduced from 79% in controls to 34% in surgicallytreated patients, and poor functional outcome from 95% to 50%. The mean interval
between the onset of symptoms and surgery was 39 hours.154 In a subsequent study,
in which hemicraniectomy was performed in 31 patients within 24 hours after the
onset of symptoms, mortality was reduced to 16%, without an increase in the number
of severely disabled survivors.155 Complications of the operative procedure were
generally not serious and had no eﬀect on patient outcome. Parenchymal bleeding
occurred more often with smaller bone resections.156 In another small prospective
series of patients (n=19), hemicraniectomy reduced mortality and improved short
term clinical outcome (Glasgow Outcome Scale at three months) as compared to a
non-randomized control group (n=15).157 Other reports suggest that decompressive
surgery is less eﬀective in elderly patients158 and that substantial recovery extends into
the second half year and thereafter.159 In recent non-controlled patient series, a better
outcome was observed in younger patients and associated with intervention as soon as
possible after160,161 or before162 the onset of clinical signs of herniation.
The results of two relatively large prospective studies154,155 suggest a substantial beneﬁt
of decompressive surgery as compared with medical therapy alone. However, groups
were not constituted by random selection. Control groups consisted of patients with a
signiﬁcantly higher age, more co-morbidity and more frequent lesions in the dominant
hemisphere than those in the surgical groups. In addition, information on functional
outcome was insuﬃcient.154,155 Randomized trials have not been performed.163
Discussion
None of the therapeutic strategies proposed to control cerebral edema formation and
to reduce tissue shifts and raised ICP after space-occupying hemispheric infarction
is supported by adequate evidence of eﬃcacy from experimental studies or clinical
trials. If any, experimental support for these strategies is derived from studies with
28
Treatment of space-occupying MCA infarction
animal models of moderately severe focal cerebral ischemia, not adequately reﬂecting
the syndrome of space-occupying infarction in man. Several studies suggest that the
beneﬁcial eﬀects of treatment with mannitol,24,164-166 hypothermia,167,168 or barbiturates73
demonstrated in transient or moderate focal cerebral ischemia may be absent in cases
of permanent or more severe infarcts.
In rats, edema formation after cerebral infarction occurs earlier than in man. On
histopathological examination of rat brains at 6, 24, and 72 hours, and at seven days
after the onset of transient focal cerebral ischemia brain water content peaked at 24
hours.169 In patients, clinical deterioration from serious edema formation usually
occurs between the second and the ﬁfth day after stroke onset.4,170-172 This diﬀerence
in the timing of edema formation may have consequences for the extrapolation of the
results of rat studies into the clinic.
Most treatments are based on the perception that a raised intracranial pressure is
the dominant cause of neurologic deterioration. However, displacement of brain tissue
rather than increased ICP is probably the most likely cause of the initial decrease in
consciousness and further neurological deterioration.170 One study, in which ICP
was monitored in 48 patients with clinical signs of increased ICP caused by large
hemispheric infarction, showed that ICP measurements were not helpful in guiding
long-term treatment.173 Reducing ICP with osmotic agents or hyperventilation might
even be harmful, because the reduction in volume of the contralateral hemisphere,
where the blood brain barrier and cerebral autoregulation are still intact, might be
more pronounced than that of the infarcted hemisphere, thereby increasing brain
tissue shifts.170 Moreover, osmotic agents like mannitol and glycerol may accumulate in
the aﬀected tissue, thereby reversing the osmotic gradient between tissue and plasma,
leading to an exacerbation of edema.12 Therefore, the outcomes of osmotic treatment
may be largely dependent on the timing and the duration of treatment.174
According to the guidelines of the American Heart Association, patients with
space-occupying cerebral infarction whose condition is deteriorating secondary to
edema formation should be treated with osmotic agents and hyperventilation.6 Other
experts recommend decompressive surgery and hypothermia for the treatment of
these patients. They suggest that early intervention generates better results in terms of
mortality and functional recovery of survivors, and that treatment should probably be
started even before clinical deterioration in patients with massive infarction.175 There
is no unequivocal evidence to support either opinion.
It remains unclear which patients should be candidates for intensive anti-edema
treatment. Several parameters have been studied as possible predictors of the
development of fatal brain edema. An increased risk was found to be associated with
clinical conditions such as a high NIH Stroke Scale score on admission, early nausea
and vomiting, hypertension, and heart failure, but the predictive value of the diﬀerent
conditions was weak.176,177 Radiological predictors of fatal brain edema include CT
29
Chapter 2
hypodensity of 50% or more of the MCA territory176,177 and lesion volume on diﬀusion
weighted MRI (DWI) exceeding 145cm3.178 Although DWI has high sensitivity and
speciﬁcity rates when performed within 14 hours after stroke onset, the sensitivity of
this parameter may be considerably lower in earlier phases of the infarct.179 In these
very early phases other parameters, such as a complete MCA territory perfusion
deﬁcit or MCA occlusion on MR angiography may be more sensitive predictors of the
development of malignant infarction.180,181 An unambiguous decision based on one or
on a combination of these parameters can not yet be made.
It remains unclear whether patients with severe aphasia should be treated as
aggressively as patients without. Despite severe language disturbances, quality of life
in these patients is not necessarily worse than in other patients.182 Therefore patients
with aphasia should not be excluded from trials testing treatment strategies in spaceoccupying infarction.
In animal studies hypothermia reduced infarct size very consistently. Furthermore,
non-randomized studies in patients with severe space-occupying edema after MCA
infarction suggested that moderate hypothermia (32°C -34°C) can help to control
critically elevated ICP values and to improve clinical outcome.136-138 Thus, hypothermia
deserves further research as a measure to prevent and treat massive edema formation.
Also surgical decompression is a promising treatment option, given the suggested
large reductions in mortality.154,155
Before implementation of the diﬀerent proposed strategies as standard treatment
modalities, data from randomized controlled clinical trials are needed. Multi-center
randomized trials of decompressive surgery for space-occupying hemispheric
infarction are on their way.183
References
1.
2.
3.
4.
5.
6.
7.
30
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
Heinsius T, Bogousslavsky J, Van Melle G. Large infarcts in the middle cerebral artery territory. Etiology
and outcome patterns. Neurology 1998;50:341-350.
Biller J, Adams-HP J, Bruno A, Love BB, Marsh EE. Mortality in acute cerebral infarction in young
adults--a ten-year experience. Angiology 1991;42:224-230.
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 1996;53:309-315.
Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction and early brain swelling:
progression and eﬀect of age on outcome. Mayo Clin Proc 1998;73:829-836.
Adams HP, Jr., Adams RJ, Brott T. Guidelines for the early management of patients with ischemic stroke:
A scientiﬁc statement from the Stroke Council of the American Stroke Association. Stroke 2003;34:10561083.
Wijdicks EF. Management of massive hemispheric cerebral infarct: is there a ray of hope? Mayo Clin
Proc 2000;75:945-952.
Treatment of space-occupying MCA infarction
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Schwarz S, Schwab S, Bertram M, Aschoﬀ A, Hacke W. Eﬀects of hypertonic saline hydroxyethyl starch
solution and mannitol in patients with increased intracranial pressure after stroke. Stroke 1998;29:15501555.
van der Worp HB, Kappelle LJ. Complications of acute ischaemic stroke. Cerebrovasc Dis 1998;8:124132.
Schwab S, Spranger M, Schwarz S, Hacke W. Barbiturate coma in severe hemispheric stroke: useful or
obsolete? Neurology 1997;48:1608-1613.
Muizelaar JP, Marmarou A, Ward JD. Adverse eﬀects of prolonged hyperventilation in patients with
severe head injury: a randomized clinical trial. J Neurosurg 1991;75:731-739.
Kaufmann AM, Cardoso ER. Aggravation of vasogenic cerebral edema by multiple-dose mannitol. J
Neurosurg 1992;77:584-589.
Schell RM, Applegate RL, Cole DJ. Salt, starch, and water on the brain. J Neurosurg Anesthesiol
1996;8:178-182.
Burke AM, Quest DO, Chien S, Cerri C. The eﬀects of mannitol on blood viscosity. J Neurosurg
1981;55:550-553.
Jafar JJ, Johns LM, Mullan SF. The eﬀect of mannitol on cerebral blood ﬂow. J Neurosurg 1986;64:754759.
Andrews RJ, Bringas JR, Muto RP. Eﬀects of mannitol on cerebral blood ﬂow, blood pressure, blood
viscosity, hematocrit, sodium, and potassium. Surg Neurol 1993;39:218-222.
Shirane R, Weinstein PR. Eﬀect of mannitol on local cerebral blood ﬂow after temporary complete
cerebral ischemia in rats. J Neurosurg 1992;76:486-492.
Muizelaar JP, Wei EP, Kontos HA, Becker DP. Mannitol causes compensatory cerebral vasoconstriction
and vasodilation in response to blood viscosity changes. J Neurosurg 1983;59:822-828.
Rosner MJ, Coley I. Cerebral perfusion pressure: a hemodynamic mechanism of mannitol and the
postmannitol hemogram. Neurosurgery 1987;21:147-156.
Suzuki J, Imaizumi S, Kayama T, Yoshimoto T. Chemiluminescence in hypoxic brain--the second
report: cerebral protective eﬀect of mannitol, vitamin E and glucocorticoid. Stroke 1985;16:695-700.
Little JR. Modiﬁcation of acute focal ischemia by treatment with mannitol and high-dose dexamethasone.
J Neurosurg 1978;49:517-524.
Little JR. Modiﬁcation of acute focal ischemia by treatment with mannitol. Stroke 1978;9:4-9.
Little JR. Treatment of acute focal cerebral ischemia with intermittent, low dose mannitol. Neurosurgery
1979;5:687-691.
Little JR. Morphological changes in acute focal ischemia: response to osmotherapy. Adv Neurol
1980;28:443-457.
Suzuki J, Tanaka S, Yoshimoto T, Seki H. Recirculation in the acute period of cerebral infarction:
experimental research on brain swelling and its suppression by using mannitol or glycerol. Acta
Neurochir Wien 1980;54:219-231.
Ehteshami S, Aspey BS, Hurst CM, McCoy AL, Harrison MJ. The combined eﬀects of hypertension,
hemodilution, and osmotherapy on the metabolic sequelae of acute experimental cerebral ischemia.
Metab Brain Dis 1988;3:235-244.
Karibe H, Zarow GJ, Weinstein PR. Use of mild intraischemic hypothermia versus mannitol to reduce
infarct size after temporary middle cerebral artery occlusion in rats. J Neurosurg 1995;83:93-98.
Kobayashi H, Ide H, Kabuto M, Handa Y, Kubota T, Ishii Y. Eﬀect of mannitol on focal cerebral ischemia
evaluated by somatosensory-evoked potentials and magnetic resonance imaging. Surg Neurol
31
Chapter 2
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
32
1995;44:55-61.
Kobayashi H, Ide H, Kodera T. Eﬀect of mannitol on focal cerebral ischemia evaluated by magnetic
resonance imaging. Acta Neurochir Suppl Wien 1994;60:228-230.
Paczynski RP, He YY, Diringer MN, Hsu CY. Multiple-dose mannitol reduces brain water content in a
rat model of cortical infarction. Stroke 1997;28:1437-1443.
Paczynski RP, Venkatesan R, Diringer MN, He YY, Hsu CY, Lin W. Eﬀects of ﬂuid management on
edema volume and midline shift in a rat model of ischemic stroke. Stroke 2000;31:1702-1708.
Ogilvy CS, Chu D, Kaplan S. Mild hypothermia, hypertension, and mannitol are protective against
infarction during experimental intracranial temporary vessel occlusion. Neurosurgery 1996;38:12021209.
Oktem IS, Menku A, Akdemir H, Kontas O, Kurtsoy A, Koc RK. Therapeutic eﬀect of tirilazad mesylate
(U-74006F), mannitol, and their combination on experimental ischemia. Res Exp Med Berl 2000;199:231242.
Koc RK, Akdemir H, Kandemir O, Pasaoglu H, Oktem IS, Pasaoglu A. The therapeutic value of naloxone
and mannitol in experimental focal cerebral ischemia. Neurological outcome, histopathological
ﬁndings, and tissue concentrations of Na+, K+ and water. Res Exp Med Berl 1994;194:277-285.
Bhardwaj A, Harukuni I, Murphy SJ. Hypertonic saline worsens infarct volume after transient focal
ischemia in rats. Stroke 2000;31:1694-1701.
Gueniau C, Oberlander C. The kappa opioid agonist niravoline decreases brain edema in the mouse
middle cerebral artery occlusion model of stroke. J Pharmacol Exp Ther 1997;282:1-6.
Kotwica Z, Persson L. Eﬀect of mannitol on intracranial pressure in focal cerebral ischemia. An
experimental study in a rat. Mater Med Pol 1991;23:280-284.
Nau R, Prins FJ, Kolenda H, Prange HW. Temporary reversal of serum to cerebrospinal ﬂuid glycerol
concentration gradient after intravenous infusion of glycerol. Eur J Clin Pharmacol 1992;42:181-185.
Nau R, Desel H, Lassek C. Slow elimination of mannitol from human cerebrospinal ﬂuid. Eur J Clin
Pharmacol 1997;53:271-274.
Manno EM, Adams RE, Derdeyn CP, Powers WJ, Diringer MN. The eﬀects of mannitol on cerebral
edema after large hemispheric cerebral infarct. Neurology 1999;52:583-587.
Videen TO, Zazulia AR, Manno EM, et al. Mannitol bolus preferentially shrinks non-infarcted brain in
patients with ischemic stroke. Neurology 2001;11:2120-2122.
Santambrogio S, Martinotti R, Sardella F, Porro F, Randazzo A. Is there a real treatment for stroke?
Clinical and statistical comparison of diﬀerent treatments in 300 patients. Stroke 1978;9:130-132.
Candelise L, Colombo A, Spinnler H. Therapy against brain swelling in stroke patients. A retrospective
clinical study on 227 patients. Stroke 1975;6:353-356.
Bereczki D, Liu M, do PG, Fekete I. Mannitol for acute stroke. Cochrane Database Syst Rev 2002
Jan;CD001153.
Antonini FM, Bertini G, Fumagalli C. Eﬀects of intravenous infusion of glycerol on regional cerebral
blood ﬂow in cerebral infarction. Gerontology 1977;23:376-380.
Ott EO, Mathew NT, Meyer JS. Redistribution of regional cerebral blood ﬂow after glycerol infusion in
acute cerebral infarction. Neurology 1974;24:1117-1126.
Meyer JS, Itoh Y, Okamoto S. Circulatory and metabolic eﬀects of glycerol infusion in patients with
recent cerebral infarction. Circulation 1975;51:701-712.
Dodson RF, Tagashira Y, Wai-Fong CL. The eﬀects of glycerol on cerebral ultrastructure following
experimentally induced cerebral ischemia. J Neurol Sci 1975;26:235-243.
Treatment of space-occupying MCA infarction
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
Popovic P, Popovic V, Schaﬀer R, Sutton CH. Treatment of experimental cerebral infarction in rats with
levodopa or with glycerol. J Neurosurg 1978;48:962-969.
Dietis A, Ehteshami S, Harrison MJ, Perinpanayagam NI. The eﬀect of isovolaemic haemodilution and
intravenous glycerol on the sequelae of middle cerebral artery occlusion in the rat. J Neurol Neurosurg
Psychiatry 1986;49:428-430.
Meyer JS, Teraura T, Marx P, Hashi K, Sakamoto K. Brain swelling due to experimental cerebral
infarction. Changes in vasomotor capacitance and eﬀects of intravenous glycerol. Brain 1972;95:833852.
Berger C, Sakowitz OW, Kiening KL, Schwab S. Neurochemical monitoring of glycerol therapy in
patients with ischemic brain edema. Stroke 2005;36:e4-e6.
Sakamaki M, Igarashi H, Nishiyama Y. Eﬀect of glycerol on ischemic cerebral edema assessed by
magnetic resonance imaging. J Neurol Sci 2003;15:69-74.
Righetti E, Celani MG, Cantisani T, Sterzi R, Boysen G, Ricci S. Glycerol for acute stroke. Cochrane
Database Syst Rev 2004;CD000096.
Zornow MH. Hypertonic saline as a safe and eﬃcacious treatment of intracranial hypertension. J
Neurosurg Anesthesiol 1996;8:175-177.
Toung TJ, Chang Y, Lin J, Bhardwaj A. Increases in lung and brain water following experimental stroke:
eﬀect of mannitol and hypertonic saline. Crit Care Med 2005;33:203-208.
Toung TJ, Hurn PD, Traystman RJ, Bhardwaj A. Global brain water increases after experimental focal
cerebral ischemia: eﬀect of hypertonic saline. Crit Care Med 2002;30:644-649.
Schwarz S, Georgiadis D, Aschoﬀ A, Schwab S. Eﬀects of hypertonic (10%) saline in patients with raised
intracranial pressure after stroke. Stroke 2002;33:136-140.
Hacke W, Schwab S, De Georgia M. Intensive Care of Acute Ischemic Stroke. Cerebrovasc Dis
1994;4:385-392.
Woster PS, LeBlanc KL. Management of elevated intracranial pressure. Clin Pharm 1990;9:762-772.
Harbaugh RD, James HE, Marshall LF, Shapiro HM, Laurin R. Acute therapeutic modalities for
experimental vasogenic edema. Neurosurgery 1979;5:656-665.
Millson C, James HE, Shapiro HM, Laurin R. Intracranial hypertension and brain oedema in albino
rabbits. Part 2: Eﬀects of acute therapy with diuretics. Acta Neurochir (Wien) 1981;56:167-181.
Nordstrom CH, Messeter K, Sundbarg G, Schalen W, Werner M, Ryding E. Cerebral blood ﬂow,
vasoreactivity, and oxygen consumption during barbiturate therapy in severe traumatic brain lesions. J
Neurosurg 1988;68:424-431.
Kassell NF, Hitchon PW, Gerk MK, Sokoll MD, Hill TR. Alterations in cerebral blood ﬂow, oxygen
metabolism, and electrical activity produced by high dose sodium thiopental. Neurosurgery 1980;7:598603.
Ochiai C, Asano T, Takakura K, Fukuda T, Horizoe H, Morimoto Y. Mechanisms of cerebral protection
by pentobarbital and nizofenone correlated with the course of local cerebral blood ﬂow changes. Stroke
1982;13:788-796.
Smith DS, Rehncrona S, Siesjo BK. Inhibitory eﬀects of diﬀerent barbiturates on lipid peroxidation in
brain tissue in vitro: comparison with the eﬀects of promethazine and chlorpromazine. Anesthesiology
1980;53:186-194.
Smith DS, Rehncrona S, Siesjo BK. Barbiturates as protective agents in brain ischemia and as free
radical scavengers in vitro. Acta Physiol Scand Suppl 1980;492:129-134.
33
Chapter 2
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
34
Hoﬀ J, Smith A, Nielsen S, Larson P. Eﬀects of barbiturate and halothane anesthesia on focal cerebral
infarction in the dog. Surg Forum 1973;24:449-451.
Smith AL, Hoﬀ JT, Nielsen SL, Larson CP. Barbiturate protection in acute focal cerebral ischemia.
Stroke 1974;5:1-7.
Hoﬀ JT, Smith AL, Hankinson HL, Nielsen SL. Barbiturate protection from cerebral infarction in
primates. Stroke 1975;6:28-33.
Moseley JI, Laurent JP, Molinari GF. Barbiturate attenuation of the clinical course and pathologic lesions
in a primate stroke model. Neurology 1975;25:870-874.
Corkill G, Sivalingam S, Reitan JA, Gilroy BA, Helphrey MG. Dose dependency of the post-insult
protective eﬀect of pentobarbital in the canine experimental stroke model. Stroke 1978;9:10-12.
Selman WR, Spetzler RF, Roessmann UR, Rosenblatt JI, Crumrine RC. Barbiturate-induced coma
therapy for focal cerebral ischemia. Eﬀect after temporary and permanent MCA occlusion. J Neurosurg
1981;55:220-226.
Hoﬀ JT, Nishimura M, Newﬁeld P. Pentobarbital protection from cerebral infarction without suppression
of edema. Stroke 1982;13:623-628.
Pappas TN, Mironovich RO. Barbiturate-induced coma to protect against cerebral ischemia and
increased intracranial pressure. Am J Hosp Pharm 1981;38:494-498.
Shapiro HM. Barbiturates in brain ischaemia. Br J Anaesth 1985;57:82-95.
Selman WR, Spetzler RF, Roski RA, Roessmann U, Crumrine R, Macko R. Barbiturate coma in focal
cerebral ischemia. Relationship of protection to timing of therapy. J Neurosurg 1982;56:685-690.
Woodhurst WB, Robertson WD, Thompson GB. Carotid injury due to intraoral trauma: case report and
review of the literature. Neurosurgery 1980;6:559-563.
Belopavlovic M, Buchthal A. Barbiturate therapy in the management of cerebral ischaemia. Anaesthesia
1980;35:271-278.
Caseby NG. Postpartum stroke successfully treated with high-dose pentobarbitone therapy: a case
report. Can Anaesth Soc J 1983;30:77-83.
Woodcock J, Ropper AH, Kennedy SK. High dose barbiturates in non-traumatic brain swelling: ICP
reduction and eﬀect on outcome. Stroke 1982;13:785-787.
Rockoﬀ MA, Marshall LF, Shapiro HM. High-dose barbiturate therapy in humans: a clinical review of
60 patients. Ann Neurol 1979;6:194-199.
Hall ED. The neuroprotective pharmacology of methylprednisolone. J Neurosurg 1992;76:13-22.
French L. The use of steroids in the treatment of cerebral oedema. Bull N Y Acad Med 1966;42:301311.
Donley RF, Sundt-TM J. The eﬀect of dexamethasone on the edema of focal cerebral ischemia. Stroke
1973;4:148-155.
Lee MC, Mastri AR, Waltz AG, Loewenson RB. Ineﬀectiveness of dexamethasone for treatment of
experimental cerebral infarction. Stroke 1974;5:216-218.
de-la-Torre JC, Surgeon JW. Dexamethasone and DMSO in experimental transorbital cerebral
infarction. Stroke 1976;7:577-583.
Slivka AP, Murphy EJ. High-dose methylprednisolone treatment in experimental focal cerebral ischemia.
Exp Neurol 2001;167:166-172.
Klatzo I. Neuropathological aspects of brain edema. J Neuropathol Exp Neurol 1967;26:1-14.
Patten BM, Mendell J, Bruun B, Curtin W, Carter S. Double-blind study of the eﬀects of dexamethasone
on acute stroke. Neurology 1972;22:377-383.
Treatment of space-occupying MCA infarction
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
Bauer RB, Tellez H. Dexamethasone as treatment in cerebrovascular disease. 2. A controlled study in
acute cerebral infarction. Stroke 1973;4:547-555.
Kaste M, Fogelholm R, Waltimo O. Combined dexamethasone and low-molecular-weight dextran in
acute brain infarction: double-blind study. Br Med J 1976 11;2:1409-1410.
Norris JW. Steroid therapy in acute cerebral infarction. Arch Neurol 1976;33:69-71.
Mulley G, Wilcox RG, Mitchell JR. Dexamethasone in acute stroke. Br Med J 1978;7:994-996.
Norris JW, Hachinski VC. High dose steroid treatment in cerebral infarction. Br Med J Clin Res Ed
1986;4:21-23.
Ogun SA, Odusote KA. Eﬀectiveness of high dose dexamethasone in the treatment of acute stroke.
West Afr J Med 2001;20:1-6.
Hetzel BS, Lander H, Robson H. Immediate treatment of apoplexy. BMJ 1957;1:1122.
Zuliani G, Cherubini A, Atti AR. Prescription of anti-oedema agents and short-term mortality in older
patients with acute ischaemic stroke. Drugs Aging 2004;21:273-278.
Qizilbash N, Lewington SL, Lopez-Arrieta JM. Corticosteroids for acute ischaemic stroke. Cochrane
Database Syst Rev 2002;CD000064.
Marshall LF, Smith RW, Shapiro HM. The outcome with aggressive treatment in severe head injuries.
Part I: the signiﬁcance of intracranial pressure monitoring. J Neurosurg 1979;50:20-25.
Muizelaar JP, van der Poel HG, Li ZC, Kontos HA, Levasseur JE. Pial arteriolar vessel diameter and CO2
reactivity during prolonged hyperventilation in the rabbit. J Neurosurg 1988;69:923-927.
Nariai T, Senda M, Ishii K. Posthyperventilatory steal response in chronic cerebral hemodynamic stress:
a positron emission tomography study. Stroke 1998;29:1281-1292.
Yundt KD, Diringer MN. The use of hyperventilation and its impact on cerebral ischemia in the
treatment of traumatic brain injury. Crit Care Clin 1997;13:163-184.
Stringer WA, Hasso AN, Thompson JR, Hinshaw DB, Jordan KG. Hyperventilation-induced cerebral
ischemia in patients with acute brain lesions: demonstration by xenon-enhanced CT. AJNR 1993;14:475484.
Soloway M, Moriarty G, Fraser JG, White RJ. Eﬀect of delayed hyperventilation on experimental
cerebral infarction. Neurology 1971;21:479-485.
Michenfelder JD, Milde JH. Failure of prolonged hypocapnia, hypothermia, or hypertension to favorably
alter acute stroke in primates. Stroke 1977;8:87-91.
Christensen MS, Paulson OB, Olesen J. Cerebral apoplexy (stroke) treated with or without prolonged
artiﬁcial hyperventilation. 1. Cerebral circulation, clinical course, and cause of death. Stroke
1973;4:568-631.
Simard D, Paulson OB. Artiﬁcal hyperventilation in stroke. Trans Am Neurol Assoc 1973;98:309310.
Chopp M, Knight R, Tidwell CD, Helpern JA, Brown E, Welch KM. The metabolic eﬀects of mild
hypothermia on global cerebral ischemia and recirculation in the cat: comparison to normothermia
and hyperthermia. J Cereb Blood Flow Metab 1989;9:141-148.
Berntman L, Welsh FA, Harp JR. Cerebral protective eﬀect of low-grade hypothermia. Anesthesiology
1981;55:495-498.
Dietrich WD, Busto R, Halley M, Valdes I. The importance of brain temperature in alterations of the
blood-brain barrier following cerebral ischemia. J Neuropathol Exp Neurol 1990;49:486-497.
Toyoda T, Suzuki S, Kassell NF, Lee KS. Intraischemic hypothermia attenuates neutrophil inﬁltration
in the rat neocortex after focal ischemia-reperfusion injury. Neurosurgery 1996;39:1200-1205.
35
Chapter 2
113. Busto R, Globus MY, Dietrich WD, Martinez E, Valdes I, Ginsberg MD. Eﬀect of mild hypothermia on
ischemia-induced release of neurotransmitters and free fatty acids in rat brain. Stroke 1989;20:904910.
114. Huang FP, Zhou LF, Yang GY. Eﬀects of mild hypothermia on the release of regional glutamate and
glycine during extended transient focal cerebral ischemia in rats. Neurochem Res 1998;23:991-996.
115. Winfree CJ, Baker CJ, Connolly-ES J, Fiore AJ, Solomon RA. Mild hypothermia reduces penumbral
glutamate levels in the rat permanent focal cerebral ischemia model. Neurosurgery 1996;38:12161222.
116. Goto Y, Kassell NF, Hiramatsu K, Soleau SW, Lee KS. Eﬀects of intraischemic hypothermia on cerebral
damage in a model of reversible focal ischemia. Neurosurgery 1993;32:980-984.
117. Jiang Q, Chopp M, Zhang ZG. The eﬀect of hypothermia on transient focal ischemia in rat brain
evaluated by diﬀusion- and perfusion-weighted NMR imaging. J Cereb Blood Flow Metab 1994;14:732741.
118. Karibe H, Chen J, Zarow GJ, Graham SH, Weinstein PR. Delayed induction of mild hypothermia to
reduce infarct volume after temporary middle cerebral artery occlusion in rats. J Neurosurg
1994;80:112-119.
119. Moyer DJ, Welsh FA, Zager EL. Spontaneous cerebral hypothermia diminishes focal infarction in rat
brain. Stroke 1992;23:1812-1816.
120. Kawai N, Okauchi M, Morisaki K, Nagao S. Eﬀects of delayed intraischemic and postischemic
hypothermia on a focal model of transient cerebral ischemia in rats. Stroke 2000;31:1982-1989.
121. Yanamoto H, Hong SC, Soleau S, Kassell NF, Lee KS. Mild postischemic hypothermia limits cerebral
injury following transient focal ischemia in rat neocortex. Brain Res 1996;29:207-211.
122. Yanamoto H, Nagata I, Niitsu Y, et al. Prolonged mild hypothermia therapy protects the brain against
permanent focal ischemia. Stroke 2001;32:232-239.
123. Markarian GZ, Lee JH, Stein DJ, Hong SC. Mild hypothermia: therapeutic window after experimental
cerebral ischemia. Neurosurgery 1996;38:542-550.
124. Zhang ZG, Chopp M, Chen H. Duration dependent post-ischemic hypothermia alleviates cortical
damage after transient middle cerebral artery occlusion in the rat. J Neurol Sci 1993;117:240-244.
125. Zhang RL, Chopp M, Chen H, Garcia JH, Zhang ZG. Postischemic (1 hour) hypothermia signiﬁcantly
reduces ischemic cell damage in rats subjected to 2 hours of middle cerebral artery occlusion. Stroke
1993;24:1235-1240.
126. Huh PW, Belayev L, Zhao W, Koch S, Busto R, Ginsberg MD. Comparative neuroprotective eﬃcacy of
prolonged moderate intraischemic and postischemic hypothermia in focal cerebral ischemia. J
Neurosurg 2000;92:91-99.
127. Kader A, Brisman MH, Maraire N, Huh JT, Solomon RA. The eﬀect of mild hypothermia on permanent
focal ischemia in the rat. Neurosurgery 1992;31:1056-1060.
128. Baker CJ, Onesti ST, Solomon RA. Reduction by delayed hypothermia of cerebral infarction following
middle cerebral artery occlusion in the rat: a time-course study. J Neurosurg 1992;77:438-444.
129. Maier CM, Sun GH, Kunis D, Yenari MA, Steinberg GK. Delayed induction and long-term eﬀects of
mild hypothermia in a focal model of transient cerebral ischemia: neurological outcome and infarct
size. J Neurosurg 2001;94:90-96.
130. Kamiya T, Nito C, Ueda M. Mild hypothermia enhances the neuroprotective eﬀects of a selective
thrombin inhibitor following transient focal ischemia in rats. Acta Neurochir Suppl 2003;86:195198.
36
Treatment of space-occupying MCA infarction
131. Nito C, Kamiya T, Ueda M, Arii T, Katayama Y. Mild hypothermia enhances the neuroprotective
eﬀects of FK506 and expands its therapeutic window following transient focal ischemia in rats. Brain
Res 2004;22:179-185.
132. Nito C, Kamiya T, Amemiya S, Katoh K, Katayama Y. The neuroprotective eﬀect of a free radical
scavenger and mild hypothermia following transient focal ischemia in rats. Acta Neurochir Suppl
2003;86:199-203.
133. Karibe H, Zarow GJ, Graham SH, Weinstein PR. Mild intraischemic hypothermia reduces
postischemic hyperperfusion, delayed postischemic hypoperfusion, blood-brain barrier disruption,
brain edema, and neuronal damage volume after temporary focal cerebral ischemia in rats. J Cereb
Blood Flow Metab 1994;14:620-627.
134. Kollmar R, Schabitz WR, Heiland S. Neuroprotective eﬀect of delayed moderate hypothermia after
focal cerebral ischemia: an MRI study. Stroke 2002;33:1899-1904.
135. Dempsey RJ, Combs DJ, Maley ME, Cowen DE, Roy MW, Donaldson DL. Moderate hypothermia
reduces postischemic edema development and leukotriene production. Neurosurgery 1987;21:177181.
136. Schwab S, Schwarz S, Spranger M, Keller E, Bertram M, Hacke W. Moderate hypothermia in the
treatment of patients with severe middle cerebral artery infarction. Stroke 1998;29:2461-2466.
137. Schwab S, Georgiadis D, Berrouschot J, Schellinger PD, Graﬀagnino C, Mayer SA. Feasibility and
safety of moderate hypothermia after massive hemispheric infarction. Stroke 2001;32:2033-2035.
138. Steiner T, Friede T, Aschoﬀ A, Schellinger PD, Schwab S, Hacke W. Eﬀect and feasibility of controlled
rewarming after moderate hypothermia in stroke patients with malignant infarction of the middle
cerebral artery. Stroke 2001;32:2833-2835.
139. Krieger DW, De Georgia M, Abou-Chebl A, Andrefsky JC, Sila CA, Katzan IL, Mayberg MR, Furlan
AJ. Cooling for acute ischemic brain damage (COOL AID). An open pilot study of induced hypothermia
in acute ischemic stroke. Stroke 2001;32:1847-1854.
140. Schwab S, Rieke K, Aschoﬀ A, Albert F, von Kummer R, Hacke W. Hemicraniotomy in space-occupying
hemispheric infarction: useful early intervention or desparate activism? Cerebrovasc Dis 1996;6:325329.
141. Doerﬂer A, Forsting M, Reith W. Decompressive craniectomy in a rat model of “malignant” cerebral
hemispheric stroke: experimental support for an aggressive therapeutic approach. J Neurosurg
1996;85:853-859.
142. Forsting M, Reith W, Schabitz WR. Decompressive craniectomy for cerebral infarction. An
experimental study in rats. Stroke 1995;26:259-264.
143. Engelhorn T, Doerﬂer A, Kastrup A. Decompressive craniectomy, reperfusion, or a combination for
early treatment of acute “malignant” cerebral hemispheric stroke in rats? Potential mechanisms
studied by MRI. Stroke 1999;30:1456-1463.
144. 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 2002;33:617-622.
145. Rengachary SS, Batnitzky S, Morantz RA, Arjunan K, Jeﬀries B. Hemicraniectomy for acute massive
cerebral infarction. Neurosurgery 1981;8:321-328.
146. Robertson SC, Lennarson P, Hasan DM, Traynelis VC. Clinical course and surgical management of
massive cerebral infarction. Neurosurgery 2004;55:55-61.
147. Reddy AK, Saradhi V, Panigrahi M, Rao TN, Tripathi P, Meena AK. Decompressive craniectomy for
37
Chapter 2
stroke : indications and results. Neurol India 2002;50 Suppl:S66-S69.
148. Walz B, Zimmermann C, Bottger S, Haberl RL. Prognosis of patients after hemicraniectomy in
malignant middle cerebral artery infarction. J Neurol 2002;249:1183-1190.
149. Carter BS, Ogilvy CS, Candia GJ, Rosas HD, Buonanno F. One-year outcome after decompressive
surgery for massive nondominant hemispheric infarction. Neurosurgery 1997;40:1168-1175.
150. Delashaw JB, Broaddus WC, Kassell NF. Treatment of right hemispheric cerebral infarction by
hemicraniectomy. Stroke 1990;21:874-881.
151. Kondziolka D, Fazl M. Functional recovery after decompressive craniectomy for cerebral infarction.
Neurosurgery 1988;23:143-147.
152. Kalia KK, Yonas H. An aggressive approach to massive middle cerebral artery infarction. Arch Neurol
1993;50:1293-1297.
153. Koh MS, Goh KY, Tung MY, Chan C. Is decompressive craniectomy for acute cerebral infarction of
any beneﬁt? Surg Neurol 2000;53:225-230.
154. Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
155. Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
156. Wagner S, Schnippering H, Aschoﬀ A, Koziol JA, Schwab S, Steiner T. Suboptimum hemicraniectomy
as a cause of additional cerebral lesions in patients with malignant infarction of the middle cerebral
artery. J Neurosurg 2001;94:693-696.
157. Mori K, Aoki A, Yamamoto T, Horinaka N, Maeda M. Aggressive decompressive surgery in patients
with massive hemispheric embolic cerebral infarction associated with severe brain swelling. Acta
Neurochir (Wien) 2002;143:483-491.
158. Holtkamp M, Buchheim K, Unterberg A. Hemicraniectomy in elderly patients with space occupying
media infarction: improved survival but poor functional outcome. J Neurol Neurosurg Psychiatry
2001;70:226-228.
159. Manai R, Srour A, Crozier S, Vandamme X, Samson Y, Cornu P, Rancurel G. Long-term functional
outcome of hemicraniectomy in middle cerebral artery malignant infarcts [abstract]. J Neurol 2001;
248 (suppl 2):121-122.
160. Kilincer C, Asil T, Utku U. Factors aﬀecting the outcome of decompressive craniectomy for large
hemispheric infarctions: a prospective cohort study. Acta Neurochir (Wien ) 2005;147:587-594.
161. Kastrau F, Wolter M, Huber W, Block F. Recovery from aphasia after hemicraniectomy for infarction
of the speech-dominant hemisphere. Stroke 2005;36:825-829.
162. Mori K, Nakao Y, Yamamoto T, Maeda M. Early external decompressive craniectomy with duroplasty
improves functional recovery in patients with massive hemispheric embolic infarction: timing and
indication of decompressive surgery for malignant cerebral infarction. Surg Neurol 2004;62:420-429.
163. Morley NC, Berge E, Cruz-Flores S, Whittle IR. Surgical decompression for cerebral oedema in acute
ischaemic stroke. Cochrane Database Syst Rev 2002;CD003435.
164. Meyer FB, Anderson RE, Sundt-TM J, Yaksh TL. Treatment of experimental focal cerebral ischemia
with mannitol. Assessment by intracellular brain pH, cortical blood ﬂow, and electroencephalography.
J Neurosurg 1987;66:109-115.
165. Seki H, Yoshimoto T, Ogawa A, Suzuki J. Eﬀect of mannitol on rCBF in canine thalamic ischemia--an
experimental study. Stroke 1983;14:46-50.
166. Suzuki J, Tanaka S, Yoshimoto T. Recirculation in the acute period of cerebral infarction: brain swelling
38
Treatment of space-occupying MCA infarction
and its suppression using mannitol. Surg Neurol 1980;14:467-472.
167. Ridenour TR, Warner DS, Todd MM, McAllister AC. Mild hypothermia reduces infarct size resulting
from temporary but not permanent focal ischemia in rats. Stroke 1992;23:733-738.
168. Morikawa E, Ginsberg MD, Dietrich WD, et al. The signiﬁcance of brain temperature in focal cerebral
ischemia: histopathological consequences of middle cerebral artery occlusion in the rat. J Cereb Blood
Flow Metab 1992;12:380-389.
169. Lin TN, He YY, Wu G, Khan M, Hsu CY. Eﬀect of brain edema on infarct volume in a focal cerebral
ischemia model in rats. Stroke 1993;24:117-121.
170. Frank JI. Large hemispheric infarction, deterioration, and intracranial pressure. Neurology
1995;45:1286-1290.
171. Ropper AH, Shafran B. Brain edema after stroke. Clinical syndrome and intracranial pressure. Arch
Neurol 1984;41:26-29.
172. Wardlaw J, Dennis M, Lindley R, Warlow C, Sandercock P, Sellar R. Does early reperfusion of a cerebral
infarct inﬂuence cerebral infarct swelling in the acute stage or the ﬁnal clinical outcome? Cerebrovasc
Dis 1993;3:86-93.
173. Schwab S, Aschoﬀ A, Spranger M, Albert F, Hacke W. The value of intracranial pressure monitoring
in acute hemispheric stroke. Neurology 1996;47:393-398.
174. Ziai WC, Mirski MA, Bhardwaj A. Use of hypertonic saline in ischemic stroke. Stroke 2002;33:11661167.
175. Steiner T, Ringleb P, Hacke W. Treatment options for large hemispheric stroke. Neurology 2001;57:
S61-S68.
176. Kasner SE, Demchuk AM, Berrouschot J. Predictors of fatal brain edema in massive hemispheric
ischemic stroke. Stroke 2001;32:2117-2123.
177. Krieger DW, Demchuk AM, Kasner SE, Jauss M, Hantson L. Early clinical and radiological predictors
of fatal brain swelling in ischemic stroke. Stroke 1999;30:287-292.
178. Oppenheim C, Samson Y, Manai R. Prediction of malignant middle cerebral artery infarction by
diﬀusion-weighted imaging. Stroke 2000;31:2175-2181.
179. Neumann-Haefelin T, Sitzer M, du MdR, Lanfermann H. Prediction of malignant MCA infarction
with DWI: pitfalls in hyperacute stroke. Stroke 2001;32:580-583.
180. Barber PA, Davis SM, Darby DG. Absent middle cerebral artery ﬂow predicts the presence and
evolution of the ischemic penumbra. Neurology 1999;52:1125-1132.
181. Neumann-Haefelin T, Moseley ME, Albers GW. New magnetic resonance imaging methods for
cerebrovascular disease: emerging clinical applications. Ann Neurol 2000;47:559-570.
182. de Haan RJ, Limburg M, Van der Meulen JH, Jacobs HM, Aaronson NK. Quality of life after stroke.
Impact of stroke type and lesion location. Stroke 1995;26:402-408.
183. Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Macleod MR, Kappelle LJ, van der Worp HB; the
HAMLET investigators. Hemicraniectomy After Middle cerebral artery infarction with Lifethreatening Edema Trial (HAMLET). Protocol for a randomised controlled trial of decompressive
surgery in space-occupying hemispheric infarction. Trials 2006;7:29.
39
Chapter 2
40
Rat model of space-occupying infarction
Chapter 3
The time course of ischemic damage and
perfusion in a rat model of space-occupying
cerebral infarction
Jeannette Hofmeijer, Wouter B. Veldhuis, Janneke Schepers, Klaas Nicolay, L. Jaap Kappelle, Peter R. Bär,
and H. Bart van der Worp
Based on Brain Research 2004;1013:74-82
41
Chapter 3
Summary
In this chapter a rat model of space-occupying hemispheric infarction is described.
For adequate timing of therapy in future experiments, the development of tissue
damage, edema formation, and perfusion over time were studied with diﬀerent MRI
techniques.
Permanent middle cerebral artery (MCA) occlusion was performed in 32 Fisher344 rats. Forty-six MRI experiments including diﬀusion weighted (DWI), T2-weighted
(T2W), ﬂow-sensitive alternating inversion recovery (FAIR) perfusion-weighted, and
T1-weighted (T1W) imaging before and after gadolinium were performed at 1, 3, 8, 16,
24, and 48 hours of ischemia.
MCA occlusion consistently led to infarction of the complete MCA territory.
Mortality was 75% at 48 hours after occlusion. Lesion volumes as derived from apparent
diﬀusion coeﬃcient- and T2-maps increased to maximum values of 400 ± 48mm3 at
24 hours and 420 ± 54mm3 at 48 hours of ischemia. Midline shift peaked at 24 hours.
The area with diﬀusion-perfusion deﬁcit decreased to a minimum at 24 hours after the
onset of ischemia. Perfusion of the contralateral hemisphere dropped at the same time
point. Leakage of gadolinium through the blood-brain barrier occurred within 3 hours
of ischemia within the entire infarct.
Permanent intraluminal MCA occlusion in Fisher-344 rats is an adequate model
for space-occupying cerebral infarction. Rats may beneﬁt from intervention aimed at
reducing tissue shift and intracranial pressure, and at improving cerebral blood ﬂow,
if initiated before 24 hours after MCA occlusion. The value of treatment modalities
depending on an intact blood brain barrier should be questioned.
42
Rat model of space-occupying infarction
P
atients with a hemispheric infarct and massive brain edema formation have a
poor prognosis: in prospective intensive care (IC)-based series, the case fatality
rate was about 80%, despite maximal medical therapy.1,2 Fatal space-occupying
edema occurs in 1 to 5% of the patients with a supratentorial infarct.3 Several treatment strategies have been proposed to limit brain tissue shifts and to reduce intracranial pressure, such as sedation, hyperventilation, and osmotic therapy, but none
has been proven eﬀective.4-6 Because of these limitations, there have been proposals
for decompressive surgery in patients with large hemispheric infarcts who deteriorate as a result of edema. This intervention leads to reversal of brain tissue shifts and
normalization of intracranial pressure (ICP). Non-randomized studies have suggested
that hemicraniectomy lowers mortality without increasing the rate of severely disabled survivors.7,8
Support for the eﬃcacy of diﬀerent treatment modalities has been sought in rat
models of focal cerebral ischemia. In these models, decompressive surgery has been
shown to reduce mortality and to improve histological and functional outcome.9-12
However, previous studies on the temporal evolution of cerebral infarction in rats
have been performed with models of variable or moderate focal cerebral ischemia.13-17
Moreover, the optimal timing of therapy has remained uncertain. An adequate model
of space-occupying infarction may provide information on the applicability and
optimal timing of medical or surgical intervention.
We present a model of space-occupying hemispheric infarction in rats and
investigated the temporal evolution of lesion volume, edema formation, perfusion,
and blood-brain barrier breakdown. We used permanent intraluminal occlusion of the
middle cerebral artery (MCA) in Fisher rats, because in Fisher rats, MCA occlusion
leads to larger and more reproducible infarcts than in Wistar or Sprague-Dawley rats.18
The MRI protocol, consisting of diﬀusion- (DWI), T2- (T2W), post-contrast T1- (T1W),
and non-invasive perfusion weighted imaging (PWI) with ﬂow-sensitive alternating
inversion recovery (FAIR) was performed at six diﬀerent time points in the ﬁrst 48
hours after MCA occlusion.
Methods
Animal model
The experiments were performed according to a protocol approved by the Utrecht
University Animal Experiment Ethical Committee. Animals had free access to standard
laboratory chow and water.
Thirty-two male Fisher rats (F344, Iﬀa-Credo Broekman, Someren, the Netherlands),
weighing 280-320g, were anesthetized by intraperitoneal injection of 0.16mg / kg fentanyl
citrate, 5mg / kg ﬂuanisone, and 2.5mg / kg midazolam, followed by subcutaneous
injection of 0.05mg / kg atropine sulfate. They were intubated and mechanically
43
Chapter 3
ventilated (Amsterdam Infant Ventilator, MK3, Hoekloos, the Netherlands) with
a gas mixture of 35% oxygen and 65% nitrous oxide. During all experiments, rectal
temperature was continuously monitored and maintained between 36.5 and 37.5°C
by means of a water heating pad. The left femoral artery was cannulated to monitor
arterial blood pressure (Datascope 3000 Monitor, Datascope corp), and to obtain blood
for arterial blood gas analyses before and after MCA occlusion (ABL 505 / OSM3,
Radiometer, Copenhagen, Denmark). When necessary, respiratory adjustments were
made to maintain normal blood gas levels. After surgery, the animals received 0.3mg /
kg buprenorﬁne subcutaneously for relief of pain.
MCA occlusion was achieved by a minor modiﬁcation of the intraluminal ﬁlament
technique, originally described by Koizumi et al.19 The right common (CCA), internal
(ICA) and external (ECA) carotid arteries were exposed through a ventral cervical
midline incision. The pterygopalatine artery was ligated with a 7.0 prolene suture.
After coagulation of the proximal branches of the ECA, the ECA and CCA were
ligated with a 7.0 silk suture. A microvascular clip was placed at the origin of the ICA.
A 3.0 prolene suture, with a rounded tip and coated with poly-L-lysine (0.1% wt/vol
in deionized water; Sigma) as described by Belayev et al.,20 was introduced into the
CCA. A 7.0 silk suture was tightened around the CCA and the intraluminal thread,
after which the microvascular clip was removed. The thread was gently advanced into
the ICA until a slight resistance was felt and the MCA was occluded. Immediately
thereafter the incision was closed.
The animals were randomly assigned to the MRI measurement time points of 1,
3, 8, and 16 hours after MCA occlusion, and to later time points if they were still
alive. For measurements at 1 and 3 hours after occlusion the rats remained under
anesthesia with 1% halothane in a mixture of 35% oxygen and 65 % nitrous oxide. For
measurements at later time points, animals were re-intubated and ventilated with 1%
halothane in the same gas mixture. Rats were immobilized in a stereotactic holder and
positioned in an animal cradle. During the MRI experiments, the exhaled CO2 level
and blood oxygen saturation were continuously monitored (Datascope Multinex 4200
and Pulse oximeter 8600V, Nonin Medical). If necessary, respiratory adjustments were
made.
Each day, rats were weighed and examined neurologically according to the scale
introduced by Bederson et al.21 and reﬁned by Menzies et al.22 Seven days after MCA
occlusion the surviving animals were killed by an intraperitoneal injection of 150 mg
pentobarbital. The brains of all animals were carefully evaluated for the presence of
subarachnoid hemorrhage.
MRI experiments
MRI experiments were performed on a 4.7T Varian (Palo Alto, USA) horizontal
bore spectrometer. RF-excitation and signal detection were accomplished by means
44
Rat model of space-occupying infarction
of a Helmholtz volume coil (diameter 9cm, length 10cm) and an inductively coupled
surface coil (diameter 2cm), respectively.
A spin-echo sequence was used for determination of the position of the animal in
the magnet (echo time (TE) / repetition time (TR) = 40ms / 1s, matrix (M) = 128 x 64,
ﬁeld of view (FOV) = 5.0 x 3.0cm2, 21 1.0mm thick sagital slices, number of excitations
(NEX) = 1). Eight contiguous 1.7mm thick transversal slices were planned with respect
to the center of the eyes. The ﬁrst slice was positioned 4mm anterior to the eyes, which
is consistent with approximately 5mm posterior to the bregma.
A single-scan diﬀusion-trace MRI sequence (4 b values: 100-1780s / mm2, TE / TR
= 100ms / ≥ 2s, M = 128 x 64, FOV = 3.2 x 3.2cm2, NEX = 2) was used to generate
quantiﬁed images of the tissue water trace apparent diﬀusion coeﬃcient (ADC). For
DWI a double spin-echo pulse sequence was used with four pairs of bipolar gradients
with speciﬁc predetermined signs in each of the three orthogonal directions.23 The
combination of gradient directions leads to a cancellation of all oﬀ-diagonal tensor
elements, eﬀectively measuring the trace of the diﬀusion tensor. This provides
unambiguous and rotationally invariant ADC values in one experiment, circumventing
the need for three separate experiments. To minimize the inherently high sensitivity of
diﬀusion-weighted imaging to motion, data acquisition was triggered to the respiratory
cycle.
T2W-images were acquired with a multi-echo sequence (TE / TR = 17ms + 7 x 17ms
/ 3s, M = 256 x 128, FOV = 3.2 x 3.2cm2, NEX = 2).
PWI was performed in a single slice through the infarct core (corresponding to
slice number ﬁve from DWI and T2W images), with the use of the FAIR technique.24
A slice-selective and a non-selective inversion recovery image (Mss and Mns) were
acquired with turbo-fast low angle shot (FLASH) acquisition and a suﬃcient inversion
time (TI) to allow inﬂow of labeled spins into brain tissue. MR parameters were as
follows: ﬂip angle (α) = 20º, TE / TR / TI = 3ms / 6ms / 2000ms, predelay = 2.0s (total
TR = 4.75s), M = 128 x 128, FOV = 3.2 x 3.2cm2, slice thickness = 1.7mm, NEX =
96. For normalization of the FAIR signal an equilibrium magnetization (M0) image
was acquired with the same parameters, but without inversion. For quantiﬁcation of
the FAIR signal a slice-selective T1-map was acquired by Turbo-FLASH Look-Locker
acquisition25 (α / TE / TR = 5º / 4.5ms / 11ms, 10 TIs (0.4 + 9 x 1.4s), M = 128 x 128,
FOV = 3.2 x 3.2cm2, slice thickness = 1.7mm, NEX = 8).
Finally, a pre-contrast T1W dataset (TE / TR = 12.5ms / 0.65s, M = 256 x 128, FOV
= 3.2 x 3.2cm2, NEX = 2) and, 12 min after injection of a bolus of 0.1ml / 100mg
Gadolinium-DTPA (Gd-DTPA) (469.01mg / ml), contrast-enhanced T1W images
were acquired.
To minimize interference at the slice boundaries, slices for multi-slice datasets were
acquired in alternating order, thus maximizing the time between excitation of two
neighboring slices.
45
Chapter 3
Data processing and analysis
All images were zero-ﬁlled to 256 x 256 prior to analysis. ADC and T2 maps were
generated by mono-exponential ﬁtting with IDL (Research Systems, Boulder, USA).
Parametric images were analyzed in anatomic regions of interest (ROIs) using in-house
software. Calculations of lesion volume were based on ipsilateral ADC or T2 diﬀerences
of more than 20% as compared to the mean value in the contralateral hemisphere. This
threshold corresponds to a diﬀerence of more than 2 standard deviations (SD) and is
close to the 23% drop in ADC found to correlate with ATP depletion 1 hour after MCA
occlusion.26
Measurements of midline shift were performed with the software package
ImageBrowser (SISCO / Varian) on slice number 5 of T2W images, according to
midline shift = ipsilateral diameter – (total diameter / 2), using the third ventricle as
a landmark.
PWI data were processed with IDL. Relative FAIR images were obtained by
subtracting the non-selective inversion recovery image from the slice selective
inversion recovery image and dividing the resulting image by the M0 image ((Mss –
Mns) / M0). T1 maps were obtained by mono-exponential ﬁtting with a correction for
magnetization saturation.27 From these relative FAIR images and T1 maps, CBF maps
in milliliters per minute per 100g tissue were calculated by T1 correction according to
Calamante et al.28 assuming perfect inversion, a homogeneous blood brain partition
coeﬃcient of 0.9ml / g and a blood T1 of 2.0s. Further analysis was performed on
region of interest (ROI) basis with ImageBrowser. ROIs in the ADC maps were deﬁned
as those with a ≥ 20% reduction of ADC as compared to the mean value calculated
from the contralateral hemisphere (ADC ROI), and in the CBF maps as those with a ≥
20% reduction of CBF as compared to contralateral (CBF ROI). Areas of the speciﬁc
ROIs as well as the mean CBF of the contralateral hemisphere were determined. The
area with diﬀusion-perfusion mismatch was calculated by subtracting the area of the
ADC ROI from the area of the CBF ROI.
Pre- and postcontrast T1W datasets were analyzed with ImageBrowser as described
by Blezer et al.29 A Gd-DTPA-enhancement ratio (GER) was calculated according to
GER = [(T1W MRI post contrast – T1W MRI precontrast) / T1W MRI precontrast] x
100. Pixel intensities thus display the percentage signal increase due to Gd leakage, and
can therefore be used as a semi-quantative measure for blood brain barrier leakage.
Mean blood brain barrier leakage was determined in the ADC ROIs.
Statistical analysis was carried out using SPSS 9.0 (SPSS Inc, Chicago, USA). Data
were evaluated by one way analysis of variance to analyze the eﬀect of time after
occlusion, or by Student’s t-test where appropriate. Post hoc tests were performed using
Bonferroni’s method. Outcome measures are expressed as means ± SD. Diﬀerences
were considered signiﬁcant at levels of P < 0.05.
46
Rat model of space-occupying infarction
Table 1 Intra-operative physiological parameters of rats at diﬀerent time points after MCA occlusion.
1h
3h
8h
16 h
24 h
48 h
n=8
n=7
n=6
n=9
n=9
n=7
Weight (g)
287 ± 21
287 ± 19
290 ± 20
289 ± 14
278 ± 17
289 ± 20
Temperature (°C)
36.4 ± 0.7
36.3 ± 0.6
36.7 ± 0.4
36.3 ± 0.5
36.8 ± 0.4
36.8 ± 0.3
BP syst (mmHg)
116 ± 6
119 ± 2
145 ± 56
121 ± 8
120 ± 6
136 ± 45
BP diast (mmHg)
91 ± 7
89 ± 9
99 ± 15
90 ± 11
95 ± 11
100 ± 12
pCO2 (mmHg)
49 ± 4
47 ± 3
49 ± 3
49 ± 4
49 ± 4
49 ± 4
O2 saturation (%)
97 ± 2
96 ± 2
98 ± 1
97 ± 2
98 ± 2
98 ± 2
BP syst indicates systolic blood pressure; BP diast, diastolic blood pressure
Results
Twenty of the 32 animals were measured once, 10 animals twice and 2 animals were
measured at 3 time points. All physiological variables remained within the normal
range throughout the operative procedures. There were no diﬀerences between the
groups for weight, body temperature, blood pressure, and arterial blood gases (Table 1).
Eight of the 32 rats (25%) survived up to seven days. All other animals (24 (75%)) died
within 48 hours after MCA occlusion (mean survival 24 ± 14 hours). On post-mortem
examination, none of the animals showed evidence of subarachnoid hemorrhage. All
surviving animals had a continuous decline in body weight up to seven days after MCA
occlusion. At day 7, body weight was 68 ± 7% of the initial weight. Directly after MCA
occlusion, all animals had a score of 4 on Bederson’s scale, indicating spontaneous
contralateral circling.21 In the surviving animals this score started to improve after 4
(n = 5) or 5 (n = 3) days.
DWI and T2W MRI
Ischemic areas were visualized on ADC and T2 maps and covered fronto-parietal
cortical and subcortical areas in all animals. Figure 1a shows a typical example of
ADC maps acquired at diﬀerent time points of a slice through the infarct core (slice
number 5). Figure 1b shows the course of lesion development on T2 maps. Lesion
volume as deduced from ADC maps gradually increased from 200 ± 65mm3 after 1
hour to a maximum of 400 ± 48mm3 at 24 hours after MCA occlusion (P < 0.001 for
the eﬀect of time on lesion volume, Figure 2a). Lesion volume on T2 maps increased
from 41 ± 25mm3 after one hour to 420 ± 54mm3 after 48 hours of ischemia (P < 0.001,
Figure 2b).
47
Chapter 3
Figure 1 ADC (a) and T2 (b) maps calculated from slice number 5 at diﬀerent
time-points after MCA occlusion.
a
b
Midline shift
Midline shift increased from 0.2 ± 0.1mm after one hour to 1.3 ± 0.3mm after 24 hours
of ischemia (P < 0.001, Figure 3).
Perfusion deﬁcit and diﬀusion-perfusion mismatch
In 12 of the 46 MRI experiments FAIR images could not be analyzed as a result of
poor signal : noise ratios due to low global CBF. In the CBF maps of the other 34
experiments, regions with perfusion deﬁcits covering the entire MCA territory were
clearly visible (Figure 4a). The mean area of the perfusion deﬁcit on slice number 5
was 64 ± 14mm2. The area of the perfusion deﬁcit did not change signiﬁcantly over
48
Rat model of space-occupying infarction
Figure 2 Lesion volume as calculated from ADC (a) and T2 (b) maps at diﬀerent time-points after
MCA occlusion. Error bars indicate mean ± standard deviation.
a
b
Figure 3
Midline shift at diﬀerent time-points after MCA
occlusion. Error bars indicate mean ± standard
deviation.
time (P = 0.1). The mean perfusion in the area of perfusion deﬁcit was 28 ± 10ml /
min / 100g, and did not change signiﬁcantly over time (P = 0.7). Since the area with
the perfusion deﬁcit remained constant over time, and the area of the lesion on ADC
maps increased (from 36 ± 11mm2 to 57 ± 16mm2), the area with diﬀusion perfusionmismatch decreased from 31 ± 12mm2 after one hour to 4 ± 5mm2 after 24 hours of
ischemia (P = 0.007, Figure 4b). Mean CBF in the contralateral hemisphere was lowest
at 24 hours after MCA occlusion: 46 ± 8ml / min / 100g compared with 74 ± 25ml /
min / 100g at other time points (P = 0.024).
49
Chapter 3
Figure 4
Example of ADC and CBF maps (a) and error bar of
the area with diﬀusion-perfusion mismatch at diﬀerent
time-points after MCA occlusion (b). Error bars indicate
mean ± standard deviation.
a
b
T1W MRI before and after Gd-DTPA
GER increased from 7 ± 2% at 1 hour to 31 ± 12% at 3 hours of ischemia (P = 0.023), and
remained essentially the same thereafter (P = 0.81). There was obvious enhancement in
cortical and subcortical infarct regions in all animals. In the contralateral hemisphere
we found no increase of Gd enhancement over time (P = 0.5).
Discussion
We studied changes in lesion volume, tissue shift, cerebral perfusion, and blood brain
barrier disturbances after permanent MCA occlusion in Fisher-344 rats. In these rats,
intraluminal occlusion of the MCA consistently led to complete, space-occupying
MCA infarction. Lesion volumes from ADC and T2 maps gradually increased up to
24 and 48 hours respectively. Brain tissue shifts peaked at 24 hours after occlusion,
whereas perfusion levels of the contralateral non-aﬀected hemisphere tended to be
lowest and regions with diﬀusion perfusion mismatch were smallest at 24 hours of
ischemia. Gadolinium leakage across the blood-brain barrier increased in the ﬁrst 3
50
Rat model of space-occupying infarction
hours of ischemia, and remained essentially the same thereafter.
In this study, permanent intraluminal MCA occlusion in Fisher-344 rats proved to be
an adequate model of space-occupying cerebral infarction. The procedure consistently
resulted in infarction of the complete MCA territory. The observed lesion volumes
of 400 ± 48mm3 on DWI at 24 hours were substantially larger than those of 161 to
225mm3 reported in previous experimental studies on space-occupying cerebral
infarction, in which other rat strains were used.9,10,12 The gradual increase in midline
shift to a maximum of 1.3 mm also reﬂects the clinical syndrome of space-occupying
infarction in man. In addition, the high mortality of 75% corresponds with mortality
levels found in clinical studies.1,2 However, the mortality rate observed in our study
may have been inﬂuenced by the MRI experiments under general anesthesia.
Our ﬁnding of consistently large infarcts in Fisher-344 rats is in line with a previous
study18 in which infarct volumes of diﬀerent rat strains were compared after MCA
occlusion according to Tamura et al.30 The smaller variability of infarct volume after
MCA occlusion in this strain is probably a result of the smaller variability of the
cerebral vasculature as compared to Wistar31 and Sprague Dawley32 rats and the lack
of proximal MCA collaterals. The consistently large infarct volumes make Fisher-344
probably the preferred strain for studies on space-occupying hemispheric infarction.
Histological studies have shown that the volume of the infarct does not grow in
size beyond 3 to 4 hours after permanent MCA occlusion.33,34 In the present model
lesion volume as calculated from the ADC maps reached its maximum at 24 hours
after MCA occlusion. Whereas DWI may suﬀer from contamination of for instance T2
contrast, contrast on the calculated ADC maps is brought about purely by diﬀerences
in the apparent diﬀusion coeﬃcient of water. The ischemia-induced ADC reduction
is thought to be related to cytotoxic edema formation and the associated decrease in
extracellular space, as a result of a malfunctioning of energy-dependent ion pumps.35
Studies comparing MRI with histopathological data found that a marked ADC
reduction after 2 hours of ischemia is related to irreversible tissue damage.35 Several
studies of permanent MCA occlusion in rats have reported a gradual extension of the
area of DWI hyperintensity.13-17 The fast growth of lesion volume during the ﬁrst hours
is in line with previous reports.13,36
Lesion size on T2W images increased up to 48 hours after occlusion and, at this time
point, exceeded the size of the DWI abnormalities. Prolongation of brain tissue water
T2, measured by T2W MRI, is thought to reﬂect vasogenic edema and is probably most
pronounced in irreversibly damaged tissue.37-41 It has been demonstrated that the size
of the infarct does not increase any more after 24 hours of ischemia, so infarct volume
on T2W14 and DWI42 24 hours after occlusion can be considered as the ﬁnal lesion
size. The (non-signiﬁcant) increase in T2 abnormality between 24 and 48 hours of
ischemia therefore may reﬂect an increase in brain edema. This is in line with other
ﬁndings suggesting overestimation of infarct volume as a result of the development of
51
Chapter 3
brain edema, especially in the ﬁrst 3 days of ischemia.43,44 Edema formation and the
consequent tissue shifts and ICP increase may compromise CBF in other vascular
territories and thereby increase ischemic damage.
We used FAIR for PWI. FAIR has been shown to quantify normal CBF values24 and
decreases in CBF in a model of forebrain ischemia in gerbils45 as well as in patients with
various cerebral diseases.46,47 In contrast to continuous arterial spin labeling, FAIR had
not been applied in experimental focal cerebral ischemia before.
Analysis of the CBF maps revealed a region with a clear perfusion deﬁcit covering
the entire MCA territory on all 34 measurement points. The area with the perfusion
deﬁcit remained constant over time, indicating permanent complete MCA occlusion.
Since lesion volume increased on ADC maps, the area with diﬀusion-perfusion
mismatch decreased up to 24 hours of ischemia. At 24 hours of ischemia, when midline
shift reached a maximum, perfusion levels of the initially non-injured contralateral
hemisphere were lowest. In patients with space-occupying cerebral infarction,
tissue shifts may lead to compression of other arteries and thereby to additional
ischemic damage.48 Despite the substantial midline shift and impaired perfusion of
the contralateral hemisphere, we did not ﬁnd evidence of ischemic damage beyond
the ipsilateral MCA territory at 48 hours in this model. The increase of the area with
diﬀusion-perfusion mismatch at 48 hours is caused by the decrease of lesion volume
calculated from ADC maps. Since some animals died before the MRI measurement
points of 24 and 48 hours, it is possible that the rats that were studied at these time
points had less severe ischemic damage. This may partially explain the stabilization
of lesion volume after 24 hours. However, these animals also had complete spaceoccupying MCA infarction and a severe neurological deﬁcit, underscoring the validity
of the model.
The gradual extension of tissue damage as observed on the ADC and T2W maps
and substantiated by the decrease in diﬀusion-perfusion mismatch is consistent
with earlier ﬁndings of time-dependent perfusion thresholds for tissue damage,49,50
and is not caused by tissue shifts or a rise in ICP. Although CBF of the contra-lateral
hemisphere is not the same in all groups of animals, we deﬁned the CBF ROI as a
percentage of the CBF in the contralateral hemisphere. We did not use absolute values,
because of the ﬂuctuations in resting CBF with this technique. The reduction of 20%
corresponds with a diﬀerence of about two SD. With the 20% reduction we did not
aim to deﬁne ischemic tissue, but tissue that is hemodynamicaly compromised, or so
called ‘penumbra’ or ‘tissue at risk’
The Gd-DTPA-enhancement ratio, a semi-quantitative measure for blood-brain
barrier disruption,29 increased up to 3 hours of ischemia and remained essentially
the same thereafter. This ﬁnding is consistent with previous reports showing that the
blood-brain barrier remains intact for only a few hours after permanent ischemia.51-53
After reperfusion, blood-brain barrier damage is probably even more pronounced.54
52
Rat model of space-occupying infarction
This ﬁnding supports theoretical concerns on the use of osmotic agents in patients
with massive space-occupying infarction. As a result of the destruction of the blood
brain barrier, osmotic agents may accumulate in the aﬀected tissue, thereby reversing
the osmotic gradient between tissue and plasma and leading to an exacerbation of
edema.55 Moreover, as osmotic agents are supposed to exert their main eﬀect in the
contralateral hemisphere where the blood brain barrier is still intact, the induced
reduction in volume of this hemisphere could lead to an additional increase in brain
tissue shifts.48
In conclusion, several observations in our study indicate that permanent intraluminal
MCA occlusion in Fisher-344 rats is an adequate model of space-occupying cerebral
infarction and may be superior to MCA occlusion in other rat strains. The consistently
large infarct volume, substantial midline shift, and high mortality adequately reﬂect
this clinical syndrome in man. In rats, midline shift was maximal at 24 hours, whereas
perfusion of the contralateral hemisphere was lowest at this time point. In this model,
potential therapies against ICP increase and tissue shifts should therefore be started
before 24 hours of ischemia. The usefulness of treatment modalities depending on
an intact blood-brain barrier, such as osmotherapy, should be questioned since the
integrity of the blood-brain barrier is disturbed from the ﬁrst hours of ischemia
onwards.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
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 1996;53:309-315.
Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction and early brain swelling:
progression and eﬀect of age on outcome. Mayo Clin Proc 1998;73:829-836.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
Adams HP, Jr., Adams RJ, Brott T. Guidelines for the early management of patients with ischemic stroke:
A scientiﬁc statement from the Stroke Council of the American Stroke Association. Stroke
2003;34:1056-1083.
van der Worp HB, Kappelle LJ. Complications of acute ischaemic stroke. Cerebrovasc Dis 1998;8:124132.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral infarction. Crit
Care Med 2003;31:617-625.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Doerﬂer A, Forsting M, Reith W. Decompressive craniectomy in a rat model of “malignant” cerebral
hemispheric stroke: experimental support for an aggressive therapeutic approach. J Neurosurg
1996;85:853-859.
53
Chapter 3
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
54
Forsting M, Reith W, Schabitz WR. Decompressive craniectomy for cerebral infarction. An
experimental study in rats. Stroke 1995;26:259-264.
Engelhorn T, Doerﬂer A, Kastrup A. Decompressive craniectomy, reperfusion, or a combination for
early treatment of acute “malignant” cerebral hemispheric stroke in rats? Potential mechanisms
studied by MRI. Stroke 1999;30:1456-1463.
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 2002;33:617-622.
Hoehn-Berlage M, Norris DG, Kohno K, Mies G, Leibfritz D, Hossmann KA. Evolution of regional
changes in apparent diﬀusion coeﬃcient during focal ischemia of rat brain: the relationship of
quantitative diﬀusion NMR imaging to reduction in cerebral blood ﬂow and metabolic disturbances.
J Cereb Blood Flow Metab 1995;15:1002-1011.
Quast MJ, Huang NC, Hillman GR, Kent TA. The evolution of acute stroke recorded by multimodal
magnetic resonance imaging. Magn Reson Imaging 1993;11:465-471.
Knight RA, Ordidge RJ, Helpern JA, Chopp M, Rodolosi LC, Peck D. Temporal evolution of ischemic
damage in rat brain measured by proton nuclear magnetic resonance imaging. Stroke 1991;22:802808.
Dardzinski BJ, Sotak CH, Fisher M, Hasegawa Y, Li L, Minematsu K. Apparent diﬀusion coeﬃcient
mapping of experimental focal cerebral ischemia using diﬀusion-weighted echo-planar imaging.
Magn Reson Med 1993;30:318-325.
Verheul HB, Berkelbach van der Sprenkel JW, Tulleken CA, Tamminga KS, Nicolay K. Temporal
evolution of focal cerebral ischemia in the rat assessed by T2-weighted and diﬀusion-weighted
magnetic resonance imaging. Brain Topogr 1992;5:171-176.
Duverger D, MacKenzie ET. The quantiﬁcation of cerebral infarction following focal ischemia in the
rat: inﬂuence of strain, arterial pressure, blood glucose concentration, and age. J Cereb Blood Flow
Metab 1988;8:449-461.
Koizumi J, Yoshida Y, Nakazawa T, Ooneda G. Experimental studies of ischemic brain edema, I: a new
experimental model of cerebral embolism in rats in which recirculation can be introduced in the
ischemic area. Jpn J Stroke 1986;8:1-8.
Belayev L, Alonso OF, Busto R, Zhao W, Ginsberg MD. Middle cerebral artery occlusion in the rat by
intraluminal suture. Neurological and pathological evaluation of an improved model. Stroke
1996;27:1616-1622.
Bederson JB, Pitts LH, Tsuji M, Nishimura MC, Davis RL, Bartkowski H. Rat middle cerebral artery
occlusion: evaluation of the model and development of a neurologic examination. Stroke 1986;17:472476.
Menzies SA, Hoﬀ JT, Betz AL. Middle cerebral artery occlusion in rats: a neurological and pathological
evaluation of a reproducible model. Neurosurgery 1992;31:100-106.
de Graaf RA, Braun KP, Nicolay K. Single-shot diﬀusion trace (1)H NMR spectroscopy. Magn Reson
Med 2001;45:741-748.
Kim SG. Quantiﬁcation of relative cerebral blood ﬂow change by ﬂow- sensitive alternating inversion
recovery (FAIR) technique: application to functional mapping. Magn Reson Med 1995;34:293-301.
Haase A, Matthaei D, Bartkowski R, Duhmke E, Leibfritz D. Inversion recovery snapshot FLASH MR
imaging. J Comput Assist Tomogr 1989;13:1036-1040.
Olah L, Wecker S, Hoehn M. Relation of apparent diﬀusion coeﬃcient changes and metabolic
Rat model of space-occupying infarction
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
disturbances after 1 hour of focal cerebral ischemia and at diﬀerent reperfusion phases in rats. J Cereb
Blood Flow Metab 2001;21:430-439.
Deichmann R, Haase A. Quantiﬁcation of T1 values by SNAPSHOT-FLASH NMR imaging. J Magn
Reson 1992;96:608-612.
Calamante F, Williams SR, van Bruggen N, Kwong KK, Turner R. A model for quantiﬁcation of
perfusion in pulsed labelling techniques. NMR Biomed 1996;9:79-83.
Blezer EL, Nicolay K, Goldschmeding R, Koomans HA, Joles JA. Reduction of cerebral injury in
stroke-prone spontaneously hypertensive rats by amlodipine. Eur J Pharmacol 2002;444:75-81.
Tamura A, Graham DI, McCulloch J, Teasdale GM. Focal cerebral ischaemia in the rat: 1. Description
of technique and early neuropathological consequences following middle cerebral artery occlusion. J
Cereb Blood Flow Metab 1981;1:53-60.
Herz RC, Jonker M, Verheul HB, Hillen B, Versteeg DH, De Wildt DJ. Middle cerebral artery occlusion
in Wistar and Fischer-344 rats: functional and morphological assessment of the model. J Cereb Blood
Flow Metab 1996;16:296-302.
Fox G, Gallacher D, Shevde S, Loftus J, Swayne G. Anatomic variation of the middle cerebral artery in
the Sprague-Dawley rat. Stroke 1993;24:2087-2092.
Memezawa H, Smith ML, Siesjo BK. Penumbral tissues salvaged by reperfusion following middle
cerebral artery occlusion in rats. Stroke 1992;23:552-559.
Kaplan B, Brint S, Tanabe J, Jacewicz M, Wang XJ, Pulsinelli W. Temporal thresholds for neocortical
infarction in rats subjected to reversible focal cerebral ischemia. Stroke 1991;22:1032-1039.
Moseley ME, Cohen Y, Mintorovitch J. Early detection of regional cerebral ischemia in cats:
comparison of diﬀusion- and T2-weighted MRI and spectroscopy. Magn Reson Med 1990;14:330346.
van Dorsten FA, Olah L, Schwindt W. Dynamic changes of ADC, perfusion, and NMR relaxation
parameters in transient focal ischemia of rat brain. Magn Reson Med 2002;47:97-104.
Brant-Zawadzki M, Pereira B, Weinstein P. MR imaging of acute experimental ischemia in cats. AJNR
1986;7:7-11.
Kato H, Kogure K, Ohtomo H. Characterization of experimental ischemic brain edema utilizing
proton nuclear magnetic resonance imaging. J Cereb Blood Flow Metab 1986;6:212-221.
Hoehn-Berlage M, Tolxdorﬀ T, Bockhorst K, Okada Y, Ernestus RI. In vivo NMR T2 relaxation
of experimental brain tumors in the cat: a multiparameter tissue characterization. Magn Reson
Imaging 1992;10:935-947.
Olah L, Wecker S, Hoehn M. Secondary deterioration of apparent diﬀusion coeﬃcient after 1-hour
transient focal cerebral ischemia in rats. J Cereb Blood Flow Metab 2000;20:1474-1482.
Horikawa Y, Naruse S, Tanaka C, Hirakawa K, Nishikawa H. Proton NMR relaxation times in ischemic
brain edema. Stroke 1986;17:1149-1152.
Loubinoux I, Volk A, Borredon J. Spreading of vasogenic edema and cytotoxic edema assessed by
quantitative diﬀusion and T2 magnetic resonance imaging. Stroke 1997;28:419-426.
Lin TN, He YY, Wu G, Khan M, Hsu CY. Eﬀect of brain edema on infarct volume in a focal cerebral
ischemia model in rats. Stroke 1993;24:117-121.
Gerriets T, Stolz E, Walberer M. Noninvasive quantiﬁcation of brain edema and the space-occupying
eﬀect in rat stroke models using magnetic resonance imaging. Stroke 2004;35:566-571.
Pell GS, Lythgoe MF, Thomas DL. Reperfusion in a gerbil model of forebrain ischemia using serial
magnetic resonance FAIR perfusion imaging. Stroke 1999;30:1263-1270.
55
Chapter 3
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56
Arbab AS, Aoki S, Toyama K. Brain perfusion measured by ﬂow-sensitive alternating inversion
recovery (FAIR) and dynamic susceptibility contrast-enhanced magnetic resonance imaging:
comparison with nuclear medicine technique. Eur Radiol 2002;11:635-641.
Arbab AS, Aoki S, Toyama K. Quantitative Measurement of Regional Cerebral Blood Flow with FlowSensitive Alternating Inversion Recovery Imaging: Comparison with [Iodine 123]-Iodoamphetamin
Single Photon Emission CT. AJNR 2002;23:381-388.
Frank JI. Large hemispheric infarction, deterioration, and intracranial pressure. Neurology
1995;45:1286-1290.
Dijkhuizen RM, Berkelbach van der Sprenkel JW, Tulleken KA, Nicolay K. Regional assessment of
tissue oxygenation and the temporal evolution of hemodynamic parameters and water diﬀusion
during acute focal ischemia in rat brain. Brain Res 1997;750:161-170.
Hossmann KA. Viability thresholds and the penumbra of focal ischemia. Ann Neurol 1994;36:557565.
Menzies SA, Betz AL, Hoﬀ JT. Contributions of ions and albumin to the formation and resolution of
ischemic brain edema. J Neurosurg 1993;78:257-266.
Gotoh O, Asano T, Koide T, Takakura K. Ischemic brain edema following occlusion of the middle
cerebral artery in the rat. I: The time courses of the brain water, sodium and potassium contents and
blood-brain barrier permeability to 125I-albumin. Stroke 1985;16:101-109.
Olsson Y, Crowell RM, Klatzo I. The blood-brain barrier to protein tracers in focal cerebral ischemia
and infarction caused by occlusion of the middle cerebral artery. Acta Neuropathol (Berl) 1971;18:89102.
Cole DJ, Matsumura JS, Drummond JC, Schultz RL, Wong MH. Time- and pressure-dependent
changes in blood-brain barrier permeability after temporary middle cerebral artery occlusion in rats.
Acta Neuropathol (Berl) 1991;82:266-273.
Kaufmann AM, Cardoso ER. Aggravation of vasogenic cerebral edema by multiple-dose mannitol. J
Neurosurg 1992;77:584-589.
Perfusion in rats with MCA occlusion
Chapter 4
Perfusion MRI by ﬂow-sensitive alternating
inversion recovery and dynamic susceptibility
contrast in rats with permanent middle cerebral
artery occlusion
Jeannette Hofmeijer, Janneke Schepers, H. Bart van der Worp, L. Jaap Kappelle, and Klaas Nicolay
Based on NMR in Biomedicine 2005;18:390-394
57
Chapter 4
Summary
In this chapter cerebral blood ﬂow (CBF) parameters obtained by dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) are compared with those
obtained by ﬂow-sensitive alternating inversion recovery (FAIR) in brain regions
with diﬀerent perfusion levels in rats with permanent middle cerebral artery (MCA)
occlusion.
MCA occlusion was performed in 19 rats. T2-weighted MRI, FAIR, and DSC-MRI
were performed within 48 hours after occlusion. CBF parameters were analyzed in
regions of interest with either prolonged or less prolonged mean transit time (MTT).
Ratios of ipsi- versus contralateral CBF values were calculated and tested for correlation
and diﬀerences between FAIR and DSC-MRI.
FAIR-aCBF ratios correlated signiﬁcantly with DSC-rCBF ratios (r2 = 0.719, P <
0.001). Mean FAIR-aCBF ratio was signiﬁcantly lower than mean DSC-rCBF ratio in
the area with prolonged MTT (0.39 ± 0.14 and 0.50 ± 0.14 respectively, P = 0.008).
In the area with less prolonged MTT, mean FAIR-aCBF ratio and mean DSC-rCBF
values did not diﬀer signiﬁcantly (0.78 ± 0.16 and 0.80 ± 0.19 respectively, P = 0.7).
It is concluded that FAIR correlates with DSC-MRI if perfusion is preserved.
FAIR provides lower CBF values than DSC-MRI if perfusion is reduced and MTT
is prolonged. This probable underestimation of perfusion may be caused by transit
delays. Care should be taken when quantifying CBF with FAIR and when comparing
the results of FAIR-and DSC-MRI in areas with hypoperfusion.
58
Perfusion in rats with MCA occlusion
M
easurement of perfusion may guide therapeutic decisions in cerebral ischemia.1
Magnetic resonance imaging (MRI) allows perfusion-weighted imaging
(PWI) by the use of exogenous contrast agents, and by endogenous contrast
mechanisms. PWI by dynamic susceptibility contrast (DSC)-MRI depends on signal
changes induced by the passage of a bolus of injected paramagnetic agent, such as a
gadolinium chelate. Maps of hemodynamic parameters, including cerebral blood ﬂow
(CBF), cerebral blood volume (CBV), and mean transit time (MTT) can be calculated.2
Hemodynamic parameters can be quantiﬁed only if an arterial input function can be
measured.3
Arterial spin labeling (ASL) is an endogenous contrast method, based on the signal
diﬀerence detected between magnetically labeled and unlabeled spins in two imaging
experiments. ASL is increasingly used for CBF measurements in experimental stroke
research.4-6 ASL has been validated for PWI in healthy rat brain7 and after transient
ischemia,8 but not in permanent cerebral ischemia.
We compared CBF parameters obtained with Flow-sensitive Alternating Inversion
Recovery (FAIR), a pulsed ASL technique, with those obtained with DSC-MRI in brain
regions with diﬀerent levels of hypoperfusion in rats with permanent middle cerebral
artery (MCA) occlusion.
Methods
Animal model
The protocol was approved by the Utrecht University Animal Experiment Ethical
Committee. Animals had free access to food and water. Nineteen male Fisher rats
(F344/Ico, Iﬀa-Credo Broekman, Someren, the Netherlands), weighing 280-320g,
were anesthetized by intraperitoneal injection of 0.16 mg/kg fentanyl citrate, 5 mg/
kg ﬂuanisone, and 2.5 mg/kg midazolam, followed by subcutaneous injection of 0.05
mg/kg atropine sulfate. They were intubated and mechanically ventilated (Amsterdam
Infant Ventilator, MK3, Hoekloos, the Netherlands) with 35% oxygen and 65% nitrous
oxide. During all experiments, rectal temperature was maintained between 36.5 and
37.5°C by means of a water heating pad, and the exhaled CO2 concentration was
monitored. The left femoral artery was cannulated to monitor arterial blood pressure
(Datascope 3000 Monitor, Datascope corp), and to obtain blood for arterial blood gas
analyses before and after MCA-occlusion (ABL 505/ OSM3, Radiometer, Copenhagen,
Denmark). When necessary, respiratory adjustments were made. After surgery, the
animals received 0.3 mg/kg buprenorﬁne subcutaneously for relief of pain.
MCA-occlusion was achieved by a minor modiﬁcation of the intraluminal ﬁlament
technique.9 In short, the right common (CCA), internal (ICA), and external (ECA)
carotid arteries were exposed through a ventral cervical midline incision. The
pterygopalatine artery, ECA and CCA were ligated with a 7.0 silk suture. A 3.0 prolene
59
Chapter 4
suture, with a rounded tip and coated with poly-L-lysine (0.1% wt/vol in deionized
water; Sigma)10 was introduced into the CCA and advanced into the ICA until a
slight resistance was felt. A 7.0 silk suture was tightened around the CCA and the
intraluminal thread. Thereafter the incision was closed.
Twenty-three MRI experiments were performed within 48 hours after MCA
occlusion. The animals were re-induced by halothane inhalation followed by
intubation. Anesthesia was maintained with 1.0% halothane in a 35% oxygen and 65%
nitrous oxide mixture. During MRI experiments the animals were immobilized in an
animal cradle using incisor and ear bars. A tail vein was cannulated for injection of
gadolinium diethylene triamine pentaacetic acid (Gd-DTPA, Magnevist, Schering,
Berlin, Germany).
MR experiments
Experiments were performed on a 4.7 T Varian INOVA system (Palo Alto, USA).
RF excitation and signal detection were accomplished by a Helmholtz volume coil
(diameter 9cm, length 10cm) and an inductively coupled surface coil (diameter 2cm),
respectively. A spin-echo sequence was used for determination of the position of the
animal.
Single coronal slice FAIR, T1-weighted and DSC-MRI images were acquired
approximately 5mm posterior to bregma. For FAIR PWI a slice-selective and a nonselective inversion recovery image (Mss and Mns) were acquired with turbo-fast low
angle shot (FLASH) acquisition and a suﬃcient inversion time (TI) to allow inﬂow of
labeled spins into brain tissue.8 MRI parameters were as follows: ﬂip angle (α) = 20º,
echo time (TE) / repetition time (TR) / TI = 3ms / 6ms / 2000ms, pre-delay = 2.0s (total
TR = 4.75s), resolution = 250 x 250um2, slice thickness = 2mm, number of excitations
(NEX) = 96. For normalization of the FAIR signal an equilibrium magnetization (M0)
image was acquired with the same parameters, without inversion. For quantiﬁcation
of the FAIR signal a slice-selective and non selective T1-map were acquired by TurboFLASH Look-Locker acquisition11 (α / TE / TR = 5º / 3ms / 11ms, 10 TIs (0.4 + 9 x 1.4s),
resolution = 250 x 250um2, slice thickness = 2mm, NEX = 8). The FAIR experiment
took 15 minutes.
For DSC-MRI (acquired immediately after FAIR PWI) FLASH acquisition (α / TE /
TR = 12º / 7ms / 11ms, resolution = 500 x 500um2, slice thickness = 2 mm, and NEX
= 1) was used following injection of 0.5 mmol / kg Gd-DTPA as a bolus of 0.1 mg /
100mg, injected after twelve images in less than 0.7s, with a temporal resolution of
0.7s per image.
Generation of hemodynamic parameter maps
All images were zero-ﬁlled to 256 x 256. T1 maps were generated by mono-exponential
ﬁtting on a pixel-by-pixel basis, and corrected for longitudinal saturation as described
60
Perfusion in rats with MCA occlusion
by Deichmann et al.12 Relative FAIR images (FAIR-rCBF in %) were obtained by
subtracting the non-selective inversion recovery image from the slice selective
inversion recovery image and dividing the result by the M0 image ((Mss–Mns) / M0).
From the relative FAIR images and T1 maps, CBF maps in milliliters per minute per 100g
tissue were calculated by T1 correction according to Calamante et al.,13 assuming perfect
inversion, a homogeneous blood-brain partition coeﬃcient of 0.9ml / g, and an arterial
blood T1 of 2.0s. The resulting FAIR-CBF values will be referred to as FAIR apparent
CBF values (FAIR-aCBF in ml / min / 100g tissue), to emphasize that no correction
was made for transit delays and that quantiﬁcation may thus not be absolute.
A gamma-variate ﬁt was used to derive hemodynamic parameters from DSC-MRI
images, yielding maps of relative CBF (DSC-rCBF), relative CBV, and relative MTT
in arbitrary units. The annotation “relative” indicates that no absolute quantiﬁcation
was performed by deconvolution with an arterial input function. T1 and FAIR image
calculations were performed in IDL (Interactive Data Language, Research Systems,
Boulder, CO). The gamma-variate ﬁts of DSC-MRI data were ﬁtted as described in
Kluytmans et al. using in-house software.14
Further analysis was performed on region of interest (ROI) basis with the software
package ImageBrowser (Varian). ROIs were deﬁned using an MTT map. Two ROIs were
drawn in the ipsilateral hemisphere. The ﬁrst ROI was drawn according to MTT (ROI)
> 1.5 x mean MTT (contralateral), and will be referred to as the ‘ROI with prolonged
MTT’. The second ROI included the rest of the hemisphere, and will be referred to as
the ‘ROI with less prolonged MTT’. Parts of the hemisphere in which perfusion was
too low to ﬁt were excluded. Large cerebral venous structures were excluded using the
CBV map, according to CBV > mean CBV (contralateral) + 2 standard deviations (SD).
Ipsilateral ROIs were mirrored in the mid-sagittal line to obtain contralateral ROIs.
Ratios of ipsi- versus contralateral CBF values were calculated for these ROIs.
Statistical analysis
Data are expressed as mean ± SD. For analysis, data from measurements at diﬀerent
time-points were pooled. The diﬀerence between the data obtained from the two
perfusion MRI techniques was tested for signiﬁcance using a Student’s t-test. Linear
(Pearson’s) correlation coeﬃcients between FAIR and DSC-CBF parameters were
calculated. Correlation and diﬀerences were considered signiﬁcant if P < 0.05.
Results
Intraoperative physiological variables remained within the normal range in all animals
(data not shown). Ischemic areas were visualized on T2, MTT, DSC- and FAIR-perfusion
maps, and covered the complete MCA territory in all animals. The development of
tissue damage and edema formation in these rats has been described in a separate
61
Chapter 4
article.15 In the lesion core signal-to-noise ratio (SNR) was too low to determine MTT.
This area was excluded from analysis in all rats. Regions with prolonged MTT values
included the watershed areas between the MCA and anterior cerebral artery (ACA).
Regions with less prolonged MTT were found in the ACA territory (Figure 1).
FAIR-aCBF ipsi- versus contralateral CBF ratios were signiﬁcantly correlated
with DSC-rCBF ratios (r2 = 0.719, P < 0.001; Figure 2). The mean MTT ipsi- versus
contralateral ratio was 2.0 ± 0.4 in ROIs with prolonged MTT and 1.0 ± 0.1 in ROIs
with less prolonged MTT (P < 0.001). In the area with prolonged MTT, mean FAIRaCBF ratio was signiﬁcantly lower than mean DSC-rCBF ratio (0.39 ± 0.14 and 0.50 ±
0.14 respectively, P = 0.008). In the area with less prolonged MTT, mean FAIR-aCBF
ratio and mean DSC-rCBF values did not diﬀer signiﬁcantly (0.78 ± 0.16 and 0.80 ±
0.19 respectively, P = 0.7; Figure 3).
Figure 1 Examples of maps of (a) local anatomy with regions of interest, (b) FAIR-aCBF, (c) DSC-rCBF, and
(d) MTT.
a
b
c
d
In Figure 1a, the infarct core is colored in black (excluded from analysis), the ROI with prolonged MTT
(watershed area) in dark grey, and the ROI with less prolonged MTT (ACA area) in light grey.
Figure 2 Scatter plot of ipsiversus contralateral FAIR-aCBF
as a function of ipsi- versus
contralateral DSC-rCBF per rat
(each dot).
62
Perfusion in rats with MCA occlusion
Figure 3 Ipsi- versus contralateral ratios for FAIR-aCBF and DSC-rCBF in areas with
diﬀerent MTT values.
Error bars indicate means and standard deviations; *, a statistically signiﬁcant
diﬀerence (P < 0.05).
Discussion
In this study, both FAIR and DSC-MRI were sensitive to spatial diﬀerences in CBF
in rats with permanent MCA occlusion. A signiﬁcant linear correlation was found
between the ratios of FAIR values and relative DSC-CBF indices. However, in regions
with prolonged MTT, FAIR ipsi- versus contralateral CBF ratios were lower than
DSC-MRI ratios. This indicates that FAIR detected larger CBF reductions than DSCrCBF in those regions.
Underestimation of CBF with ASL in hypoperfused brain regions has been reported
before.16,17 This is generally attributed to transit delays, which are longer in areas with
reduced perfusion. These transit delays cause loss of label in ASL-based imaging, and
may lead to a lower measured CBF.18 We found prolonged MTT values mainly in
watershed areas. In permanent vessel occlusion, transit delays in watershed areas are
probably longer, since blood supply is dependent on collateral pathways.
Among the ASL methods, FAIR is reported to suﬀer only moderately from transit
delays, because the distance between the site of labeling and the acquisition slice is
relatively small. However, in regions with reduced perfusion, especially in regions
supplied by collateral ﬂow, transit delays may be signiﬁcant even for FAIR imaging.
They may lead to a reduction of measured CBF of up to 20% with transit delays of
500 ms.18 DSC-MRI may be considered as standard MRI modality in clinical practice,
and in diﬀerent animal models. Our study demonstrates that (semi-) quantitative
63
Chapter 4
ASL results cannot always be compared with qualitative or quantitative DSC-MRI
parameters, especially in areas with low CBF.19
A reduction of the gap between the site of labeling and the acquisition slice may
improve FAIR PWI, even in regions with CBF reductions.20 This would also improve
SNR, which is known to be low for ASL-based PWI.20,21 Alternatively, transit
delays may be ﬁtted by analysis of FAIR signal acquired with multiple TI values.
However, acquisitions with multiple TI values are time consuming. In addition, ASL
implementations that compensate intrinsically for transit delays by means of saturation
pulses may be used.22-24 These sequences set the transit delay, such that it is a known
parameter in the quantiﬁcation of perfusion. However, with very long transit delays (as
can be found in severe stroke), these methods are less applicable.
Still, the beneﬁts of FAIR-based PWI may outweigh the above-mentioned
disadvantages for most experimental purposes. Quantiﬁcation of perfusion with DSCMRI is probably equally diﬃcult, since determination of an arterial input function is
problematic in small rodents, given the small diameter of the feeding arteries. FAIR
images can be acquired with a better spatial resolution, since spatial resolution is
restricted by the need for a good temporal resolution in bolus passage sampling, such
as DSC-MRI. Moreover, ASL based imaging is non-invasive and has the potential of
unlimited repeats without delays.
The role of ASL in clinical stroke care is still limited, because of its low sensitivity
in regions with reduced CBF, and the relatively long duration of scanning. However,
in follow-up studies of stroke patients, ASL may be of use, especially if repeated
measurements are necessary. FAIR has been used to measure reserve capacity of
cerebral auto-regulation after acetazolamide challenge, which would have been
diﬃcult with injections of contrast agent.25 Continuous ASL oﬀers the possibility to
measure perfusion territories using a single labeling coil placed on one of the carotid
arteries.26 Transit delays may even be used to detect hemodynamic compromise in an
early stage.14 Moreover, ASL may be used to detect activation patterns in functional
MRI.27
Our results emphasize that FAIR correlates with the commonly used DSC-MRI if
perfusion is relatively preserved. Care should be taken when quantifying CBF with
FAIR in regions with reduced perfusion, and further research is required to determine
the nature of the FAIR signal under diﬀerent levels of hypoperfusion. The use of ASL
sequences that limit the eﬀect of transit delays is recommended.
References
1.
2.
64
Wu O, Koroshetz WJ, Ostergaard L. Predicting tissue outcome in acute human cerebral ischemia
using combined diﬀusion- and perfusion-weighted MR imaging. Stroke 2001;32:933-942.
Rosen BR, Belliveau JW, Vevea JM, Brady TJ. Perfusion imaging with NMR contrast agents. Magn
Reson Med 1990;14:249-265.
Perfusion in rats with MCA occlusion
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Rempp KA, Brix G, Wenz F, Becker CR, Guckel F, Lorenz WJ. Quantiﬁcation of regional cerebral blood
ﬂow and volume with dynamic susceptibility contrast-enhanced MR imaging. Radiology 1994;193:637-641.
Olah L, Wecker S, Hoehn M. Secondary deterioration of apparent diﬀusion coeﬃcient after 1-hour
transient focal cerebral ischemia in rats. J Cereb Blood Flow Metab 2000;20:1474-1482.
Hilger T, Niessen F, Diedenhofen M, Hossmann KA, Hoehn M. Magnetic resonance angiography of
thromboembolic stroke in rats: indicator of recanalization probability and tissue survival after
recombinant tissue plasminogen activator treatment. J Cereb Blood Flow Metab 2002;22:652-662.
Hofmeijer J, Schepers J, Veldhuis WB, Nicolay K, Kappelle LJ, Bar PR, van der Worp HB. Delayed
decompressive surgery increases apparent diﬀusion coeﬃcient and improves peri-infarct perfusion in
rats with space-occupying cerebral infarction. Stroke 2004;35:1476-1481.
Calamante F, Lythgoe MF, Pell GS. Early changes in water diﬀusion, perfusion, T1, and T2 during focal
cerebral ischemia in the rat studied at 8.5 T. Magn Reson Med 1999;41:479-485.
Schepers J, Veldhuis WB, Pauw RJ, de Groot JW, van Osch MJP, Nicolay K, van der Sanden BPJ.
Comparison of FAIR perfusion kinetics with DSC-MRI and functional histoloy, in a model of transient
ischemia. Mag Reson Med 2004;51:312-320.
Koizumi J, Yoshida Y, Nakazawa T, Ooneda G. Experimental studies of ischemic brain edema, I: a new
experimental model of cerebral embolism in rats in which recirculation can be introduced in the
ischemic area. Jpn J Stroke 1986;8:1-8.
Belayev L, Alonso OF, Busto R, Zhao W, Ginsberg MD. Middle cerebral artery occlusion in the rat by
intraluminal suture. Neurological and pathological evaluation of an improved model. Stroke
1996;27:1616-1622.
Haase A, Matthaei D, Bartkowski R, Duhmke E, Leibfritz D. Inversion recovery snapshot FLASH MR
imaging. J Comput Assist Tomogr 1989;13:1036-1040.
Deichmann R, Haase A. Quantiﬁcation of T1 values by SNAPSHOT-FLASH NMR imaging. J Magn
Reson 1992;96:608-612.
Calamante F, Williams SR, van Bruggen N, Kwong KK, Turner R. A model for quantiﬁcation of
perfusion in pulsed labelling techniques. NMR Biomed 1996;9:79-83.
Kluytmans M, van der Grond J, Viergever MA. Gray matter and white matter perfusion imaging in
patients with severe carotid artery lesions. Radiology 1998;209:675-682.
Hofmeijer J, Veldhuis WB, Schepers J, Nicolay K, Kappelle LJ, Bar PR, van der Worp HB. The time
course of ischemic damage and cerebral perfusion in a rat model of space-occupying cerebral
infarction. Brain Res 2004;1013:74-82.
Hunsche S, Sauner D, Schreiber WG, Oelkers P, Stoeter P. FAIR and dynamic susceptibility contrastenhanced perfusion imaging in healthy subjects and stroke patients. J Magn Reson Imaging
2002;16:137-146.
Yoneda K, Harada M, Morita N, Nishitani H, Uno M, Matsuda T. Comparison of FAIR technique with
diﬀerent inversion times and post contrast dynamic perfusion MRI in chronic occlusive cerebrovascular
disease. Magn Reson Imaging 2003;21:701-705.
Buxton RB, Frank LR, Wong EC, Siewert B, Warach S, Edelman RR. A general kinetic model for
quantitative perfusion imaging with arterial spin labeling. Magn Reson Med 1998;40:383-396.
Perthen JE, Calamante F, Gadian DG, Connelly A. Is quantiﬁcation of bolus tracking MRI reliable
without deconvolution? Magn Reson Med 2002;47:61-67.
Sidaros K, Andersen IK, Gesmar H, Rostrup E, Larsson HB. Improved perfusion quantiﬁcation in
FAIR imaging by oﬀset correction. Magn Reson Med 2001;46:193-197.
65
Chapter 4
21.
22.
23.
24.
25.
26.
27.
66
Schepers J, Garwood M, van der SB, Nicolay K. Improved subtraction by adiabatic FAIR perfusion
imaging. Magn Reson Med 2002;47:330-336.
Wong EC, Buxton RB, Frank LR. Quantitative imaging of perfusion using a single subtraction (QUIPSS
and QUIPSS II). Magn Reson Med 1998;39:702-708.
Wong EC, Buxton RB, Frank LR. Implementation of quantitative perfusion imaging techniques for
functional brain mapping using pulsed arterial spin labeling. NMR Biomed 1997;10:237-249.
Luh WM, Wong EC, Bandettini PA, Hyde JS. QUIPSS II with thin-slice TI1 periodic saturation: a
method for improving accuracy of quantitative perfusion imaging using pulsed arterial spin labeling.
Magn Reson Med 1999;41:1246-1254.
Yen YF, Field AS, Martin EM. Test-retest reproducibility of quantitative CBF measurements using
FAIR perfusion MRI and acetazolamide challenge. Magn Reson Med 2002;47:921-928.
Zaharchuk G, Ledden PJ, Kwong KK, Reese TG, Rosen BR, Wald LL. Multislice perfusion and
perfusion territory imaging in humans with separate label and image coils. Magn Reson Med
1999;41:1093-1098.
Detre JA, Wang J. Technical aspects and utility of fMRI using BOLD and ASL. Clin Neurophysiol
2002;113:621-634.
Decompressive surgery in rats with space-occupying infarction
Chapter 5
Delayed decompressive surgery improves
peri-infarct perfusion in rats with
space-occupying cerebral infarction
Jeannette Hofmeijer, Janneke Schepers, Wouter B. Veldhuis, Klaas Nicolay, L. Jaap Kappelle, Peter R. Bär,
and H. Bart van der Worp
Based on Stroke 2004;35:1476-81
67
Chapter 5
Summary
There is no conclusive support that decompressive surgery in late stages of spaceoccupying cerebral infarction will improve outcome. In this chapter the eﬀects of
delayed decompressive surgery on the development of tissue damage, edema formation,
and cerebral perfusion are studied with diﬀerent magnetic resonance imaging (MRI)
techniques in a rat model of space-occupying cerebral infarction.
Permanent middle cerebral artery (MCA) occlusion was performed in 6 Fisher-344
rats. Decompressive surgery was performed 17 hours after the occlusion. Each animal
was assessed before, 2 hours after, and 4 hours after surgery by means of diﬀusionweighted (DW), T2-weighted (T2W), and ﬂow-sensitive alternating inversion recovery
(FAIR) perfusion-weighted MRI. Ischemic damage was also evaluated in hematoxylineosin-stained brain sections.
Lesion volume as derived from ADC maps decreased from 522 ± 98 mm3 before to
405 ± 100 mm3 (P = 0.016) 4 hours after decompressive surgery, whereas lesion volume
from T2 maps increased from 420 ± 66 mm3 before to 510 ± 92 mm3 (P = 0.048).
Midline shift decreased from 1.4 ± 0.1 mm to 0.5 ± 0.2 mm (P = 0.001). Blood ﬂow in
the non-infarcted area of the ipsilateral hemisphere improved from 25 ± 9 ml/min/
100g tissue to 38 ± 9 ml/min/100g tissue (P = 0.035). Despite the pseudonormalization
of ADC, irreversible damage was found in the entire MCA territory on histological
evaluation.
In rats with space-occupying cerebral infarction, delayed decompressive surgery
leads to increased perfusion of non-infarcted areas, possibly preventing secondary
damage. The observed decrease in lesion volume derived from ADC maps is probably
a result of increased extracellular water formation. There are no signs that this reﬂects
rescue of ischemic tissue.
68
Decompressive surgery in rats with space-occupying infarction
F
atal brain edema occurs in 1 to 5% of patients with a supratentorial infarct.1
Patients with a hemispheric infarct and space-occupying brain edema have a
poor prognosis: in prospective series, mortality was about 80%.2 There are no
medical treatment strategies of proven eﬃcacy.3 Non-randomized studies have suggested that decompressive surgery lowers mortality without increasing the percentage
of severely disabled survivors.4,5
Studies of decompressive surgery in rat models of brain infarction have shown
that this procedure reduces infarct volume and improves cerebral blood ﬂow when
performed within 12 hours after occlusion of the middle cerebral artery (MCA).6-9
Infarct volume increased with time between MCA-occlusion and decompressive
surgery.7,9 Clinical studies also favor early over delayed hemicraniectomy.5 Still, most
clinicians will be inclined to perform this procedure only in later stages, when the
patient has signs of impending herniation. There is no experimental support that
therapy in this stage will improve outcome.
In the present study, we investigated the eﬀect of delayed hemicraniectomy on lesion
volume, midline shift, and cerebral blood ﬂow in rats with space-occupying cerebral
infarction. MRI experiments, consisting of diﬀusion- (DW), perfusion- (PW), and T2weighted (T2W) imaging, were performed directly before and at two time points after
decompression.
Methods
Animal model
The experiments were performed according to a protocol approved by the Utrecht
University Animal Experiment Ethical Committee. Animals had free access to food
and water.
Six male Fisher344 rats (Iﬀa-Credo Broekman, Someren, the Netherlands), weighing
280-320 g, were anesthetized by intraperitoneal injection of 0.16 mg/kg fentanyl citrate,
5 mg/kg ﬂuanisone, and 2.5 mg/kg midazolam, followed by subcutaneous injection
of 0.05 mg/kg atropine sulfate. They were intubated and mechanically ventilated
(Amsterdam Infant Ventilator, MK3, Hoekloos, the Netherlands) with 35% oxygen
and 65% nitrous oxide. During all experiments, rectal temperature was maintained
between 36.5 and 37.5°C by means of a water heating pad. The left femoral artery was
cannulated to monitor arterial blood pressure (Datascope 3000 Monitor, Datascope
corp), and to obtain blood for arterial blood gas analysis before and after MCAocclusion (ABL 505/ OSM3, Radiometer, Copenhagen, Denmark). When necessary,
respiratory adjustments were made. After surgery, the animals received 0.3 mg/kg
buprenorﬁne subcutaneously for relief of pain.
MCA-occlusion was achieved by a minor modiﬁcation of the intraluminal ﬁlament
technique.10 In short, the right common (CCA), internal (ICA), and external (ECA)
69
Chapter 5
carotid arteries were exposed through a ventral cervical midline incision. The
pterygopalatine artery, ECA and CCA were ligated with a 7.0 silk suture. A 3.0 prolene
suture, with a rounded tip and coated with poly-L-lysine (0.1% wt/vol in deionized
water; Sigma)11 was introduced into the CCA and advanced into the ICA until a
slight resistance was felt. A 7.0 silk suture was tightened around the CCA and the
intraluminal thread. Thereafter the incision was closed.
Sixteen hours after MCA occlusion the rats were neurologically examined according
to the scale introduced by Bederson et al.,12 and reﬁned by Menzies et al.13 Thereafter,
they were intubated and ventilated with halothane 1% in a mixture of 35% oxygen and
65% nitrous oxide for MRI experiments and hemicraniectomy. They remained under
anesthesia during the surgical procedure and all measurements.
Decompressive surgery was performed 17 hours after MCA-occlusion as described
by Forsting et al.7 Rats were immobilized in a stereotactic holder, the skull was exposed
through a dorsal midline incision, and the temporal muscle was removed partially. A
bone ﬂap of 10 x 5 mm2 was created in the parietal and temporal bone using a highspeed mini-drill. The dura was opened by a large incision. Thereafter, the temporal
muscle and the skin were adapted and sutured in place.
MRI experiments
MRI experiments were performed just before, and at 2 and 4 hours after decompressive
surgery on a 4.7 T Varian (Palo Alto, USA) horizontal bore spectrometer. RF excitation
and signal detection were accomplished by a Helmholtz volume coil (diameter 9 cm,
length 10 cm) and an inductively coupled surface coil (diameter 2 cm), respectively.
A spin-echo sequence was used for determination of the position of the animal (echo
time (TE) / repetition time (TR) = 40 ms / 1 s, matrix (M) = 128 x 64, ﬁeld of view
(FOV) = 5.0 x 3.0 cm2, 21 1.0 mm thick sagital slices, number of excitations (NEX) =
1). Eight contiguous 1.7 mm thick transversal slices were planned, with the ﬁrst slice
4mm anterior to the center of the eyes.
A single-scan diﬀusion-trace MRI sequence (4 b values: 100-1780 s / mm2, TE / TR
= 100 ms / 2 s, M = 128 x 64, FOV = 3.2 x 3.2 cm2, NEX = 2) was used to generate
quantiﬁed images of the tissue water trace apparent diﬀusion coeﬃcient (ADC). For
DW imaging a double spin-echo pulse sequence was used with four pairs of bipolar
gradients with speciﬁc predetermined signs in each of the three orthogonal directions.14
The combination of gradient directions leads to a cancellation of all oﬀ-diagonal tensor
elements, measuring the trace of the diﬀusion tensor. This provides unambiguous and
rotationally invariant ADC values in one experiment. To minimize the high sensitivity
of DW imaging to motion, the acquisition was triggered to the respiratory cycle.
T2W-images were acquired using a multi-echo sequence (TE / TR = 17 ms + 7 x 17
m s / 3 s, M = 256 x 128, FOV = 3.2 x 3.2 cm2, NEX = 2). To minimize interference at
the slice boundaries, slices were acquired in alternating order.
70
Decompressive surgery in rats with space-occupying infarction
PWI was performed in a single slice through the infarct core (corresponding to
slice number ﬁve from diﬀusion and T2-weighted images), with use of Flow-sensitive
Alternating Inversion Recovery (FAIR).15 A slice-selective and a non-selective inversion
recovery image (Mss and Mns) were acquired with turbo-fast low angle shot (FLASH)
acquisition and a suﬃcient inversion time (TI) to allow inﬂow of labeled spins into
brain tissue.16 MR parameters were as follows: ﬂip angle (α) = 20º, TE / TR / TI = 3
ms / 6 ms / 2000 ms, pre-delay = 2.0 s (total TR = 4.75 s), M = 128 x 128, FOV = 3.2
x 3.2 cm2, slice thickness = 1.7 mm, NEX = 96. For normalization of the FAIR signal
an equilibrium magnetization (M0) image was acquired with the same parameters,
without inversion. For quantiﬁcation of the FAIR signal a slice-selective T1-map was
acquired by Turbo-FLASH Look-Locker acquisition17 (α / TE / TR = 5º / 3 ms / 11 ms,
10 TIs (0.4 + 9 x 1.4 s), M = 128 x 128, FOV = 3.2 x 3.2 cm2, slice thickness = 1.7 mm,
NEX = 8).
Data processing and analysis
All images were zero-ﬁlled to 256 x 256. ADC and T2 maps were generated by monoexponential ﬁtting with IDL (Interactive Data Language, Research Systems, Boulder,
USA). Parametric images were analyzed in anatomic regions of interest using in-house
software. Calculations of lesion volume were based on ipsilateral ADC or T2 diﬀerences
of more than 20% as compared to the mean value in the contralateral hemisphere. This
threshold corresponds to a diﬀerence of more than 2 standard deviations (SD) but is
less inﬂuenced by noise than the SD. Moreover, this threshold is close to the 23% drop
in ADC found to correlate with ATP depletion 1 hour after MCA occlusion.18
Measurements of midline shift were performed with the software package
ImageBrowser (SISCO/Varian) at the level of the infarct core, corresponding to slice
number ﬁve of T2W images, according to the formula midline shift = (total diameter /
2) – contralateral diameter, using the third ventricle as a landmark.
PWI data were processed with IDL. Relative FAIR images were obtained by
subtracting the non-selective inversion recovery image from the slice selective
inversion recovery image and dividing the result by the M0 image ((Mss – Mns) / M0). T1
maps were obtained by mono-exponential ﬁtting with a correction for magnetization
saturation.19 From these relative FAIR images and T1 maps, CBF maps in milliliters
per minute per 100g tissue were calculated by T1 correction according to Calamante
et al.,20 assuming perfect inversion, a homogeneous blood brain partition coeﬃcient
of 0.9 ml / g and a blood T1 of 2.0 s. Further analysis was performed on region of
interest (ROI) basis with ImageBrowser. Infarct ROIs were drawn in the T2 maps and
deﬁned as those regions with a 20% increase of T2 as compared to the mean value of
the contralateral hemisphere. Peri-infarct ROIs were deﬁned as those covering the
non-infarcted parts of the aﬀected hemisphere, excluding the ventricles. Contralateral
ROIs consisted of the whole contralateral hemisphere. These ROIs were transferred to
71
Chapter 5
the CBF maps and CBF was determined in the infarct core, in the peri-infarct region,
and in the contralateral hemisphere.
Statistical analyses of MR parameters before and 2 and 4 hours after surgical
treatment were performed by means of repeated measures analysis of variance.
Outcome measures are expressed as means ± SD. Diﬀerences were considered
signiﬁcant at levels of P < 0.05.
Histology
After the MRI experiments the rats were killed by an intraperitoneal injection of 150
mg pentobarbital. The brains were removed and stored in a 4% phosphate-buﬀered
formaldehyde solution. After dehydration in a phosphate-buﬀered 25%-sucrose
solution, coronal cryosections (25 μm) were cut and stained with hematoxylin and
eosin for histopathological evaluation. Infarct was deﬁned as the area of pallor caused
by loss of aﬃnity for hematoxylin aﬀecting all cell types except inﬁltrated inﬂammatory
cells. Cells were considered to be damaged irreversibly when the cytoplasm had become
intensively eosinophilic while the nucleus had become pyknotic, or when the cellular
nucleic acids had completely lost their aﬃnity for hematoxylin (ghost neurons).21
Counts of neurons and glial cells were performed in speciﬁed areas of infarcted cortex
and striatum, and in the corresponding areas of the contralateral hemisphere, in slices
corresponding to slice number ﬁve of the MRI experiments.
Results
Intraoperative physiological variables remained within the normal range (data not
shown). Before hemicraniectomy all rats had a neurological score of 4 according to
Bederson et al., indicating spontaneous contralateral circling.12
DW and T2W MRI
Before surgery, ischemic areas were visualized on ADC and T2 maps and covered the
complete MCA territory in all animals (Figure 1). Mean ADC and T2 values of the
contralateral hemisphere, threshold values, and values within the infarct and in periinfarct tissue are presented in Table 1. Values of the contralateral hemisphere (and
thereby threshold values) did not change signiﬁcantly over time. Mean lesion volume
as calculated from the ADC maps decreased from 522 ± 98 mm3 before to 458 ± 92
mm3 at two hours, and 405 ± 100 mm3 at four hours after surgery (P = 0.016, Figure
2a). Of total lesion volume calculated from ADC maps, the percentage with increased
ADC values increased from 4.6 ± 2.7% to 10.4 ± 6.7% (P = 0.05). In contrast to ADC
lesion volumes, lesion volume as deduced from T2 maps increased from 420 ± 66 mm3
before to 457 ± 94 mm3 at two hours, and 510 ± 92 mm3 at four hours after surgery (P
= 0.048, Figure 2b).
72
Decompressive surgery in rats with space-occupying infarction
Figure 1 ADC (a), T2 (b), and CBF (c) maps of a slice through the infarct
core before and four hours after hemicraniectomy.
Midline shift
Mean midline shift decreased from 1.4 ± 0.1 mm before to 0.7 ± 0.3 mm at two hours,
and 0.5 ± 0.2 mm at four hours after surgery (P = 0.001, Figure 3). The total volume of
the aﬀected hemisphere as deduced from T2 maps increased from 767 ± 19 mm3 to 820
± 58 mm3 at two hours, and 830 ± 52 mm3 at four hours after surgery (P = 0.033).
PWI
From 18 FAIR experiments the results of 5 could not be analyzed due to poor signalto-noise ratios as a result of low global CBF. The calculated perfusion maps of the
remaining experiments showed regions with perfusion deﬁcits covering the entire
MCA territory. Figure 1c shows an example of a perfusion map before and four hours
after surgery. Perfusion before and after surgery could be analyzed in 5 out of six rats.
The mean CBF in the peri-infarct region increased from 25 ± 9 ml / min / 100 g tissue
73
Chapter 5
Table 1 ADC and T2 values of the contralateral hemisphere, threshold values, and values within the lesion and
in peri-infarct tissue.
Before surgery Two hours after
surgery
Four hours after
surgery
P
ADC contralateral (x10-4mm2/s)
7.5 ± 0.5
7.3 ± 0.4
7.5 ± 0.5
0.4
ADC threshold for decreased values
(x10-4mm2/s)
6.0 ± 0.4
5.8 ± 0.3
6.0 ± 0.4
0.4
ADC within part of lesion with
decreased values (x10-4mm2/s)
4.4 ± 0.2
4.5 ± 0.3
4.4 ± 0.7
0.3
ADC threshold for increased values
(x10-4mm2/s)
9.0 ± 0.5
8.8 ± 0.5
9.0 ± 0.6
0.4
ADC within part of lesion with
increased values (x10-4mm2/s)
10.6 ± 0.5
11.1 ± 0.4
12.1 ± 1.6
0.8
ADC peri-infarct (x10-4mm2/s)
7.1 ± 0.4
7.4 ± 0.4
7.6 ± 0.4
0.7
T2 contralateral (ms)
53 ± 2
53 ± 4
53 ± 2
0.2
T2 threshold value (ms)
63 ± 3
64 ± 5
64 ± 3
0.6
T2 within lesion (ms)
80 ± 4
81 ± 3
83 ± 3
0.5
T2 peri-infarct (ms)
51 ± 1
52 ± 2
52 ± 2
0.6
Data are expressed as mean ± SD
Figure 2 Lesion volume from ADC (a) and T2 (b) maps before and after hemicraniectomy.
a
Error bars indicate mean ± SD.
74
b
Decompressive surgery in rats with space-occupying infarction
Figure 3 Midline shift before and after decompressive
surgery. Error bars indicate mean ± SD.
Figure 4 Regional cerebral blood ﬂow (rCBF) in noninfarcted tissue of the ipsilateral hemisphere before and
after decompressive surgery. Error bars indicate mean ±
SD.
before to 37 ± 9 ml / min / 100 g tissue two hours, and 38 ± 5 ml / min / 100 g tissue
four hours after surgery (P = 0.035, Figure 4). There were no diﬀerences in CBF of
the infarcted region and the contralateral hemisphere before and after decompressive
surgery (14 ml / min / 100 g tissue before and 12 ml / min / 100 g tissue four hours after
surgery, P = 0.6, and 46 ml / min / 100 g tissue before and 44 ml / min / 100 g tissue
four hours after surgery, P = 0.4 respectively).
75
Chapter 5
Histology
Signs of irreversible tissue damage, including cytoplasmic eosinophilia aﬀecting both
neurons and glial cells, clumping of nuclear chromatin, nuclear pyknosis, and ghost
neurons, were observed in the entire MCA territory, including the parts where ADC
had normalized. Cell density in the infarct as compared to contralateral decreased
with 40 ± 14% in the cortex (P = 0.001) and 43 ± 14% in the striatum (P = 0.001).
Discussion
In this model of space-occupying cerebral infarction, hemicraniectomy and durotomy
performed when herniation was impending gave rise to a reversal of midline shift
and an increase in blood ﬂow in non-infarcted tissue of the ipsilateral hemisphere.
Apparent lesion volume as measured on ADC maps decreased, suggesting a beneﬁcial
eﬀect on ischemic damage. However, there was an increase in lesion volume calculated
from T2 maps, and histological evaluation demonstrated irreversible damage in the
complete MCA territory, including areas with (pseudo)normalized ADC values.
ADC reductions in ischemic tissue are thought to be caused by cytotoxic edema
and the associated decrease in extracellular water,22 and are a very sensitive indicator
of early ischemic brain damage.23 Several investigators have observed reversal of DWI
abnormalities after transient focal ischemia followed by delayed recurrence.23,24 It
has been suggested that this phenomenon is caused by secondary ischemic damage,
after an initial recovery of the energy status.24 This is the ﬁrst report that describes an
increase in ADC after hemicraniectomy in space-occupying cerebral infarction. As
irreversible injury was observed on histological evaluation in the entire MCA territory,
including the parts where ADC had (pseudo)normalized, the decrease in apparent
lesion volume on ADC maps did not indicate tissue recovery. Given the increase in
apparent lesion volume as deduced from the T2 maps, and the increase in volume of
the entire hemisphere, we suggest that the increase of ADC values may be caused by
an increased extracellular space by vasogenic edema.
T2 prolongation of brain tissue water, measured by T2W MRI, is thought to reﬂect
vasogenic edema and is probably most pronounced in irreversibly damaged tissue.25An
increase in extracellular space in our model may be explained by a decreased interstitial
pressure caused by the bony decompression. This is in line with earlier ﬁndings. Cooper
et al. showed that bony decompression resulted in exacerbation of edema in a cold
lesion model in dogs.26 In a rat model of MCA infarction, Engelhorn and colleagues
found that lesion volume after a combination of reperfusion and decompressive
surgery was signiﬁcantly larger than lesion volume after reperfusion alone.8
Analysis of the CBF maps revealed a region with a clear perfusion deﬁcit covering
the entire MCA territory in each animal. After surgery, the area with the perfusion
deﬁcit slightly increased, probably as a result of the increase of lesion volume by edema.
76
Decompressive surgery in rats with space-occupying infarction
CBF in the non-infarcted parts of the ipsilateral hemisphere improved signiﬁcantly,
supporting one of the rationales of decompressive surgery, which is to decrease ICP to
improve cerebral perfusion.4,5 However, in contrast to a previous study,8 we observed
no increase in CBF in the territory of the occluded artery. In this previous study,
treatment was applied one hour after MCA occlusion, when there was no evidence
of edema yet. It is unlikely that delayed hemicraniectomy as performed in our study
(and in most patients), exerts a beneﬁcial eﬀect by saving penumbral tissue in the area
of the occluded artery. Protection from secondary ischemic damage, if present, may
be explained by the increase of CBF in the non-aﬀected vascular territories, thereby
preventing recruitment of these territories in the infarct.
We did not compare ADC, T2, and CBF values after decompressive surgery with
those in controls. However, in a previous study on the natural course of tissue damage
in rats with space-occupying infarction we did not observe normalization of ADC
values, nor reversal of midline shift or increase of CBF in the peri-infarct region in the
time period between 16 and 24 hours after MCA-occlusion.
In conclusion, delayed decompressive surgery might be beneﬁcial by reversal of
tissue shifts with improvement of CBF in vascular territories surrounding the infarct.
We found no beneﬁcial eﬀect of surgery on infarcted tissue within the MCA territory.
Therefore, the intervention may only prevent additional damaged in already severely
aﬀected patients. Randomized clinical studies focusing on functional outcome are
required before implementing this strategy as a standard treatment modality.27
References
1.
2.
3.
4.
5.
6.
7.
8.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
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 1996;53:309-315.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral infarction. Crit
Care Med 2003;31:617-625.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Doerﬂer A, Forsting M, Reith W. Decompressive craniectomy in a rat model of “malignant” cerebral
hemispheric stroke: experimental support for an aggressive therapeutic approach. J Neurosurg
1996;85:853-859.
Forsting M, Reith W, Schabitz WR. Decompressive craniectomy for cerebral infarction. An
experimental study in rats. Stroke 1995;26:259-264.
Engelhorn T, Doerﬂer A, Kastrup A. Decompressive craniectomy, reperfusion, or a combination for
early treatment of acute “malignant” cerebral hemispheric stroke in rats? Potential mechanisms
studied by MRI. Stroke 1999;30:1456-1463.
77
Chapter 5
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
78
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 2002;33:617-622.
Koizumi J, Yoshida Y, Nakazawa T, Ooneda G. Experimental studies of ischemic brain edema, I: a new
experimental model of cerebral embolism in rats in which recirculation can be introduced in the
ischemic area. Jpn J Stroke 1986;8:1-8.
Belayev L, Alonso OF, Busto R, Zhao W, Ginsberg MD. Middle cerebral artery occlusion in the rat by
intraluminal suture. Neurological and pathological evaluation of an improved model. Stroke
1996;27:1616-1622.
Bederson JB, Pitts LH, Tsuji M, Nishimura MC, Davis RL, Bartkowski H. Rat middle cerebral artery
occlusion: evaluation of the model and development of a neurologic examination. Stroke 1986;17:472476.
Menzies SA, Hoﬀ JT, Betz AL. Middle cerebral artery occlusion in rats: a neurological and pathological
evaluation of a reproducible model. Neurosurgery 1992;31:100-106.
de Graaf RA, Braun KP, Nicolay K. Single-shot diﬀusion trace (1)H NMR spectroscopy. Magn Reson
Med 2001;45:741-748.
Kim SG. Quantiﬁcation of relative cerebral blood ﬂow change by ﬂow- sensitive alternating inversion
recovery (FAIR) technique: application to functional mapping. Magn Reson Med 1995;34:293-301.
Schepers J, Veldhuis WB, Pauw RJ, de Groot JW, van Osch MJP, Nicolay K, van der Sanden BPJ.
Comparison of FAIR perfusion kinetics with DSC-MRI and functional histology, in a model of transient
ischemia. Magn Reson Med 2004;51:312-320.
Haase A, Matthaei D, Bartkowski R, Duhmke E, Leibfritz D. Inversion recovery snapshot FLASH MR
imaging. J Comput Assist Tomogr 1989;13:1036-1040.
Olah L, Wecker S, Hoehn M. Relation of apparent diﬀusion coeﬃcient changes and metabolic
disturbances after 1 hour of focal cerebral ischemia and at diﬀerent reperfusion phases in rats. J Cereb
Blood Flow Metab 2001;21:430-439.
Deichmann R, Haase A. Quantiﬁcation of T1 values by SNAPSHOT-FLASH NMR imaging. J Magn
Reson 1992;96:608-612.
Calamante F, Williams SR, van Bruggen N, Kwong KK, Turner R. A model for quantiﬁcation of
perfusion in pulsed labelling techniques. NMR Biomed 1996;9:79-83.
Garcia JH, Yoshida Y, Chen H. Progression from ischemic injury to infarct following middle cerebral
artery occlusion in the rat. Am J Pathol 1993;142:623-635.
Moseley ME, Cohen Y, Mintorovitch J. Early detection of regional cerebral ischemia in cats:
comparison of diﬀusion- and T2-weighted MRI and spectroscopy. Magn Reson Med 1990;14:330346.
Dijkhuizen RM, Berkelbach van der Sprenkel JW, Tulleken KA, Nicolay K. Regional assessment of
tissue oxygenation and the temporal evolution of hemodynamic parameters and water diﬀusion
during acute focal ischemia in rat brain. Brain Res 1997;750:161-170.
Neumann-Haefelin T, Kastrup A, de Crespigny A. Serial MRI after transient focal cerebral ischemia
in rats: dynamics of tissue injury, blood-brain barrier damage, and edema formation. Stroke
2000;31:1965-1972.
Olah L, Wecker S, Hoehn M. Secondary deterioration of apparent diﬀusion coeﬃcient after 1-hour
transient focal cerebral ischemia in rats. J Cereb Blood Flow Metab 2000;20:1474-1482.
Cooper PR, Hagler H, Clark WK, Barnett P. Enhancement of experimental cerebral edema after
Decompressive surgery in rats with space-occupying infarction
27.
decompressive craniectomy: implications for the management of severe head injuries. Neurosurgery
1979;4:296-300.
Hofmeijer J, Amelink GJ, Algra A, van Gijn, Macleod MR, Kappelle LJ, van der Worp HB; the
HAMLET investigators. Hemicraniectomy After Middle cerebral artery infarction with Lifethreatening Edema Trial (HAMLET). Protocol for a randomised controlled trial of decompressive
surgery in space-occupying hemispheric infarction. Trials 2006 Sep 11;7:29.
79
Part II
Clinical studies
Predictors of life-threatening edema in MCA infarction
Chapter 6
Predictors of life-threatening brain edema in
middle cerebral artery infarction.
Jeannette Hofmeijer, Ale Algra, L. Jaap Kappelle, and H. Bart van der Worp
Submitted
83
Chapter 6
Summary
Clinical and preclinical data suggest that outcome after decompressive surgery in
patients with middle cerebral artery (MCA) infarction and life-threatening edema
formation is better if surgery is performed as early as possible. Early treatment requires
identiﬁcation of patients at risk. In this chapter a systematic review is presented
to identify predictors of life-threatening edema formation in patients with MCA
infarction.
Medline and Embase have been searched from January 1966 and January 1974,
respectively, to August 2006 for cohort and case-control studies on predictors of lifethreatening edema formation in patients with MCA infarction. Data were extracted by
means of a predeﬁned data extraction form and crude data were used to calculate risk
ratios, odds ratios, or weighted mean diﬀerences.
Infarct size was the major determinant for life-threatening edema formation. Cut-oﬀ
values of 50% of the MCA territory for infarct size and 66% for perfusion deﬁcit were
found. Other associated signs were early mass eﬀect, involvement of other vascular
territories, higher body temperature, internal carotid artery occlusion, and need for
mechanical ventilation. However, predictive values were moderate.
In patients with MCA infarction, prediction of life-threatening edema formation
remains tempting. The size of the ischemic area is the major determinant. However,
single predictors lack suﬃcient predictive value to select candidates for surgical
decompression before the onset of clinical signs of herniation.
84
Predictors of life-threatening edema in MCA infarction
L
arge middle cerebral artery (MCA) infarcts may be complicated by spaceoccupying and life-threatening edema formation, usually between the second
and ﬁfth day after stroke onset. Case fatality rates as high as 78% have been
reported.1 Fatal edema occurs in 1 to 5% of all patients with a supratentorial infarct.2,3
Because of the limitations of medical treatment strategies to reduce edema
formation,4 there have been proposals for decompressive surgery consisting of a
large hemicraniectomy and a duraplasty.5,6 In rats with MCA infarction, surgical
decompression reduced infarct volume when performed early after the onset of
ischemia, but not when start of treatment was delayed.7 Although a recent systematic
review of retrospective and uncontrolled clinical studies failed to show a signiﬁcant
eﬀect of timing of surgery on outcome,8 a large prospective study not included in
this review also suggested that functional outcome is better if treatment is started
early, even before clinical deterioration.6 If this suggestion is conﬁrmed in ongoing
clinical trials, patients at risk for developing fatal edema should be identiﬁed as early
as possible.
Diﬀerent early predictors of life-threatening edema are described, ranging from the
presence of vomiting to infarct volume.9-30 In this chapter the results of a systematic
review to identify predictors of life-threatening edema formation in patients with
MCA infarction are presented.
Methods
Search strategy and selection criteria
We searched Medline from January 1966 to August 2006 and Embase from January 1974
to August 2006 for cohort and case control studies on predictors of life-threatening
cerebral edema formation in patients with MCA infarction. The search terms we
used were (predict* OR prognos*) AND (stroke OR (middle cerebral artery)) AND
(edema OR oedema OR swelling OR malignant OR herniation OR space-occupying).
Publications were identiﬁed independently by the ﬁrst and by the last author. Reference
lists of all relevant publications were checked for additional articles. This method of
cross-checking was continued until no further studies were found.
Inclusion criteria
Criteria for inclusion of studies were (1) cohort or case-control studies conducted
among patients with acute MCA infarction, published as full articles and written in
English, French, or German; (2) clear description of the cohort in cohort studies and
the control group in case-control studies; and (3) outcome deﬁned as neurological
deterioration with a decrease in consciousness and signs of cerebral herniation as a
result of cerebral edema formation. Studies were also included if patients died or if
invasive therapeutic procedures, such as decompressive surgery, were performed, after
85
Chapter 6
the above outcome criteria had been met. For inclusion in the meta-analysis, the studies
had to present crude data to allow recalculations in our analyses. However, we were
not able to use individual patient data. Studies were excluded if (1) only a radiological
outcome was described, without clinical information, or if (2) brain imaging was not
performed and patients with other lesions than MCA infarcts could therefore not be
excluded. Data were included only once if used in multiple publications.
Data extraction
Data were systematically extracted with use of a predeﬁned data extraction form.
Diﬃculties were resolved by consensus discussions between the ﬁrst and the last
author.
Data analysis
Cochrane Collaboration’s Review Manager software (version 4.2) was used for
analysis. For analysis of crude data, data of at least two studies had to be available on
the same potential risk factor. Crude dichotomous data were used to construct 2 x 2
tables and to calculate risk ratios (RR) or odds ratios (OR) where appropriate, as well
as positive and negative predictive values. Data from cohort and case control studies
were not combined. Crude continuous data were used to calculate weighted mean
diﬀerences (WMD). If there was statistically signiﬁcant heterogeneity among the
results of included studies (P < 0.05), we used random- instead of ﬁxed-eﬀect models,
because these include both within-study variance and between-study variation in the
assessment of uncertainty of the meta-analysis.31 Data are presented as point estimates
with 95% conﬁdence intervals (CI). The boxes in the ﬁgures describe the study size,
with larger boxes for larger studies.
Results
Of 955 citations identiﬁed after searching Medline and Embase with the above search
terms, 84 were considered possibly relevant after screening of titles and abstracts, and
reviewed in detail. Hand searching of the reference lists of these articles identiﬁed 8
additional publications that were reviewed in detail. Sixty-nine articles were excluded
for one or more of the following reasons: no crude patient data (n=6), duplicate
publication (n=3), intervention study without analysis of predictors of edema
formation (n=2), inclusion of stroke subtypes diﬀerent from MCA infarction (n=12),
death as an outcome measure without adequate speciﬁcation of its cause (n=20),
other clinical outcome measure diﬀerent from neurological deterioration as a result
of life-threatening edema formation (n=23), insuﬃcient deﬁnition of the outcome
measure (n=6), and outcome deﬁned as severe edema on CT or MRI without clinical
information (n=11). Eventually, 23 articles were included in the present review. Details
86
Predictors of life-threatening edema in MCA infarction
of these studies are given in Table 1.
We were able to perform a meta-analysis of crude data on the following potentially
predictive factors: age, sex, gaze palsy on admission, need for mechanical ventilation,
infarct size, early signs of infarction on CT, side of infarction, involvement of
other vascular territories besides the MCA territory, results of perfusion-weighted
imaging, thrombolysis, recanalization, hemorrhagic transformation, blood glucose
on admission, diabetes mellitus, temperature on admission, infarct etiology, and the
presence of MCA or ICA occlusion. Because of the use of diﬀerent cut-oﬀ points, it
was not possible to combine data for some other predictive factors. For this reason,
data on infarct severity, level of consciousness, blood pressure, fever, brain atrophy,
neuro-monitoring, cerebral perfusion, and midline shift are presented in a narrative
synthesis.
Meta-analysis
The results of the meta-analysis are summarized in Table 2. We found a total of 27
factors evaluated for their relation with life-threatening edema formation, of which
12 were statistically signiﬁcant predictors of this complication. For infarct size, cut-oﬀ
values of 50%, 66%, and 100% of the MCA territory were used,9,18,20,21,23,26,27,29,30 and for
perfusion deﬁcit a cut-oﬀ value of 66% of the MCA territory. A need for mechanical
ventilation, infarction of more than 66% of the MCA territory, and a perfusion deﬁcit
larger than 66% of the MCA territory were found to be the strongest predictors in this
meta-analysis. However, predictive values were only moderate, and highly dependent
on the incidence of life-threatening edema, which ranged from 10 % to 78 % in the
cohort studies included in this review (Table 2). MCA occlusion was associated with
a signiﬁcantly reduced risk. Meta-analyses of age, infarct size, involvement of other
vascular territories, and thrombolysis are presented in Figures 1-4.
Narrative synthesis
The investigators of 11 studies reported on severity of the neurological deﬁcits at
admission and risk of life-threatening edema formation. Severity was expressed as a
score on either the National Institutes of Health Stroke Scale (NIHSS)11,12,14,21,26,27,30 or
the Scandinavian Stroke Scale (SSS)1,10,15,29 and results were reported either on the basis
of mean or median scores, or dichotomized with diﬀerent cut-oﬀ points. This made a
formal meta-analysis impossible. In general, patients with life-threatening edema had
higher severity scores, either with statistical signiﬁcance,26,27,30 or not.1,11,12,14,15,21,29 Level
of consciousness on admission was not signiﬁcantly diﬀerent in most studies.10,22,24,25,28
In only one study level of consciousness on admission was signiﬁcantly lower in patients
going on to life-threatening edema, but the deﬁnition of this factor was unclear.18
Neither mean arterial pressure on admission,11 nor systolic blood pressure 12 hours
87
Chapter 6
Mechanical ventilation
Gaze palsy
Severity
Gender
Age
Time from onset of symptoms
(hours)
Population
Incidence of life-threatening
edema (%)
Controlsb
Cohorta
Cases
Year
Table 1 Details of included studies.
Berrouschot9
1998
11
108
10
ICU
6
-
-
-
-
-
Berrouschot10
1998
37
53
70
ICU
12
+
-
+
+
-
Bosche,11
2003
ICU
12
+
+
+
-
-
-
-
-
-
14
31
45
Dohmen,12
17
34
50
Heiss17
16
34
47
1998
6
20
30
New
6
-
Foerch
2004
16
51
31
New
6
+
+
+
-
+
Gerriets15
2001
12
42
29
New
16
+c
+c
+
-
-
ICU
12
+
c
+
+
+
+
New
18
+c
+
-
-
-
ICU
48
+
+
-
-
-
Firlik
13
14
1
Hacke and
1996
43
55
78
Haring16
1999
31
18
2001
94
19
Krieger
1999
23
New
6
+
-
-
-
-
Kucinski20
1998
17
74
23
New
6
+c
-
-
-
-
Lee21
2004
10
31
32
New
6
+
-
+
-
-
Limburg22
1990
6
26
23
New
24
-
-
-
+
-
Manno23
2003
22
36
61
New
12
+
+
-
-
-
Maramattom24
2004
14
24
58
ICU
?
+c
+
-
+
-
Mori25
2001
34
55
62
New
6
+
+
-
-
-
Oppenheim26
2000
10
28
36
New
14
+
+
+
-
-
27
Ryoo
2004
11
27
41
New
6
-
-
+
-
-
Stolz29
2002
9
21
43
New
120
+
+
+
-
-
+
+
-
Schwab28
Kasner
30
Thomalla
2003
11
31
201
47
112
37
30
New
6
+
a
c
b
Variables were included in this table only if studied by at least two authors. indicates cohort studies; , casecontrol studies; c, not included in the meta-analysis because of lack of suﬃcient crude data; d, not included in
88
Predictors of life-threatening edema in MCA infarction
Perfusion-weighted imaging
Thrombolysis
Recanalization
Hemorrhagic transformation
Atrophy
Blood glucose/ DM
Temperature/ fever
Blood pressure/ hypertension
-
+d
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
-
-
-
-
-
-
-
+
-
+d
+
-
+
-
-
+
+
-
-
+
-
-
+c
-
-
+d
+c
-
+
-
-
-
-
-
-
-
-
-
+
+
-
+
-
-
-
+
-
+
+
-
-
-
+
c
c
Midline shift
Involvement of other territories
-
Neuro-monitoring
Infarct side
-
Vessel occlusion
Early signs
+
Etiology/ atrial ﬁbrillation
Infarct size
Table 1 – Continued –
-
-
-
-
-
-
-
-
-
-
-
-
+
-
+
-
-
+
-
-
+
+
-
-
-
-
-
+d
+
+
+
+
+
-
-
-
-
-
-
-
-
-
-
+
+
-
+
+
+
-
-
-
+
-
+
+
+
+
+
+
+
-
-
+
-
-
+
-
-
-
-
+
+
+
+
+
+
-
-
+
+
-
-
-
-
+
-
-
-
-
-
-
+
-
-
+
-
+
+
+
-
-
-
-
-
-
-
-
+
-
-
-
-
-
-
+d
-
-
-
-
-
-
-
-
-
-
-
+
-
+
-
-
+
-
-
-
-
-
-
-
+
-
-
-
-
+d
+
-
-
-
-
-
-
-
-
-
+
-
+
-
-
+d
+
-
+
+
+
-
-
-
-
-
-
-
+
+
-
+
+
-
-
-
-
-
-
-
-
-
+
-
-
+
+d
-
+
+
-
-
-
-
-
-
-
-
+
-
-
+
-
-
+
-
-
-
-
-
-
-
-
-
-
-
+
+
-
+
-
+c
+
-
-
-
-
-
-
+
+
-
-
the meta-analysis because it was not possible to combine data; ICU, target population admitted to an intensive
care unit; new, target population consists of newly admitted stroke patients; time from onset of symptoms, time
window within which factors were assessed.
89
90
20,25,28
10,18,19
12,13,18,25
History of diabetes mellitus
14,18
Blood glucose levels on admission
Hemorrhagic transformation
Recanalization
Thrombolysis
11,14,18,23,25,28,30
Low perfusion levels in other vascular territories
a
RR 1.2 (0.9-1.6)
WMD 0.20 mmol/l (–0.7-1.0)
RR 1.8 (0.9-3.7)
RR 0.9 (0.2-2.5)
RR 0.8 (0.6-1.1)
RR 3.7 (2.0-6.6)
a
RR 7.7 (2.5-24) a
Perfusion deﬁcit larger than 66% of the MCA territory21,27
c 21,27
RR 2.6 (2.0-3.2) a
Involvement of other vascular territories c 14,21,24-27,29
Early mass eﬀect
RR 1.1 (0.8-1.3)
a
RR 1.5 (1.2-2.0)
RR 0.9 (0.7-1.2)
RR 7.0 (2.5-19.4)
Left-sided versus right-sided1,11,10,21,23,26,30
18,20
b 18,20
Obscuration of the lentiform nucleus on admission
Infarction of the complete MCA territory
RR 7.5 (3.9-14.3)
Infarct size larger than 66% of the MCA territory
9,26,29,30
OR 9.2 (4.0-20.9)
9,21,27,29
Infarct size larger than 50% of the MCA territory
a
a
RR 2.0 (1.5-2.6)
16,19
Infarct size larger than 50% of the MCA territory
a
18,21,23
RR 10 (2.1-51)
RR 1.2 (0.8-1.7)
Conjugate gaze palsy on admission10,22,24
Mechanical ventilation during admission
RR 0.9 (0.8-1.1)
Men versus women1,11,14,18,23-26,29,30
a
WMD –3.2 (–6.4-0.0)e
Age10,11,14,18,19,21,23,25,26,29
d 1,14
Estimates (95% CI)
Predictor
Table 2 Results of the meta-analysis.
ﬁxed
ﬁxed
random
random
random
ﬁxed
ﬁxed
ﬁxed
ﬁxed
ﬁxed
ﬁxed
random
ﬁxed
ﬁxed
ﬁxed
ﬁxed
ﬁxed
ﬁxed
random
44
40
49
45
50
36
36
44
47
40
40
21
22
46
56
55
50
42
49
75
30
43
85
73
86
47
51
39
71
86
65
72
60
48
60
54
43
47
78
86
69
53
64
58
91
90
68
96
55
48
Eﬀect
Overall incidence Positive predictive Negative
modeling life-threatening value (%)
predictive
edema (%)
value (%)
Chapter 6
ﬁxed
ﬁxed
random
ﬁxed
ﬁxed
ﬁxed
ﬁxed
ﬁxed
45
37
46
38
31
31
61
42
63
51
21
25
33
55
82
57
40
63
70
Eﬀect
Overall incidence Positive predictive Negative
modeling life-threatening value (%)
predictive
edema (%)
value (%)
indicates statistically signiﬁcant diﬀerence; b, Sylvian ﬁssure obscuration, eﬀacement of sulci, and lateral ventricle compression on admission; c, anterior or posterior
cerebral artery; d, the time between the onset of symptoms and the need for ventilation was between 3 hours and 5 days in one study;28 e,the mean age of patients with
malignant edema was 3.2 years lower than the mean age of patients without.
Perfusion-weighted imaging was performed by CT,13,21,3 MRI,30 positron emission tomography (PET),11,12,17 or single photon emission CT (SPECT).9,22
Recanalization was measured by digital subtraction angiography (DSA).
Hemorrhagic transformation had to be well deﬁned and primary intra-parenchymal hemorrhages had to be excluded.
MCA occlusion was measured by DSA,1,20 CT angiography,29 or MR angiography,30 and internal cerebral artery (ICA) occlusion by DSA1,20 or MR angiography.26,30
a
RR 1.0 (0.6-1.6)
RR 2.8 (1.9-4.1)
OR 4.0 (0.5-35.7)
Hyperdense ICA sign on CT18,21
ICA occlusion
1,20,26,30
Hyperdense MCA sign on CT
a
RR 1.2 (0.9-1.5)
a
16,19
RR 0.5 (0.3-0.7)
RR 0.7 (0.4-1.2)
RR 1.1 (0.5-2.3)
18,21,23,24
Hyperdense MCA sign on CT
MCA occlusion
1,20,27,30
Cardioembolic infarction
14,30
Athero-thrombotic infarction
WMD 0.3°C (0.0-0.6) a
Core temperature on admission10,11
14,30
Estimates (95% CI)
Predictor
Table 2 – Continued –
Predictors of life-threatening edema in MCA infarction
91
Chapter 6
Figure 1 Weighted mean diﬀerences in age between patients with and without life-threatening edema.
Figure 2 Early infarct size and the risk of life-threatening edema formation.
a
b
Figure 3 Involvement of other vascular territories besides the MCA territory and the risk of life-threatening
edema formation.
92
Predictors of life-threatening edema in MCA infarction
Figure 4 Thrombolysis and the risk of life-threatening edema formation.
after the onset of symptoms,19 nor peak or through systolic pressure18 were identiﬁed
as a risk factor for life-threatening edema. Cerebral atrophy, a history of hypertension,
fever on admission, and atrial ﬁbrillation did not lead to signiﬁcantly altered risk in
one cohort,14 and one case-control19 study.
Lower levels of extracellular non-transmitter amino acids were found in non-infarcted
ipsilateral tissue in the ﬁrst 12 hours after the onset of symptoms in patients who
developed life-threatening edema.12,17 In the ﬁrst 12 hours after the onset of symptoms
intra-cranial pressure (ICP) values could not discriminate between patients who did
and did not go on to life-threatening edema.11,17 In one study higher ICP values were
found in patients with life-threatening edema, but it is unclear at what time points
these ICP measurements were done.1 Measurements of midline shift were presented
in 5 cohort studies1,15,24,25,29 and 1 case-control study.16 Measurements were performed
at diﬀerent time points and several cut-oﬀ points were used. Generally, diﬀerences in
midline shift between life-threatening and ‘non-life-threatening’ infarctions became
statistically signiﬁcant in the later stages of the disease. Maximum midline shift was
found earlier in patients with life-threatening edema than in patients without (day 2-4
versus day 3-7),1 and the progression of midline shift was faster.29
Independent predictors
Multivariate analysis of predictors of life-threatening edema formation was performed
in 9 studies.9,10,18,19,21,23,26,27,29 Variables found as independent predictors are summarized
in Table 3.
93
Chapter 6
Table 3 Variables found as independent predictors of life-threatening edema formation.
Predictor
Interval from onset Estimates (95% CI)
of symptoms (hours)
SPECT activity deﬁcit of the complete MCA territory9
6
SPECT graded scale > 1409
6
RR 79 (11-569)
18
a
Attenuated cortico-med contrast
16
18
RR 40 (10-161)
History of hypertension
OR 3.0 (1.2-7.6)
History of congestive heart failure18
OR 2.1 (1.5-3.0)
WBC
18
48
OR 1.07 / 1000
WBC/ ul (1.01-1.14)
48
OR 3.9 (2.4-6.4)
>50% MCA on CT
48
OR 6.3 (3.5-11.6)
Additional vascular territories on CT18
48
OR 3.3 (1.2-9.4)
Nausea or vomiting
24
OR 5.1 (1.7-15.3)
Systolic BP > 18019
12
OR 4.2 (1.4-12.9)
Hyperdense MCA sign
12
OR 29 (1.6-524)
>50% MCA CT23
12
OR 14 (1.04-189)
Volume DWI > 145 cm
14
a
Hypoperfusion on CT time to peak map27
6
OR 150 (a)
WBC > 10.000/ul18
18
19
23
3 26
SPECT indicates single photon emission computed tomography; MCA, middle cerebral artery; WBC, white
blood cells; BP, blood pressure; DWI, diﬀusion-weighted imaging; a, exact estimate could not be derived from
the publication.
Discussion
The major determinants of developing fatal brain edema after MCA infarction are
the size of the infarct and the size of the area with perfusion deﬁcit. In this systematic
review, infarct size larger than 50% of the MCA territory and a perfusion deﬁcit larger
than 66% were identiﬁed as risk factors. Also involvement of additional vascular
territories besides that of the MCA was signiﬁcantly associated with life-threatening
edema formation. However, predictive values were low. To date, selection of patients
for surgical decompression before the onset of clinical signs of herniation remains
tempting.
In two recent studies, lesion volume of more than 145 ml on diﬀusion-weighted
imaging (DWI) within 14 hours26 and more than 82 ml on apparent diﬀusion coeﬃcient
maps in the ﬁrst six hours30 predicted life-threatening edema formation with high
predictive values. These ﬁndings could not be subjected to meta-analysis, because
they were not tested in diﬀerent studies. Still, measurement of infarct volume with
94
Predictors of life-threatening edema in MCA infarction
DWI is simple, reliable, and promising and should be used in future studies on the
prediction of life-threatening edema formation.
Several ﬁndings were anticipated. First, in general, greater stroke severity was
associated with an increased risk of the development of life-threatening edema.
Second, we found a borderline signiﬁcant association with age, in which lower age
was associated with a greater risk. The lack signiﬁcance is probably the result of the
heterogeneity between the included studies. Third, body temperature on admission
was marginally higher in patients with life-threatening edema, than in patients without.
Stroke patients with fever have a worse prognosis than patients with a normal or low
body temperature.32 Moreover, it has been shown that moderate hypothermia can help
to control critically elevated ICP values in severe space-occupying edema after MCA
infarction.33
Paradoxically, MCA occlusion was associated with a lower risk. This is probably a
result of selection bias within the studies included in the meta-analysis, because the
group of patients with an MCA occlusion consisted mainly of patients with distal MCA
occlusions, whereas occlusions of the MCA trunk have been associated with poor
outcome.34,35 The presence of a hyperdense MCA sign is usually indicative of MCA
trunk occlusion36 and in the present meta-analysis a hyperdense MCA sign on CT led
to an increased RR and OR, but the results were not statistically signiﬁcant.16,18,19,21,23,24
Patients with ICA occlusions had a higher risk. This group had mainly intracranial
carotid occlusions, which is considered a poor prognostic sign.1,20
The association between levels of non-transmitter and transmitter amino acids in
peri-infarct tissue and life-threatening edema formation has been evaluated in two
studies. Whereas levels of transmitter amino acids discriminated only in later stages
of the disease,12,17 a statistically signiﬁcant association was found between lower levels
of extracellular non-transmitter amino acids within 12 hours after stroke onset and
formation of life-threatening edema.11 This is possibly a result of dilution due to excessive
vasogenic edema formation within the infarct, spreading into the extracellular space of
peri-infarct tissue.11 Sensitivities and speciﬁcities of approximately 80% were found at
speciﬁc cut-oﬀ values. These observations may emphasize the relevance of vasogenic
edema in peri-infarct zones, but it is unlikely that invasive neuro-monitoring will be
used on a large scale in future stroke units. For this reason, the clinical applicability of
these results is probably low.
The most important limitation of this systematic review and meta-analysis is the
large variety of inclusion criteria and methods of analysis among the included studies,
although we studied only articles on severe MCA infarction. The diﬀerent incidences
of life-threatening brain edema in the cohort studies reﬂect the large diﬀerences in
inclusion criteria of these studies. As compared with newly admitted patients, predictive
values in patients admitted to a neuro-critical care unit were higher, because a priori
chances were higher (Table 1). Moreover, the more time passes after stroke onset, the
95
Chapter 6
more precise the prognostic reliability of the diﬀerent parameters becomes. Especially
signs that are directly related to edema formation, such as raised ICP, midline shift,
and need for mechanical ventilation, are late rather than early outcome predictors.
Early prediction of the development of life-threatening edema formation and
selection of patients for decompressive surgery is diﬃcult in patients with MCA
infarction. The size of the ischemic area appears to be the major determinant. MRI with
DWI is probably the most reliable tool. However, none of the clinical and radiological
variables has suﬃcient predictive value to be useful for individual patients. Prediction
would probably improve if combinations of symptoms and signs would be used. Future
prognostic studies, in which all of the determinants described in the present review
are evaluated, may help to establish reliable predictive models.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
96
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 1996;53:309-315.
Heinsius T, Bogousslavsky J, Van Melle G. Large infarcts in the middle cerebral artery territory.
Etiology and outcome patterns. Neurology 1998;50:341-350.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral infarction. Crit
Care Med 2003;31:617-625.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Doerﬂer A, Forsting M, Reith W. Decompressive craniectomy in a rat model of “malignant” cerebral
hemispheric stroke: experimental support for an aggressive therapeutic approach. J Neurosurg
1996;85:853-859.
Gupta R, Connolly ES, Mayer S, Elkind MS. Hemicraniectomy for massive middle cerebral artery
territory infarction: a systematic review. Stroke 2004;35:539-543.
Berrouschot J, Barthel H, von-Kummer R, Knapp WH, Hesse S, Schneider D. 99m technetium-ethylcysteinate-dimer single-photon emission CT can predict fatal ischemic brain edema. Stroke
1998;29:2556-2562.
Berrouschot J, Sterker M, Bettin S, Koster J, Schneider D. Mortality of space-occupying (‘malignant’)
middle cerebral artery infarction under conservative intensive care. Intensive Care Med 1998;24:620623.
Bosche B, Dohmen C, Graf R. Extracellular concentrations of non-transmitter amino acids in periinfarct tissue of patients predict malignant middle cerebral artery infarction. Stroke 2003;34:29082913.
Dohmen C, Bosche B, Graf R. Prediction of malignant course in MCA infarction by PET and
microdialysis. Stroke 2003;34:2152-2158.
Firlik AD, Yonas H, Kaufmann AM. Relationship between cerebral blood ﬂow and the development of
Predictors of life-threatening edema in MCA infarction
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
swelling and life-threatening herniation in acute ischemic stroke. J Neurosurg 1998;89:243-249.
Foerch C, Otto B, Singer OC. Serum S100B predicts a malignant course of infarction in patients with
acute middle cerebral artery occlusion. Stroke 2004;35:2160-2164.
Gerriets T, Stolz E, Konig S. Sonographic monitoring of midline shift in space-occupying stroke: an
early outcome predictor. Stroke 2001;32:442-447.
Haring HP, Dilitz E, Pallua A. Attenuated corticomedullary contrast: An early cerebral computed
tomography sign indicating malignant middle cerebral artery infarction. A case-control study. Stroke
1999;30:1076-1082.
Heiss WD, Dohmen C, Sobesky J. Identiﬁcation of malignant brain edema after hemispheric stroke by
PET-imaging and microdialysis. Acta Neurochir Suppl 2003;86:237-240.
Kasner SE, Demchuk AM, Berrouschot J. Predictors of fatal brain edema in massive hemispheric
ischemic stroke. Stroke 2001;32:2117-2123.
Krieger DW, Demchuk AM, Kasner SE, Jauss M, Hantson L. Early clinical and radiological predictors
of fatal brain swelling in ischemic stroke. Stroke 1999;30:287-292.
Kucinski T, Koch C, Grzyska U, Freitag HJ, Kromer H, Zeumer H. The predictive value of early CT and
angiography for fatal hemispheric swelling in acute stroke. AJNR 1998;19:839-846.
Lee SJ, Lee KH, Na DG. Multiphasic helical computed tomography predicts subsequent development
of severe brain edema in acute ischemic stroke. Arch Neurol 2004;61:505-509.
Limburg M, van Royen EA, Hijdra A, de Bruine JF, Verbeeten BW, Jr. Single-photon emission
computed tomography and early death in acute ischemic stroke. Stroke 1990;21:1150-1155.
Manno EM, Nichols DA, Fulgham JR, Wijdicks EF. Computed tomographic determinants of neurologic
deterioration in patients with large middle cerebral artery infarctions. Mayo Clin Proc 2003;78:156160.
Maramattom BV, Bahn MM, Wijdicks EF. Which patient fares worse after early deterioration due to
swelling from hemispheric stroke? Neurology 2004;63:2142-2145.
Mori K, Aoki A, Yamamoto T, Horinaka N, Maeda M. Aggressive decompressive surgery in patients
with massive hemispheric embolic cerebral infarction associated with severe brain swelling. Acta
Neurochir (Wien) 2002;143:483-491.
Oppenheim C, Samson Y, Manai R. Prediction of malignant middle cerebral artery infarction by
diﬀusion-weighted imaging. Stroke 2000;31:2175-2181.
Ryoo JW, Na DG, Kim SS. Malignant middle cerebral artery infarction in hyperacute ischemic stroke:
evaluation with multiphasic perfusion computed tomography maps. J Comput Assist Tomogr
2004;28:55-62.
Schwab S, Spranger M, von-Kummer R, Hacke W. The ‘malignant MCA infarction’: Syndrome or
artifact? Aktuelle Neurologie 1996;23:155-162.
Stolz E, Gerriets T, Babacan SS, Jauss M, Kraus J, Kaps M. Intracranial venous hemodynamics in
patients with midline dislocation due to postischemic brain edema. Stroke 2002;33:479-485.
Thomalla GJ, Kucinski T, Schoder V. Prediction of malignant middle cerebral artery infarction by early
perfusion- and diﬀusion-weighted magnetic resonance imaging. Stroke 2003;34:1892-1899.
DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-188.
Reith J, Jorgensen HS, Pedersen PM. Body temperature in acute stroke: relation to stroke severity,
infarct size, mortality, and outcome. Lancet 1996;347:422-425.
Schwab S, Schwarz S, Spranger M, Keller E, Bertram M, Hacke W. Moderate hypothermia in the
treatment of patients with severe middle cerebral artery infarction. Stroke 1998;29:2461-2466.
97
Chapter 6
34.
35.
36.
98
Bozzao L, Bastianello S, Fantozzi LM, Angeloni U, Argentino C, Fieschi C. Correlation of angiographic
and sequential CT ﬁndings in patients with evolving cerebral infarction. AJNR 1989;10:1215-1222.
von-Kummer R, Meyding LU, Forsting M. Sensitivity and prognostic value of early CT in occlusion of
the middle cerebral artery trunk. AJNR 1994;15:9-15.
Bastianello S, Pierallini A, Colonnese C. Hyperdense middle cerebral artery CT sign. Comparison
with angiography in the acute phase of ischemic supratentorial infarction. Neuroradiology 1991;33:207211.
Algorithm for treatment allocation in small trials
Chapter 7
A new algorithm for sequential allocation of
two treatments in small clinical trials
Jeannette Hofmeijer, Peter C. Anema and Ingeborg van der Tweel
Journal of Clinical Epidemiology, under revision
99
Chapter 7
Summary
In clinical trials, patients become available for treatment sequentially. Especially
in trials with a small number of patients, loss of power may become an important
issue, if treatments are not allocated equally or if prognostic factors diﬀer between
the treatment groups. In this chapter a new algorithm for sequential allocation of
two treatments in small clinical trials is presented, to reduce both selection bias and
imbalance.
With the algorithm, an element of chance is added to the treatment decision
according to minimization. The amount of chance depends on the actual amount of
imbalance of treatment allocations of the patients already enrolled. The sensitivity to
imbalance may be tuned.
Trial simulations were performed with diﬀerent numbers of patients and prognostic
factors, in which loss of power and selection bias were quantiﬁed. With this new
method, selection bias was smaller than with minimization, and loss of power was
lower than with pure randomization or treatment allocation according to a biased
coin principle.
This new method combines the conﬂicting aims of reduction of bias by predictability
and reduction of loss of power as a result of imbalance. The method may be of use in
small trials.
100
Algorithm for treatment allocation in small trials
I
n clinical trials, patients become available for treatment sequentially. To assure
baseline comparability with regard to prognosis, the treatments under study are
usually allocated at random. However, in small trials, randomization may result in
unbalance of patient numbers or prognostic factors between treatment groups. Thus,
in patient allocation there are two important characteristics to optimize.1,2 One is the
loss of power, which will increase if treatments are not allocated equally or if prognostic factors diﬀer between the treatment groups.3 The other is the selection bias, caused
by the knowledge of the treatment that is to be allocated next.4 Both should be as small
as possible.
In pure randomization there is no selection bias, since - in case of two treatments
allocated equally - each patient has an equal probability of 1/2 of receiving the one
or the other. However, this strategy may lead to imbalance of treatments among
prognostic factors, especially in small trials.
Various randomization procedures have been developed to overcome this problem.
One method is random permuted blocks of treatment.5,6 In this design, blocks of
patients are formed and a random permutation of treatments is determined for each
block. This ensures that equal numbers of patients are achieved for each treatment.
To achieve balance with regard to prognostic factors, random permuted blocks within
strata have been recommended.6 However, this strategy restricts the number of
prognostic factors that can be accounted for in trials with a small amount of patients.
To prevent substantial imbalance between treatment groups in trials with a large
number of prognostic factors, minimization is considered an alternative approach.6-8
Minimization is exclusively concerned with the balance of treatment allocations over
various prognostic factors (or strata), since assignment of a treatment to a new patient is
determined by the treatment assignments of previous patients in the relevant stratum.
Therefore, at any moment the trial is as balanced as possible. However, minimization
involves no random process, except when treatments are equally distributed among
prognostic factors and minimization has been discouraged by experts because of
predictability.9-11 The Committee for Proprietary Medicinal Products advises that
dynamic allocation, such as minimization, should be strongly discouraged10 and the
International Conference on Harmonization Guidelines advise that ‘deterministic
dynamic allocation procedures should be avoided and an appropriate element of
randomization should be incorporated for each treatment allocation’.11
In the biased coin design as described by Efron12 the allocation probability (P)
is adapted on the basis of previous treatment allocations. The underrepresented
treatment is assigned with P > 1/2 if there are just two treatments and no prognostic
factors. This probability is ﬁxed (for example equal to 2/3) and independent of the
actual amount of imbalance between groups.
We propose an allocation method, Parameterized Dynamic Minimization (PDM),
in which the assigned probability is based on previous allocations. Unlike in Efron’s
101
Chapter 7
biased coin algorithm, the probability of allocating a certain treatment is not ﬁxed,
but depends on the actual level of imbalance of treatment allocations to the patients
already enrolled.
Example (ﬁrst part)
In a new trial for secondary prevention after transient ischemic attack or nondisabling stroke, testing treatment C versus treatment E, several prognostic factors for
cardiovascular complications might be taken into account. For example, one expects a
worse prognosis in patients with hypertension (HT) and in patients older than 65 years
of age (OLD). The prognostic factors are categorized into two or more levels, to end up
with several strata. In this example the four (2 x 2) strata are: (1) HT yes and OLD yes,
(2) HT yes and OLD no, (3) HT no and OLD yes, and (4) HT no and OLD no.
If random permuted blocks within strata are used, a separate randomization list is
prepared for each of the strata. If a patient enters the clinical trial, one has to identify
which stratum the patient is in, and obtain the next random treatment assignment
from the corresponding randomization list. The purpose is to reach approximately
equal numbers of each treatment modality for each type of patient.
It might be necessary to stratify by more than two factors. For instance, in the abovementioned trial also the prognostic factors of high cholesterol (HC) (yes or no) and
smoking (SM) (no or past or present) could be taken into account, resulting in 2 x 2
x 2 x 3 = 24 strata each requiring a separate randomization list. In small trials this
large number of strata inevitably results in imbalance of prognostic factors between
treatment groups.
In small trials with a considerable number of prognostic factors to be taken into
account, treatment groups with a similar distribution of the prognostic factors may
be accomplished by the minimization method, which is ultimately concerned with
balancing the marginal treatment totals for each level of each prognostic factor.
Table 1 shows the number of patients for each of the treatments C and E according
to each of the four prognostic factors after 80 inclusions. Suppose the next patient has
hypertension, is older than 65 years of age, has a normal cholesterol level, and does not
smoke. Then for each treatment one adds the number of patients in the corresponding
rows of the table and assigns the treatment with the smallest sum of marginal totals:
•
•
Treatment C: 30 + 18 + 9 + 19 = 76
Treatment E: 31 + 17 + 8 + 21 = 77
In this example the next patient is assigned treatment C.
Especially in small trials the lack of any random process in treatment assignment could
be a problem. The PDM method we propose adds an element of chance by assigning
102
Algorithm for treatment allocation in small trials
Table 1 Treatment assignments and distribution over four prognostic factors for 80 patients in a ﬁctitious trial
on secondary prevention of brain ischemia.
Prognostic factor
Level
Number on treatment C Number on treatment E
Next patient
HT
Yes
30
31
←
No
10
9
Yes
18
17
No
22
23
Yes
31
32
No
9
8
←
←
OLD
HC
SM
Sum of totals
No
19
21
Past
8
7
Present
13
12
76
77
←
HT indicates hypertension; OLD, older than 65 years of age; HC, high cholesterol; SM, smoking. If the next
patient has hypertension, is older than 65 years, has a normal cholesterol level, and does not smoke, then
treatment C is the treatment with the smallest sum of marginal totals. According to the minimization method,
this will be the next assigned treatment.
the treatment of choice (based on minimization) with P < 1. The assignment probability
depends on the amount of imbalance of treatment allocations of the patients already
enrolled. For the above-mentioned patient entering the ﬁctitious trial of secondary
prevention, treatment C might be assigned with P = 0.6, and treatment E with P =
0.4.
If the next patient to enter had no hypertension, was older than 65 years of age,
had no high cholesterol, and was a smoker, the sum of allocated patients for each
treatment would be:
• Treatment C: 10 + 18 + 9 + 13 = 50
• Treatment E: 9 + 17 + 8 + 12 = 46
With the larger imbalance, treatment C might be assigned with P = 0.3 and treatment
E with P = 0.7. For the ﬁrst patient entering a trial and with equal treatment numbers,
assignment is random. In the described algorithm, the sensitivity of the method for
imbalance may be tuned.
103
Chapter 7
Methods
Algorithm
The amount of imbalance di between treatment C and E within stratum i (before
allocation of the next patient) is deﬁned as
di = NCi - NEi
(1)
where NCi is the number of patients in stratum i who received treatment C, and NEi is
the number of patients in stratum i who received treatment E.
The weighted average imbalance davg is deﬁned as
davg = Σ ( wi di)
(2)
where wi is the pre-assigned relative importance of imbalance within stratum i and Σ
wi = 1.
The algorithm for the allocation probability P of a new patient to the most frequently
allocated treatment is deﬁned as
P = 0.5 α |davg|
(3)
where α is a chosen constant with which the sensitivity of the method for imbalance
may be tuned. Choosing α = 1 will result in pure randomization (P = 0.5), and α = 0
will result in purely deterministic allocation (P = 0) (Figure 1).
Since P ≤ 0.5 (Figure 1), the allocation probability of the most frequently allocated
treatment is calculated. This implies an allocation probability for the less frequently
allocated treatment according to
Pless allocated = 1 - P
(4)
Evaluation of the algorithm
We evaluated the algorithm with respect to the two characteristics ‘loss of power’ and
‘selection bias’ as speciﬁed by Atkinson.2
Loss of power
It is assumed that each of n patients receives one of two treatments C and E (denoted
as –1 and +1, respectively). A balance vector b is deﬁned as
b = FT . a
(5)
where F is a matrix in which the constant term and prognostic factors are included,
and a is the n x 1 vector of allocations of treatments + 1 and – 1. The superscript T
stands for the transpose of a vector or matrix. ‘b’ is equal to zero if all prognostic
factors are balanced across all treatments.
104
Algorithm for treatment allocation in small trials
Figure 1 Probability as a function of weighted average imbalance for diﬀerent values of α
=
P indicates allocation probability; davg indicates weighted average imbalance.
The loss of power Ln after the inclusion of n patients is deﬁned as a function of b, F, and
a:2
Ln = bT . (FT . F)-1 . b
(6)
The loss of power in absence of prognostic factors is zero when each treatment has
been allocated to an equal number of patients. If prognostic factors are present, the
loss of power also depends on the balance over the factors.
Selection bias
According to Atkinson, the selection bias Bn is quantiﬁed as follows:
Bn = (Gc – Gi) / n
(7)
where n is the number of patients included, Gc is the expected number of correct
guesses of the next allocated treatment in the n patients and Gi is the expected number
of incorrect guesses.2 With two treatments and completely deterministic allocation
the treatment is always guessed correctly, and thus Bn = 1. With completely random
allocation, the expected number of correct guesses equals the expected number of
incorrect guesses, so that Bn = 0.
In the proposed PDM, in which the probability of allocation of one of two treatments
is diﬀerent after every single inclusion, selection bias can be deﬁned as follows:
Bn = Σ (Pleast allocated – Pmost allocated) / n (8)
where the summation is over the patients included (k=1,…,n), with
Pmost allocated = 1 – Pleast allocated
and thus
Bn = Σ (2 Pleast allocated – 1) / n
(9)
where Pleast allocated is the allocation probability of the kth patient to the least allocated
treatment, and Pmost allocated is the allocation probability of the kth patient to the most
105
Chapter 7
allocated treatment. Thus quantiﬁed, Bn has the correct value of 1 in completely
deterministic and of 0 in completely random allocation.
Data analysis
Trial simulations were performed with diﬀerent numbers of patients (20 up to 200)
and prognostic factors (3 to 6, each consisting of 2 levels) and diﬀerent values of α
(0 to 1). Plots were made of the separate behavior of selection bias and loss of power,
as well as curves of loss of power against selection bias. The plots show averages of
100 simulations. Values for pure randomization (α = 1), minimization (α = 0), and
allocation according to Efron’s biased coin algorithm (P = 0.67) are indicated. Note
that for the simulations all weights wi were equal.
Results
Results of the simulations are shown in ﬁgures 2 to 4. Figure 2 shows the loss of
power. The largest losses occur in small trials with values of α close to one. Loss of
power values for α = 1 (random allocation) were even higher, but for the sake of clarity
scaling of the y-axis was not adjusted to include these values. With α closer to zero, the
eﬀect of trial size becomes smaller. With an increasing number of prognostic factors,
values for loss of power become higher (Figure 2a with three prognostic factors versus
Figure 2b with six such factors).
Figure 3 shows the selection bias. The highest selection bias occurs with values of
α approaching 0 (minimization). With an increasing number of prognostic factors,
values for selection bias become slightly higher. The eﬀect of trial size is marginal.
Figure 4 shows plots of loss of power against selection bias for diﬀerent values of α
and trial sizes of up to n = 200. Allocation starts with random allocation (α=1), with
high values of loss of power. Loss of power decreases and selection bias increases, with
lower levels of α. Loss of power is higher in smaller trials. Loss of power and selection
bias are smaller than in treatment allocation according to Efron’s biased coin. Since
both loss of power and selection bias should be as low as possible, optimal values for
α are represented in the lower left-hand corner of the plots and depend on the size of
the trial and on the number of prognostic factors.
Example (second part)
Consider the above-mentioned trial on secondary prevention after 80 allocations
to either treatment C or E according to Table 1. Suppose again the next patient has
hypertension, is older than 65 years, has normal cholesterol and does not smoke.
Treatment E is the most allocated treatment (example ﬁrst part), so allocation
of treatment E according to PDM will be as follows. The amount of imbalance is
determined as
106
Algorithm for treatment allocation in small trials
Figure 2 Loss of power as a function of trial size for trials with 3 (a), or 6 (b) strata.
a
b
Figure 3 Selection bias as a function of trial size for trials with 3 (a), or 6 (b) strata.
a
b
Figure 4 Loss of power as a function of selection bias for trials with 3 (a), or 6 (b) strata.
a
b
107
Chapter 7
dHT = 31-30, dOLD = 17-18, dHC = 8-9, and dSM = 21-19
(1)
davg = Σ ( wi di) = wHT * 1 + wOLD * -1 + wHC * -1 + wSM * 2
(2)
If imbalance within all strata is equally weighted, w = 0.25 for all strata, so
davg = 0.25-0.25-0.25+0.5 = 0.25
If α = 0.5 and PC = 1- PE
PE = 0.5 * 0.50.25 = 0.4 and PC = 0.6
(3) and (4)
If a lower level of α is chosen, the sensitivity of the method for imbalance is higher.
For example if α = 0.1 and PC = 1- PE
PE = 0.5 * 0.10.25 = 0.3 and PC = 0.7
(3) and (4)
In the other example treatment C is the most frequently allocated one. Allocation of
treatment C will be as follows. First, again, the amount of imbalance is determined.
dHT = 10-9, dOLD = 18-17, dHC = 9-8, and dSM = 13-12
(1)
davg = Σ ( wi di) = wHT * 1 + wOLD * 1 + wHC * 1 + wSM * 1 = 1
(Again using equal weights for al strata.)
(2)
If α = 0.5 and PE = 1- PC
PC = 0.5 * 0.51 = 0.25 and PE = 0.75
(3) and (4)
If α = 0.1 and PE = 1- PC
PC = 0.5 * 0.11 = 0.05 and PE = 0.95
(3) and (4)
Discussion
We developed a new method for assigning one of two treatments in clinical trials,
which is aimed to improve the balance across prognostic factors (thus reducing loss of
power) as well as to reduce bias by predictability (selection bias). Pure randomization
may result in imbalance of treatment numbers across prognostic factors, whereas
minimization is ultimately concerned with balance, but results in predictable treatment
allocation.10,11 With our new method, an amount of random chance is added to the
allocation decision based on minimization. The probability diﬀers for each included
patient, depending on the actual level of imbalance. It is not surprising that allocation
rules that are more random (α approaching 1) result in lower selection bias and larger
values of loss of power as a result of imbalance, whereas values of loss of power are
relatively low if allocation becomes more deterministic (α approaching 0).
108
Algorithm for treatment allocation in small trials
Minimization based allocation rules including a random element have been
described before.13,14 The most important diﬀerence with our proposed method is that
the random element in our algorithm is not ﬁxed, but depends on the actual level
of imbalance within a stratum at the moment of including the next patient. With a
smaller imbalance the probability of assigning one of two treatments approaches 50%.
This way bias by predictability is kept as small as possible.
With a large amount of patients to be included, diﬀerences in loss of power become less
important. Selection bias is less aﬀected by sample size. However, values for selection
bias are slightly higher with a greater number of prognostic factors, because with an
increasing number of strata, treatment allocation becomes more deterministic with
our method. In the absence of prognostic factors, 50 % of the treatment allocations are
random. With the presence of prognostic factors, this percentage becomes smaller.
Especially in trials with a small number of patients, loss of power may become an
important issue. In choosing a rule for a smaller trial, the plots presented in Figure
4 may be used. Depending on the pre-speciﬁed number of patients to include and
prognostic factors, α-values in the lower left-hand corner should be chosen, so that
both loss of power and selection bias are relatively low. Consider the above-mentioned
ﬁctitious secondary prevention trial with three prognostic factors. If the trial aims to
include 100 patients, an α value of 0.8 seems reasonable (Figure 4a). For trials with
more prognostic factors, 0.4 may be chosen, depending on the number of patients
(Figure 4b). In large trials, loss of power becomes less relevant, and the choice of α
should probably be based on selection bias. A reasonable value for most situations
seems to be 0.7 (Figure 4). In large trials pure randomization may be preferable.
In our simulations we used prognostic factors with two levels, but factors with more
than two levels can be implemented as well. In addition, the pre-assigned relative
importance of imbalance over prognostic factors (wi) was the same for all factors. If
balance across one or more speciﬁc prognostic factors is more important than balance
across other, wi can be adjusted, as long as Σ wi = 1.
Our proposed method may be of use in both blinded and unblinded trials, also
with centralized treatment allocation. For example, the PDM algorithm has been
implemented on a handheld computer and is currently being used in the multi-center
Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema
Trial (HAMLET).15 HAMLET aims to include 112 patients and takes four prognostics
factors into account.
In conclusion, the PDM allocation method may be of use in relatively small
clinical trials with stratiﬁcation among prognostic factors. This method is preferable
to minimization or pure randomization, since it combines the conﬂicting aims of
reduction of selection bias caused by predictability and reduction of loss of power
caused by imbalance of treatment numbers among the strata. Moreover, loss of power
is smaller than in treatment allocation according to Efron’s biased coin. The sensitivity
109
Chapter 7
to imbalance may be tuned with α. The choice of α depends on the number of patients
to be included and the number of pre-speciﬁed prognostic factors.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
110
Atkinson AC. Optimum biased-coin designs for sequential treatment allocation with covariate
information. Stat Med 1999 30;18:1741-1752.
Atkinson AC. The comparison of designs for sequential clinical trials with covariate information.
Journal of the Royal Statistical Society, Series A 2002;165:349-373.
Atkinson AC. The distribution of loss in two-treatment biased-coin designs. Biostatistics 2003;4:179193.
Blackwell D, Hodges JL. Design for the control of selection bias. Annals of mathematical statistics
1957;28:449-460.
Matts JP, Lachin JM. Properties of permuted-block randomization in clinical trials. Control Clin Trials
1988;9:327-344.
Pocock SJ. Methods of Randomization. Clinical Trials: A Practical Approach. Chichester: John Wiley
& Sons, 1983:66-89.
Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors. Biometrics
1975;31:103-115.
Taves DR. Minimization: a new method of assigning patients to treatment and. Clin Pharmacol Ther
1974;15:443-453.
International Conference on Harmonisation of Technical Requirements for Registration of
Pharmaceuticals for Human Use (ICH) adopts Consolidated Guideline on Good Clinical Practice in
the Conduct of Clinical Trials on Medicinal Products for Human Use. Int Dig Health Legis 1997;48:231234.
Committee for Proprietary Medicinal Products (CPMP): points to consider on adjustment for baseline
covariates. Stat Med 2004;15:701-709.
ICH Harmonised Tripartite Guideline. Statistical principles for clinical trials. International Conference
on Harmonisation E9 Expert Working Group. Stat Med 1999;15:1905-1942.
Efron B. Forcing a sequential experiment to be balanced. Biometrika 1971; 58 :403-417.
Scott NW, McPherson GC, Ramsay CR, Campbell MK. The method of minimization for allocation to
clinical trials. a review. Control Clin Trials 2002;23:662-674.
Brown S, Thorpe H, Hawkins K, Brown J. Minimization-reducing predictability for multi-centre trials
whilst retaining balance within centre. Stat Med 2005;30:3715-3727.
Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Macleod MR, Kappelle LJ, van der Worp HB.
Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema Trial
(HAMLET). Protocol for a randomised controlled trial of decompressive surgery in space-occupying
hemispheric infarction. Trials 2006;7:29.
HAMLET protocol
Chapter 8
Hemicraniectomy After Middle cerebral
artery infarction with Life-threatening Edema
Trial (HAMLET). Protocol of a randomized
controlled trial of decompressive surgery in
space-occupying hemispheric infarction
Jeannette Hofmeijer, G. Johan Amelink, Ale Algra, Jan van Gijn, Malcolm R. Macleod, L. Jaap Kappelle,
and H. Bart van der Worp for the HAMLET investigators
Based on Trials 2006;11:7-29
111
Chapter 8
Summary
Patients with a hemispheric infarct and massive space-occupying brain edema have
a poor prognosis. Despite maximal medical treatment, the case fatality rate may
be as high as 80%, and most survivors are left severely disabled. Non-randomized
studies have suggested that decompressive surgery reduces mortality substantially
and improves functional outcome of survivors. The Hemicraniectomy After Middle
cerebral artery infarction with Life-threatening Edema Trial (HAMLET) is designed
to compare the eﬃcacy of decompressive surgery to improve functional outcome
with that of conservative treatment in patients with supratentorial infarction, who
deteriorate as a result of brain edema.
HAMLET is a multi-center, randomized clinical trial, which will include 112 patients
aged between 18 and 60 years with a large hemispheric infarct with space-occupying
edema that leads to a decrease in consciousness. Patients will be randomized to
receive either decompressive surgery in combination with medical treatment or best
medical treatment alone. Randomization will be stratiﬁed for the intended mode of
conservative treatment (intensive care or stroke unit care). The primary outcome
measure will be functional outcome, as determined by the score on the modiﬁed
Rankin Scale, at one year.
HAMLET started recruiting September 1st 2002 (Current Controlled Trials
ISRCTN94237756). As of January 1st 2006, 58 patients had been included.
112
HAMLET protocol
L
arge cerebral infarcts are commonly associated with variable degrees of brain
edema. In severe cases, this may lead to transtentorial or uncal herniation.
Serious edema formation usually manifests itself between the second and ﬁfth
day after stroke onset.1-4 Brain tissue shifts rather than raised intracranial pressure
(ICP) are probably the most likely cause of the initial decrease in consciousness.4,5
Fatal space-occupying brain edema occurs in 1 to 5% of patients with a supratentorial
infarct.6,7 However, in younger patients with ischemic stroke, herniation accounts for
about half of the deaths in the ﬁrst month.8
Patients with a hemispheric infarct and massive space-occupying brain edema have a
poor prognosis: in recent intensive care (IC)-based prospective series, the case fatality
rate was about 80%, despite maximal conservative therapy.9,10 Several conservative
treatment strategies have been proposed to limit brain tissue shifts and reduce ICP,
such as sedation with barbiturates or propofol, hyperventilation, and osmotic therapy
with glycerol, mannitol, or hypertonic saline hydroxyethyl starch (HES).11-13 However,
no trials have addressed the eﬃcacy of these therapies to improve clinical outcome14
and several reports suggest that these are ineﬀective9,10,15 or even detrimental.16,17 The
value of ICP monitoring has also not been established. ICP monitoring may reduce
iatrogenic errors,18 but is probably not helpful in guiding long-term treatment.5 Despite
this lack of evidence hyperventilation, osmotherapy, and the use of ICP monitors are
recommended and used by experts for patients with ischemic stroke whose condition
is deteriorating secondary to edema formation.11,12,19
Because of the limitations of medical therapies, there have been proposals for
decompressive surgery in patients with neurological deterioration due to large
hemispheric infarction and edema. The rationale of this therapy is to prevent brain
tissue shifts and to normalize intracranial pressure, and thereby to preserve cerebral
blood ﬂow and to prevent secondary damage.19 The technique of decompressive
surgery is relatively simple and consists of a large hemicraniectomy and a duraplasty.
It can be performed in every neurosurgical center.
Animal studies have shown that decompressive surgery reduces mortality and
improves histological and functional outcome.20,21 Case reports and small retrospective
studies have suggested that this intervention lowers mortality without increasing
the rate of severely disabled survivors.19,22,23 These ﬁndings were supported by two
prospective series. Mortality was reduced from about 80% in controls to 34% and 16%
in surgically treated patients, respectively, and poor functional outcome from 95% to
50%.24,25 Reports from other studies suggest that decompressive surgery is less eﬀective
in elderly patients,26,27 and that substantial recovery extends into the second half year
and thereafter.28
Although the above results suggest a very substantial beneﬁt of decompressive
surgery, the studies had too many ﬂaws to be readily translated into clinical practice.
Most importantly, the groups were not constituted by random selection. The control
groups consisted of patients with a signiﬁcantly higher age, more co-morbidity,
113
Chapter 8
and more frequent lesions in the dominant hemisphere than those in the surgical
groups.24,25 In addition, information on functional outcome of the surviving patients
was insuﬃcient.24,25 The aim of the current study is to test if decompressive surgery
improves functional outcome in patients with neurological deterioration as a result
of large hemispheric infarction and edema, as compared with ‘standard’ best medical
care, on either a stroke unit (SU) or an ICU, in a randomized controlled clinical trial.
Methods and design
HAMLET is a multi-center, open, randomized treatment trial with masked outcome
assessment. Participating centers have adequate experience with the management of
acute ischemic stroke and intensive care treatment of patients with an elevated ICP,
and neurosurgical facilities are available on a 24-hours/day basis. The trial protocol
has been approved by the medical ethical committees of all participating centers.
Enrollment criteria
Patients can be enrolled in the study if they have 1. a diagnosis of acute ischemic
stroke in the territory of the middle cerebral artery, with an onset within 96 hours
prior to the planned start of the trial treatment. Within this timeframe, it is advised
to start treatment as soon as possible; 2. a score on the National Institutes of Health
Stroke Scale (NIHSS)29 ≥ 16 for right-sided lesions or ≥ 21 for left-sided lesions; 3. a
gradual decrease in consciousness to a score of 13 or lower on the Glasgow Coma
Scale (GCS) for right-sided lesions, or an eye and motor score of 9 or lower for leftsided lesions; 4. hypodensity on CT involving two thirds or more of the territory of the
middle cerebral artery (MCA), and space-occupying edema formation. (displacement
of midline structures on CT is not a requirement for inclusion); and 5. age 18 up to
and including 60 years. In addition, there must be a possibility to start trial treatment
within 3 hours after randomization, and written informed consent must be obtained
from a representative of the patient.
Patients will be excluded from the study if they have 1. ischemic stroke of the entire
cerebral hemisphere (anterior, middle, and posterior cerebral artery territories); 2. a
decrease in consciousness at least partially explained by a cause other than edema
formation, such as metabolic disturbances or medication; 3. two ﬁxed dilated pupils;
4. been treated with a thrombolytic agent in the 12 hours preceding randomization;
5. a known systemic bleeding disorder; 6. a pre-stroke score on the modiﬁed Rankin
Scale (mRS)30 of more than 1 or on the Barthel Index (BI)31 of less than 95; 7. a life
expectancy < 3 years; or 8. a serious illness that may confound treatment assessment.
Informed consent
Written informed consent for this study will be obtained from the patient’s authorized
114
HAMLET protocol
representative prior to the performance of any protocol-speciﬁc procedure. However,
several of these assessments or tests may be performed as part of the patient’s
routine clinical evaluation (i.e. not speciﬁcally performed for this trial). The study
will be conducted in accordance with the provisions of the Declaration of Helsinki, as
amended in South Africa (1996).
Treatment allocation
After informed consent is obtained, patients will be randomized to either surgical or
conservative treatment. Randomization will be stratiﬁed for the intended mode of
conservative treatment (intensive care or stroke unit care; see below). The choice of
conservative treatment is left at the discretion of the local investigator, and will usually
be the standard mode of treatment in the participating center. Allocation to treatment
will be made via a telephone call to a 24 hour randomization service. Ideally, each
treatment arm will consist of 56 patients. The total number of patients included in the
trial will be 112.
I. decompressive surgery
Decompressive surgery will consist of a large hemicraniectomy and a duraplasty. In
summary, a large (reversed) question mark-shaped skin incision based at the ear will
be made. A bone ﬂap with a diameter of at least 12 cm (including parts of the frontal,
parietal, temporal, and occipital squama) will be removed. Additional temporal bone
will be removed so that the ﬂoor of the middle cerebral fossa can be explored. The
dura will be opened and an augmented dural patch will be inserted. The dura will be
ﬁxed at the margins of the craniotomy to prevent epidural bleeding. The temporal
muscle and the skin ﬂap will then be re-approximated and secured. Infarcted brain
tissue will not be resected. A sensor for registration of intracranial pressure may be
inserted. In surviving patients, cranioplasty will be performed after at least six weeks,
using the stored bone ﬂap or acrylate. After surgery, patients will be transferred to an
IC unit. Anti-edema therapy as described under ‘IIa’ may be used but will usually not
be necessary.
IIa. intensive care
Because no mode of IC treatment has been proven superior to the others, treatment
options may vary depending on local traditions. However, to improve consistency of
treatment between centers, the following treatment modalities are recommended:
a. Osmotherapy. Osmotherapy should be started as soon as possible after
randomization. The use of mannitol or glycerol is recommended in a dose suﬃcient
to reach a serum osmolality of 315 to 320 mOsm.
b. Intubation and mechanical ventilation. Patients will be intubated if the score on the
GCS is lower than or equal to 10, or if there are signs of respiratory insuﬃciency,
115
Chapter 8
c.
d.
e.
f.
g.
h.
i.
such as a pO2 ≤ 60 mm Hg, a pCO2 ≥ 60 mm Hg, or if the airway is compromised.
However, earlier intubation is left at discretion of the treating physician. Mechanical
ventilation with use of intermittent mandatory or assist/control ventilation will
often be suﬃcient, especially in the early phase.
Hyperventilation. The use of hyperventilation is discouraged. If hyperventilation
is started, it is advised to monitor venous oxygen saturation with jugular bulb
oxymetry and to maintain oxygen saturation higher than 50%. If venous oxygen
saturation is not monitored, the pCO2 may be reduced to 28 - 32 mm Hg.
ICP monitoring. Invasive monitoring of the intracranial pressure is left at the
discretion of the treating physician. If used, the ICP monitor should preferably be
inserted ipsilateral to the side of the infarct.
Sedation. If mechanical ventilation requires sedation, and in the case of further
neurological deterioration, or an uncontrolled increase in ICP, patients may be
sedated with propofol. The use of barbiturates is discouraged, because this may
reduce cerebral perfusion pressure and does often not lead to a sustained control
of ICP. If necessary, muscle relaxants may be used.
Blood pressure control. In general, spontaneous elevation of blood pressure to a
level of 220/120 mm Hg should be accepted. Sustained higher blood pressures can
be treated with labetolol or nitroprusside. In case of sustained hypotension or a
critical reduction of cerebral perfusion pressure catecholamines may be used.
Elevation of the head (30°) to optimize venous drainage.
Aim for normothermia. Antipyretics and a cooling blanket may be used.
Aim for normoglycemia. Consider the use of insulin.
IC treatment should be continued at least until day 5 after stroke onset, or until there
is clinical or radiological improvement deemed to be suﬃcient to transfer the patient
back to the stroke unit or medium care unit.
IIb. stroke unit care
The patient will be admitted to a stroke unit or medium care unit and is treated
according to local practice, supplemented with the treatment options that can be
performed in a medium care unit as described above under ‘IIa’.
Treatment strategy ‘I’ is the experimental option, strategy ‘IIa’ is recommended by
international experts, and strategy ‘IIb’ is the most common clinical practice in the
Netherlands, and probably also in a signiﬁcant number of other European countries.
The choice of ancillary care measures, e.g., prophylactic administration of (lowmolecular-weight) heparin and the treatment of complicating illnesses will be at the
discretion of the treating physician. However, administration of low-dose aspirin is
recommended, whereas treatment with intravenous heparin and hemodilution with
plasma expanders are discouraged.
116
HAMLET protocol
Outcome assessment
The primary outcome measure will be functional outcome, as determined by the score
on the mRS, at one year. Outcome will be dichotomized as ‘good’ (mRS 0 to 3) or
poor (mRS 4 to dead). The score on the mRS will be determined in a standard way
independently by three blinded investigators, on the basis of a narrative written by
an unblinded independent study nurse, if necessary followed by a consensus meeting.
Secondary analyses will be performed in which ‘good outcome’ will be deﬁned as mRS
0 to 2, and ‘poor outcome’ as mRS 3 to dead, and with case fatality as the measure
of outcome. Other outcome measures will include the scores on the NIHSS, the BI,
the Montgomery and Asberg Depression Rating Scale (MADRS),32 quality of life as
measured with the SF36, and quality of life measured with a visual analogue scale
(VAS)33 at one year and at three years. In addition, the mRS, NIHSS, and BI will be
determined at 3 and 6 months.
Data collection
At baseline, the medical history will be assessed and a general and neurological
examination will be carried out. The scores on the GCS and NIHSS will be used to
assess stroke severity on admission and directly prior to randomization. Retrospective
scoring of the NIHSS on admission is reliable when a published algorithm is used.34
Date and time of stroke onset and of neurological deterioration (i.e. GCS ≤ 13) will be
noted. Laboratory investigations will include a blood cell count, electrolytes, serum
glucose, creatinine, and liver enzymes. Brain CT or MRI scans will be performed on
admission and prior to randomization. Lateral displacement of the pineal gland and
the septum pellucidum will be measured by the trial co-ordinator. The GCS, pupillary
reﬂexes, body temperature, and blood pressure will be assessed every 4 hours during
the ﬁrst 4 days after randomization. If ICP is recorded, ICP and cerebral perfusion
pressure (CPP) will be noted at 2-hour intervals. Dose and time of administration of
osmotic agents and sedatives, concurrent medication, and neurological and systemic
adverse events will be recorded. The GCS, NIHSS, mRS, and BI will be determined at
day 14. CT or MRI of the brain will be performed at day 7 ± 2 days for measurement of
infarct volume35,36 and lateral displacement of the pineal gland and septum pellucidum.
At three and six months, the mRS, BI, and NIHSS will be assessed. Serious adverse
events and the number of days the patient was admitted on an IC unit, in the hospital,
and / or rehabilitation center or chronic nursing home will be recorded. At one and
three years, the mRS, BI, MADRS, VAS, and SF36 will be assessed by the research
fellow or data manager. He / she will perform a standardized evaluation of each
patient, including functional status, and will include the obtained information in an
essay. Based on this essay, three members of the executive committee, who are blinded
to treatment allocation, will determine the scores on the mRS and BI.
117
Chapter 8
Statistical aspects
In the primary analysis, outcome of all patients in the surgical group will be compared
with that of all patients in the conservative treatment group. However, subgroup
analyses will be performed based on the mode of conservative treatment (intensive
care vs. stroke unit care), side of the lesion, and time of randomization (within 48
hours versus between 48 and 96 hours). The sample size is based on the desire that in
each of the two subgroups the superiority of surgical decompression over conservative
treatment can be proved. All primary analyses will be performed on an intention-totreat basis, but additional on-treatment analyses will also be performed. An interim
analysis will be performed after the one-year follow-up of the ﬁrst 30 patients.
Power calculations are hampered by the scarcity of data on clinical outcome in this
patient group. In the non-randomized trial by Rieke et al., which is most comparable
to the proposed study regarding patient group and treatment, 95% (20/21) of the
patients in the IC treatment group had a poor outcome, compared with 50% (16/32) in
the surgically treated group.24 There are no data on outcome after ‘standard’ treatment,
but this is not likely to be better than IC treatment. Based on an α of 0.05 and a β
of 0.20, extrapolation of these data suggests that only 12 patients would have to be
included in each group to demonstrate the superiority of surgical decompression over
conservative treatment in a randomized trial. Assuming that 60% of the patients in
the surgery group will have a poor outcome and 85% of the patients in each of the
two conservative treatment subgroups, a group size of 28 patients would suﬃce to
prove the superiority of decompressive surgery. Applying the same assumption to the
complete treatment, a group size of 56 patients is needed.
Safety
In the conservative treatment group, patients will receive the standard care of the
center to which they have been admitted. For this reason, the components of the
conservative treatment strategies will not be considered experimental, and need no
detailed explanation in patient information letters.
Decompressive surgery consisting of a hemicraniectomy and a duraplasty may
be complicated by the occurrence of local postoperative bleeding. In the two open,
prospective studies of decompressive surgery for space-occupying hemispheric
infarction, including a total of 63 patients, 3 patients (5%; 95% CI 0 to 13) experienced
an epidural hematoma, one (2%; 95% CI 0 to 9) a subdural hematoma, and three (5%;
95% CI 0 to 13) a space-occupying subarachnoid hygroma over the trepanation site.
None of these complications led to additional neurological deﬁcit.9,25
An adverse event is any unfavorable and unintended sign, symptom, or disease
occurring during the follow-up period of the study. Adverse events occurring after
randomization will be recorded on the adverse event page of the CRF. A serious
118
HAMLET protocol
adverse event is deﬁned as any adverse event that results in death, a life-threatening
condition, inpatient hospitalization or prolongation of existing hospitalization, or
persistent or signiﬁcant disability / incapacity. An important medical event that may
not result in one of the above conditions may be considered a serious adverse event
when, based upon medical judgment, it may jeopardize the patient and may require
medical or surgical intervention to prevent one of the outcomes above. A reasonablyrelated adverse event is deﬁned as one that is possibly, probably, or deﬁnitely related
to the trial treatment. Adverse events that are serious and reasonably related to the
trial treatment, and all deaths, require completion of the safety report, which should
be faxed to the trial co-ordination center within 5 working days of observation or
notiﬁcation of the event. The Data Monitoring Committee performs analyses of the
unblinded interim data and formulates recommendations for the Steering Committee
on the continuation of the trial. The Data Monitoring Committee may also oﬀer
unsolicited recommendations.
Publication of the trial results
The trial results will be published by the members of the Executive Committee, on
behalf of all HAMLET investigators. Before submission of any manuscript all local
principal investigators will have the opportunity to comment on the manuscript.
Discussion
We present the protocol of a randomized clinical trial designed to test the eﬃcacy
of decompressive surgery to improve functional outcome in patients with spaceoccupying hemispheric infarction. Case reports, retrospective studies and two trials
have suggested that this intervention lowers mortality without increasing the rate
of severely disabled survivors,19,22-25 but this has never been tested in a randomized,
controlled clinical trial.
The timing of surgery and the choice of enrolment criteria form the crux of the
present trial. It has been suggested that decompressive surgery is most eﬃcacious if
performed within the ﬁrst 24 hours after stroke onset and before the occurrence of
clinical signs of herniation,25 but this may carry the risk of inclusion of patients who
do not need such an invasive procedure for survival and recovery. In addition, in a
systematic review of studies on decompressive surgery in space-occupying hemispheric
infarction, the timing of surgery did not aﬀect outcome.27
In diﬀerent studies, several clinical and radiological parameters have been identiﬁed
as independent early predictors of fatal space-occupying edema formation: nausea
and vomiting within 24 hours after stroke onset,37 a systolic blood pressure of 180
mm Hg or more at 12 hours after onset,37 a history of hypertension or heart failure,38
elevated white blood cell count,38 an activity deﬁcit covering the complete MCA
119
Chapter 8
territory on SPECT imaging,39 a hypodensity of more than 50% on the initial CT
scan,37,38 attenuated corticomedullary contrast on CT,40 involvement of additional
vascular territories,38 and volume of diﬀusion-weighted imaging (DWI) abnormalities
of more than 145 ml.41 In addition, in the Lubeluzole-International-9 trial it was found
that the minimum baseline score in patients who ultimately died from brain swelling
was 15 for right-sided lesions and 20 for left-sided lesions.37 Unfortunately, all studies
were relatively small and the value of the suggested predictors has not been conﬁrmed
in other prospective series.
With the proposed timing of therapy and choice of enrolment criteria we expect to
largely prevent inclusion of patients who cannot beneﬁt from hemicraniectomy, either
because they do not need decompressive surgery since their disease will follow a more
benign course, or because of extensive and irreversible damage. In the Heidelberg
studies, complete global aphasia was an exclusion criterion for decompressive surgery,
and the majority of patients receiving surgical therapy had a lesion in the right
hemisphere.24,25 In our study, patients with a large infarct in the dominant hemisphere
and severe aphasia may be included, as it is unproven that they will fare worse than
patients with an infarct in the non-dominant hemisphere. In fact, it has been shown
that with the exception of the ability to communicate, the long-term quality of life of
patients with left-sided lesions is slightly better than that of patients with a lesion in
the right hemisphere.42
In this trial, a fully blinded evaluation of the functional outcome of patients is virtually
impossible, as patients in the decompressive surgery group will be easily recognized
by surgical marks. Even in case of follow up by telephone, patients are likely to note
the sequels of surgery. However, suﬃcient blinding will be assured by the indirect
assessment of scores on the functional outcome scales, as described above.
The trial started September 1st 2002. As of January 1st 2006, 58 patients were
included in seven participating centers in the Netherlands. Two other Dutch centers
and seven centers in the UK are expected to join the trial shortly. Other centers that
have adequate experience with the management of acute ischemic stroke and intensive
care treatment of patients with an elevated ICP, and that have neurosurgical facilities
available on a 24-hours/day basis are welcome to participate.
References
1.
2.
3.
120
Shaw C, Alvord Jr E, Berry E. Swelling of the brain following ischemic infarction with arterial occlusion.
Arch Neurol 1959;1:161-177.
Ropper AH, Shafran B. Brain edema after stroke. Clinical syndrome and intracranial pressure. Arch
Neurol 1984;41:26-29.
Wardlaw J, Dennis M, Lindley R, Warlow C, Sandercock P, Sellar R. Does early reperfusion of a cerebral
infarct inﬂuence cerebral infarct swelling in the acute stage or the ﬁnal clinical outcome? Cerebrovasc
Dis 1993;3:86-93.
HAMLET protocol
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Frank JI. Large hemispheric infarction, deterioration, and intracranial pressure. Neurology
1995;45:1286-1290.
Schwab S, Aschoﬀ A, Spranger M, Albert F, Hacke W. The value of intracranial pressure monitoring
in acute hemispheric stroke. Neurology 1996;47:393-398.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
Heinsius T, Bogousslavsky J, Van Melle G. Large infarcts in the middle cerebral artery territory.
Etiology and outcome patterns. Neurology 1998;50:341-350.
Biller J, Adams-HP J, Bruno A, Love BB, Marsh EE. Mortality in acute cerebral infarction in young
adults--a ten-year experience. Angiology 1991;42:224-230.
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 1996;53:309-315.
Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction and early brain swelling:
progression and eﬀect of age on outcome. Mayo Clin Proc 1998;73:829-836.
Adams HP, Jr., Adams RJ, Brott T. Guidelines for the early management of patients with ischemic stroke:
A scientiﬁc statement from the Stroke Council of the American Stroke Association. Stroke
2003;34:1056-1083.
Wijdicks EF. Management of massive hemispheric cerebral infarct: is there a ray of hope? Mayo Clin
Proc 2000;75:945-952.
Schwarz S, Schwab S, Bertram M, Aschoﬀ A, Hacke W. Eﬀects of hypertonic saline hydroxyethyl starch
solution and mannitol in patients with increased intracranial pressure after stroke. Stroke 1998;29:15501555.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral infarction. Crit
Care Med 2003;31:617-625.
Schwab S, Spranger M, Schwarz S, Hacke W. Barbiturate coma in severe hemispheric stroke: useful or
obsolete? Neurology 1997;48:1608-1613.
Muizelaar JP, Marmarou A, Ward JD. Adverse eﬀects of prolonged hyperventilation in patients with
severe head injury: a randomized clinical trial. J Neurosurg 1991;75:731-739.
Kaufmann AM, Cardoso ER. Aggravation of vasogenic cerebral edema by multiple-dose mannitol. J
Neurosurg 1992;77:584-589.
Ropper AH. Neurological intensive care. Ann Neurol 1992;32:564-569.
Schwab S, Rieke K, Aschoﬀ A, Albert F, von Kummer R, Hacke W. Hemicraniotomy in space-occupying
hemispheric infarction: useful early intervention or desparate activism? Cerebrovasc Dis 1996;6:325329.
Forsting M, Reith W, Schabitz WR. Decompressive craniectomy for cerebral infarction. An
experimental study in rats. Stroke 1995;26:259-264.
Doerﬂer A, Forsting M, Reith W. Decompressive craniectomy in a rat model of “malignant” cerebral
hemispheric stroke: experimental support for an aggressive therapeutic approach. J Neurosurg
1996;85:853-859.
Carter BS, Ogilvy CS, Candia GJ, Rosas HD, Buonanno F. One-year outcome after decompressive
surgery for massive nondominant hemispheric infarction. Neurosurgery 1997;40:1168-1175.
Koh MS, Goh KY, Tung MY, Chan C. Is decompressive craniectomy for acute cerebral infarction of
any beneﬁt? Surg Neurol 2000;53:225-230.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
121
Chapter 8
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
122
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Holtkamp M, Buchheim K, Unterberg A. Hemicraniectomy in elderly patients with space occupying
media infarction: improved survival but poor functional outcome. J Neurol Neurosurg Psychiatry
2001;70:226-228.
Gupta R, Connolly ES, Mayer S, Elkind MS. Hemicraniectomy for massive middle cerebral artery
territory infarction: a systematic review. Stroke 2004;35:539-543.
Manai R, Srour A, Crozier S, Vandamme X, Samson Y, Cornu P, Rancurel G. Long-term functional
outcome of hemicraniectomy in middle cerebral artery malignant infarcts [abstract]. J Neurol
2001;248(suppl 2):121-122.
Brott T, Adams HP, Jr., Olinger CP. Measurements of acute cerebral infarction: a clinical examination
scale. Stroke 1989;20:864-870.
Bamford JM, Sandercock PA, Warlow CP, Slattery J. Interobserver agreement for the assessment of
handicap in stroke patients. Stroke 1989;20:828.
Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J 1965;14:61-65.
Montgomery S, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry
1979;134:382-389.
Indredavik B, Bakke F, Slordahl SA, Rokseth R, Haheim LL. Stroke unit treatment improves long-term
quality of life: a randomized controlled trial. Stroke 1998;29:895-899.
Williams LS, Yilmaz EY, Lopez-Yunez AM. Retrospective assessment of initial stroke severity with the
NIH Stroke Scale. Stroke 2000;31:858-862.
Brott T, Marler JR, Olinger CP. Measurements of acute cerebral infarction: lesion size by computed
tomography. Stroke 1989;20:871-875.
van der Worp HB, Claus SP, Bar PR. Reproducibility of measurements of cerebral infarct volume on
CT scans. Stroke 2001;32:424-430.
Krieger DW, Demchuk AM, Kasner SE, Jauss M, Hantson L. Early clinical and radiological predictors
of fatal brain swelling in ischemic stroke. Stroke 1999;30:287-292.
Kasner SE, Demchuk AM, Berrouschot J. Predictors of fatal brain edema in massive hemispheric
ischemic stroke. Stroke 2001;32:2117-2123.
Berrouschot J, Barthel H, von-Kummer R, Knapp WH, Hesse S, Schneider D. 99m technetium-ethylcysteinate-dimer single-photon emission CT can predict fatal ischemic brain edema. Stroke
1998;29:2556-2562.
Haring HP, Dilitz E, Pallua A. Attenuated corticomedullary contrast: An early cerebral computed
tomography sign indicating malignant middle cerebral artery infarction. A case-control study. Stroke
1999;30:1076-1082.
Oppenheim C, Samson Y, Manai R. Prediction of malignant middle cerebral artery infarction by
diﬀusion-weighted imaging. Stroke 2000;31:2175-2181.
de Haan RJ, Limburg M, Van der Meulen JH, Jacobs HM, Aaronson NK. Quality of life after stroke.
Impact of stroke type and lesion location. Stroke 1995;26:402-408.
Informed consent in HAMLET
Chapter 9
Appreciation of the informed consent
procedure in a randomized trial of
decompressive surgery for space-occupying
hemispheric infarction
Jeannette Hofmeijer, G. Johan Amelink, Helena M. den Hertog, Ale Algra, L. Jaap Kappelle, and H. Bart
van der Worp, on behalf of the HAMLET and the PAIS investigators
Journal of Neurology, Neurosurgery, and Psychiatry, under revision
123
Chapter 9
Summary
The eﬃcacy of surgical decompression to improve functional outcome in patients
with space-occupying hemispheric infarction is currently studied in the randomized
Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema
Trial (HAMLET). As non-randomized studies have suggested that this treatment may
reduce mortality, randomization may be considered unethical. In this chapter it is
studied how representatives of the patients had experienced the informed consent
procedure and what information they could recall in comparison with representatives
of patients participating in the randomized trial of Paracetamol (Acetaminophen) In
Stroke (PAIS).
One year after study inclusion, 60 consecutive representatives who had given
informed consent for participation of their relative in HAMLET or PAIS were
contacted. Appreciation of the informed consent procedure and recall of trial details
were investigated with standardized questionnaires and compared between the two
groups.
For HAMLET and PAIS 28 and 30 representatives were interviewed, respectively.
86% of HAMLET representatives and 40% of PAIS representatives remembered
participation of their relative in a clinical trial (p< 0.001). HAMLET representatives
remembered more trial details. With respect to appreciation of the informed
consent procedure there were no diﬀerences between the groups: in each trial four
representatives had considered the question of randomization unacceptable.
Recall of information about a clinical trial by representatives of patients with a
neurological emergency is better if the potential implications of the treatment under
study are greater. Participation of patients in a randomized controlled trial of surgical
decompression for space-occupying infarction is generally considered acceptable by
their legal representatives.
124
Informed consent in HAMLET
I
nformed consent from patients or their representatives is essential in any clinical trial. In studies of acute stroke,1 myocardial infarction,2 or other diseases,3-8
the amount of information of the informed consent procedure that study participants could understand and recall has varied. If treatment and thereby study inclusion
require urgency because of an acute life-threatening situation, it may be diﬃcult to
obtain truly informed consent.
The Hemicraniectomy After Middle cerebral artery infarction with Life-threatening
Edema Trial (HAMLET) is an ongoing randomized controlled trial to study the eﬀect
of decompressive surgery on functional outcome in patients with middle cerebral
artery (MCA) infarction and life-threatening edema formation.9 Non-randomized
studies have suggested that this intervention may reduce mortality from 78% to 50%
or 16%.10,11 The inclusion criteria for participation in HAMLET require a reduced
level of consciousness. Consequently, informed consent is asked from patients’ legal
representatives. Since patients can deteriorate within hours, there is little time to
contemplate participation. The life-threatening condition and time pressure may
compromise the ability for relatives to adequately perceive all relevant details of the
trial. Moreover, because of the reputed reduction of mortality caused by the treatment
under study, randomized treatment allocation may be considered unethical and
unacceptable by the patients’ legal representatives.
Legal representatives of patients included in HAMLET were interviewed to ascertain
what information about the trial they could recall and how they had experienced the
informed consent procedure. To study the impact of the life-threatening condition and
the invasive experimental treatment, also legal representatives of patients included in a
trial in which less vital issues are at stake, the randomized Paracetamol (Acetaminophen)
In Stroke (PAIS) trial,12 were interviewed and results were compared.
Methods
Subjects
All legal representatives of patients who had given informed consent for participation
in HAMLET at the University Medical Center in Utrecht or the Academic Medical
Center in Amsterdam, the Netherlands, between 1 September 2002 and 1 August
2005, and those who had given informed consent for patients included in PAIS
between 1 January 2005 and 1 January 2006 at the same hospital in Utrecht and the
Erasmus Medical Center in Rotterdam, the Netherlands were contacted. In PAIS,
informed consent should be obtained from the patient but may also be given by a legal
representative in case of a reduced consciousness or aphasia. We interviewed the ﬁrst
30 representatives who could be contacted from both trials. The interviews took place
approximately 1 year after inclusion. Both trial protocols had been approved by the
ethics committees of the participating hospitals.
125
Chapter 9
Trial physicians had sought written informed consent as soon as possible after a
reduction in consciousness caused by edema formation (HAMLET) or after admission
because of acute stoke (PAIS). The time window for inclusion in HAMLET is 96 hours
after stroke onset, whereas this is 12 hours in PAIS. However, in both trials treatment
is required to start as soon as possible after neurological deterioration or stroke onset,
respectively. The information given about each of the trials included an explanation
about the disease and the intervention under study, an explanation of the possible
beneﬁts and side eﬀects of the treatment, the controlled design, the voluntary nature
of participation, the duration of the trial, and the right to withdraw at any moment.
Data collection
Legal representatives who had originally given consent were interviewed by telephone
by one investigator (JH). The ﬁrst question was whether the representative could recall
having given informed consent for participation of his / her relative in a clinical trial.
If not, the interviewer disclosed that the relative had participated in a trial by referring
to the meeting with a study physician shortly after clinical deterioration (HAMLET)
or admission to the hospital (PAIS). Second, with a standard questionnaire (tables
1 and 2), recall of trial details and appreciation of the informed consent procedure
were investigated. Finally, representatives were asked whether they would again give
permission for participation in a clinical trial under similar circumstances.
Data analysis
Answers from representatives of patients participating in HAMLET and PAIS are
presented as frequencies (n) and proportions (%). A Chi square test was used for
testing diﬀerences between the groups. Diﬀerences in age were tested by means of
Student’s t-test and diﬀerences in initial stroke severity as expressed by the score on
the National Institutes of Health Stroke Scale (NIHSS)13 with the Mann-Whitney U
test. P < 0.05 was considered statistically signiﬁcant.
Results
Of 30 representatives of HAMLET patients contacted, 28 were interviewed;
one representative refused to participate and one spoke Dutch insuﬃciently to
communicate by telephone. Of 30 representatives of PAIS patients we contacted, all
were interviewed. The median time between the informed consent procedure and
the interview did not diﬀer between the groups (HAMLET, 13 months (range 10-16);
PAIS, 13 months (range 9-17)). As expected, because of the maximum age limit of 60
years in HAMLET, mean age of the patients at study inclusion was signiﬁcantly lower
in HAMLET (48 years ± 8) than in PAIS (69 years ± 13; mean diﬀerence 22, 95% CI
126
Informed consent in HAMLET
Table 1 Recall of trial details.
HAMLET (n (%))
PAIS (n (%))
n = 28
n = 30
P
Do you remember your family member having participated in a study?
Yes
23 (82)
21 (70)
No
5 (18)
9 (30)
0.22
Do you remember the treatment under study? *
Correct
24 (86)
12 (40)
Incorrect
0 (0)
2 (7)
Don’t know
4 (14)
16 (53)
0.001
Do you remember the presumed eﬀects of the treatment under study? *
Correct
17 (61)
1 (3)
Incorrect
5 (18)
10 (33)
Don’t know
6 (21)
19 (64)
Correct
20 (71)
0 (0)
Incorrect
1 (4)
2 (6)
Don’t know
7 (25)
28 (94)
Yes
26 (93)
26 (88)
No
0 (0)
1 (2)
Don’t know
2 (7)
3 (10)
Yes
24 (86)
18 (60)
No
3 (10)
6 (20)
Don’t know
1 (4)
6 (20)
Yes
21 (75)
12 (40)
No
2 (7)
3 (10)
Don’t know
5 (18)
15 (50)
<0.001
How was treatment allocation achieved? *
<0.001
Was participation voluntary?
0.57
Did you receive written information?
0.068
Did you read the written information? *
0.022
Did you understand the written information? *
Yes
18 (64)
11 (37)
No
2 (7)
0 (0)
Don’t know
8 (29)
19 (63)
0.017
* indicates a statistically signiﬁcant diﬀerence between HAMLET and PAIS.
127
Chapter 9
Table 2 Trial appreciation.
HAMLET (n (%))
PAIS (n (%))
n = 28
n = 30
P
Was the information provided clear to you? *
Yes
27 (96)
14 (47)
No
0 (0)
1 (3)
Don’t know
1 (4)
15 (50)
<0.001
Did you perceive the outcome of treatment allocation as a matter of life or death? *
Yes
23 (82)
10 (33)
No
4 (14)
16 (54)
Don’t know
1 (4)
4 (13)
0.001
Do you have negative feelings about the informed consent procedure? *
Yes
9 (32)
1 (3)
No
18 (64)
28 (94)
Don’t know
1 (4)
1 (3)
0.014
Do you feel that the question for participation and randomization was acceptable at that moment?
Yes
19 (68)
16 (54)
No
4 (14)
4 (13)
Don’t know
5 (18)
10 (33)
0.40
Were you capable of deciding about participation? *
Yes
17 (61)
13 (43)
No
8 (29)
4 (14)
Don’t know
3 (10)
13 (43)
0.018
Did you feel participation was voluntary?
Yes
18 (65)
23 (77)
No
8 (28)
3 (10)
Don’t know
2 (7)
4 (13)
0.18
Do you regret having given permission?
Yes
3 (11)
0 (0)
No
23 (82)
25 (83)
Don’t know
2 (7)
5 (17)
0.12
Are you glad with the treatment your relative received?
Yes
14 (50)
19 (63)
No
5 (18)
7 (23)
Don’t know
9 (32)
4 (13)
128
0.17
Informed consent in HAMLET
Table 2 – Continued –
HAMLET (n (%))
PAIS (n (%))
n = 28
n = 30
P
Would you have preferred the doctor having taken the responsibility for treatment choices?
Yes
11 (39)
6 (20)
No
15 (54)
17 (57)
Don’t know
2 (7)
7 (23)
0.12
Would you have preferred the doctor having taken the responsibility for study participation?
Yes
3 (11)
0 (0)
No
21 (75)
20 (67)
Don’t know
4 (14)
10 (33)
0.063
Would you approve participation again under similar circumstances?
Yes
17 (61)
20 (67)
No
2 (7)
1 (3)
Don’t know
9 (32)
9 (30)
0.78
How do you feel about being interviewed again?
Positive
23 (82)
23 (77)
Negative
3 (11)
2 (7)
Neutral
2 (7)
5 (17)
0.49
* indicates a statistically signiﬁcant diﬀerence between HAMLET and PAIS.
16-27). In both trials, informed consent was most frequently given by the patients’
partner. In PAIS, informed consent was given more frequently by patients’ (grand)
children (n = 10) than in HAMLET (n = 3, P = 0.046). Mean stroke severity as assessed
with the score on the NIHSS was signiﬁcantly higher in patients included in HAMLET
(median, 23; range, 20-31) than in PAIS (median, 14; range, 3-26; P = 0.02) and in
HAMLET the survival rate at follow up (n = 18, 60%) was signiﬁcantly lower than in
PAIS (n = 26, 87%, P = 0.03). The results of the interviews on recall of the details of
the trial are summarized in Table 1 and the results of the interviews on appreciation
of the informed consent procedure are summarized in Table 2. Relatives of HAMLET
participants remembered signiﬁcantly more trial details. With respect to appreciation
of the informed consent procedure, we found fewer diﬀerences between the groups. Of
the representatives of HAMLET participants, 8 (28%) felt participation to be obligatory,
as compared with 3 (10%) of those in PAIS (P = 0.1). Upon inquiry, it appeared that
these representatives felt this obligation only with respect to their relatives, because of
the reputed reduction in mortality. Of the representatives of HAMLET participants, 9
(32%) had negative feelings about the informed consent procedure, as compared with
1 (3%) of those in PAIS (P = 0.01). Upon inquiry, these negative feelings concerned the
acute and emotional situation and not the question of study participation in all but one
129
Chapter 9
representatives. With regard to these negative feelings in representatives of HAMLET
patients, there were no diﬀerences between representatives of patients who received
surgical or best medical treatment (n = 4 (27%) and n = 5 (36%) respectively (P = 0.4))
or between representatives of patients who died or survived (n = 3 (27% of all deaths),
and n = 6 (33% of all survivors) respectively (P = 0.6).
Discussion
Despite the reputed large impact on survival of decompressive surgery, representatives
of patients included in HAMLET were equally satisﬁed with the informed consent
procedure as those in a study assessing the eﬀect of acetaminophen on stroke
outcome. In both trials, most representatives considered trial participation and
randomization acceptable, despite the serious illness of their relatives. The majority
would give informed consent again in similar circumstances. HAMLET representatives
remembered signiﬁcantly more trial details.
Several factors have been put forward to explain poor recall of study details of
patients and representatives participating in clinical trials.1,14 Being admitted with an
acute and life-threatening disease has been proposed as the major factor compromising
the ability to retain information.1 It has been suggested that knowledge of details of a
clinical trial may be better in relatives than in the patients themselves, not only when
the relative had given consent, but also when the patient had done so.1 The present
study implies that recall of information may be better in life-threatening situations
than in other neurological emergencies. The impact of the treatment under study
undoubtedly plays an important role.
Previous studies have shown low recall rates of complex information, such as
randomization15 and other methodological details.16 None of the representatives of
PAIS participants could recall the process of random treatment allocation, whereas
this was recalled by two thirds of the representatives in HAMLET. Since treatment
allocation in HAMLET is not blinded, the concepts of drawing lots and probability are
much more explicit and probably easier to retain.
HAMLET diﬀers from other acute neurological emergencies by the high case
fatality rate of the disease and the presumed large eﬀect on the death rate of the
treatment under study. Three quarters of the representatives indeed considered the
outcome of randomization a matter of life or death. Still, about two thirds considered
participation in a clinical trial and randomization acceptable and only about one third
expressed negative feelings with respect to the informed consent procedure or study
participation. It has been suggested that positive answers with regard to appreciation
of trials may be biased by the hospital connection of the interviewer.1 Still, the results
of the present interviews show that even in a study like HAMLET, in which the clinical
situation is perceived as life-threatening, asking permission for study inclusion is
generally considered acceptable.
130
Informed consent in HAMLET
The present study has limitations. First, bias may have been introduced by not
interviewing representatives of patients who had refused to participate in either of
the trials. However, of the 56 representatives asked for participation of a relative in
HAMLET up to 1 August 2006, only three refused, in all cases to avoid survival with
severe disability. Second, patients included in HAMLET were younger than those
included in PAIS. In PAIS informed consent was more frequently given by the patients’
children. Nevertheless, it is likely that most representatives having given informed
consent for HAMLET were younger than those having given consent for PAIS. Age
may have inﬂuenced understanding and recall of the information. Third, we do not
have data on understanding of the trial information at the time of inclusion, so it is
impossible to ascertain whether the information was initially understood and then
forgotten.
We conclude that perception of information about a clinical trial by representatives
of patients with a neurological emergency is probably better if the condition is more
severe and if the reputed eﬀect of the treatment under study is larger. Participation
of patients in a randomized controlled trial of surgical decompression for spaceoccupying infarction is generally considered acceptable by their legal representatives.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
Schats R, Brilstra EH, Rinkel GJ, Algra A, van Gijn J. Informed consent in trials for neurological
emergencies: the example of subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 2003;74:988991.
Williams BF, French JK, White HD. Informed consent during the clinical emergency of acute
myocardial infarction (HERO-2 consent substudy): a prospective observational study. Lancet
2003;361:918-922.
Griﬃn JM, Struve JK, Collins D, Liu A, Nelson DB, Bloomﬁeld HE. Long term clinical trials: How
much information do participants retain from the informed consent process? Contemp Clin Trials
2006.
Sugarman J, McCrory DC, Hubal RC. Getting meaningful informed consent from older adults: a
structured literature review of empirical research. J Am Geriatr Soc 1998;46:517-524.
Joﬀe S, Cook EF, Cleary PD, Clark JW, Weeks JC. Quality of informed consent in cancer clinical trials:
a cross-sectional survey. Lancet 2001;358:1772-1777.
Cassileth BR, Zupkis RV, Sutton-Smith K, March V. Informed consent - why are its goals imperfectly
realized? N Engl J Med 1980;302:896-900.
Harrison K, Vlahov D, Jones K, Charron K, Clements ML. Medical eligibility, comprehension of the
consent process, and retention of injection drug users recruited for an HIV vaccine trial. J Acquir
Immune Deﬁc Syndr Hum Retrovirol 1995;10:386-390.
Howard JM, DeMets D. How informed is informed consent? The BHAT experience. Control Clin
Trials 1981;2:287-303.
Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Kappelle LJ, van der Worp HB. Hemicraniectomy After
Middle cerebral artery infarction with Life-threatening Edema Trial (HAMLET). Protocol for a
131
Chapter 9
10.
11.
12.
13.
14.
15.
16.
132
randomised controlled trial of decompressive surgery in space-occupying hemispheric infarction.
Trials 2006;7:29.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
van Breda EJ, van der Worp HB, van Gemert HM. PAIS: paracetamol (acetaminophen) in stroke;
protocol for a randomized, double blind clinical trial [ISCRTN 74418480]. BMC Cardiovasc Disord
2005;5:24.
Brott T, Adams HP, Jr., Olinger CP. Measurements of acute cerebral infarction: a clinical examination
scale. Stroke 1989;20:864-870.
Oddens BJ, Algra A, van Gijn J. Informing patients about clinical trials. Clin Investig 1993;71:572573.
Ross S, Grant A, Counsell C, Gillespie W, Russell I, Prescott R. Barriers to participate in randomised
controlled trials: a systematic review. J Clin Epidemiol 1999:52:1143-1156.
Miller C, Searight H, Grable D, Schwartz R, Barbarash R. Comprehension and recall of the
informational content of the informed consent document: an evaluation of 168 patients in a controlled
clinical trial. J Clin Res Drug Dev 1994;8:237-248.
Cognitive outcome after decompressive surgery
Chapter 10
Long-term cognitive outcome after
decompressive surgery for space-occupying
hemispheric infarction
Jeannette Hofmeijer, H. Bart van der Worp, G. Johan Amelink, L. Jaap Kappelle, Gudrun M.S. Nys, and
Martine J.E. van Zandvoort
In preperation
133
Chapter 10
Summary
In patients with space-occupying hemispheric infarction, decompressive surgery
may reduce mortality without increasing the number of severely disabled survivors.
However, cognitive functioning in surviving patients has been largely neglected.
Ten consecutive patients with a space-occupying hemispheric infarction, who had
been included in the surgical arm of the randomized Hemicraniectomy After spaceoccupying Middle cerebral artery (MCA) infarction with Life-threatening Edema
Trial (HAMLET) underwent neuropsychological testing approximately one year after
decompressive surgery. The tests comprised the following cognitive domains: general
cognitive functioning, language, memory, visual perception, neglect, and executive
functioning.
Testing of general cognitive functioning was severely hampered in all ﬁve patients
with infarcts in the dominant hemisphere, and in three of the ﬁve patients with infarcts
in the non-dominant hemisphere cognitive functioning was severely impaired. Testing
of language or verbal memory was not feasible or disclosed severe impairments in
patients with infarction in the dominant hemisphere. Non-verbal memory was
impaired in all patients with infarction in the non-dominant hemisphere and in three
of the patients with infarction in the dominant hemisphere. Visual perception was
impaired in four patients and below average in two. All patients with infarction in
the non-dominant hemisphere suﬀered profound hemispatial neglect. Executive
functioning was below average or impaired in all patients. In general, patients with
infarction in the non-dominant hemisphere performed better than patients with
infarction in the dominant hemisphere.
In addition to the expected severe focal deﬁcits, such as aphasia and neglect, global
cognitive decline was found in the majority of patients. Cognitive functioning in patients
with infarction in the dominant hemisphere was worse than in patients with infarction
in the non-dominant hemisphere. Before implementing decompressive surgery as
standard treatment in patients with space-occupying hemispheric infarction, results
of cognitive assessments of patients included in ongoing trials should be awaited.
134
Cognitive outcome after decompressive surgery
L
arge cerebral infarcts may cause space-occupying edema that puts the patient at
risk of transtentorial or uncal herniation.1 Fatal space-occupying brain edema
occurs in 1-5% of patients with a supratentorial infarct.2,3 The case fatality rate
of space-occupying infarcts has been reported to be as high as 78%, despite maximal
medical therapy on an intensive care unit.1,4
The aim of decompressive surgery in patients with space-occupying hemispheric
infarction is to revert brain tissue shifts and to normalize intracranial pressure, thereby
presumably preserving cerebral blood ﬂow and preventing secondary damage.5 Case
reports and non-randomized patient series have suggested a substantial reduction in
mortality after decompressive surgery without an increase in the number of severely
disabled survivors.6-10 For assessment of functional outcome, most investigators
used the modiﬁed Rankin Scale (mRS) and deﬁned favorable outcome as a score of
4 or lower, representing moderately severe disability at best.11 Reports on cognitive
outcome after decompressive surgery are scarce. The only study so far reports on 14
patients subjected to decompressive surgery for space-occupying infarction, however,
this study selectively included patients with right-sided infarction and a relatively mild
handicap.12
Over the last few years, cognitive functioning has become an important outcome
measure in stroke research, as this has shown to be crucial to dependence in daily
life,13,14 long-term survival,15 and more subjective measures such as quality of life.16
To study long-term cognitive impairments in unselected patients after decompressive
surgery for space-occupying hemispheric infarction, we performed comprehensive
neuropsychological testing one year after surgery.
Methods
Patients
Between September 2003 and November 2006, all consecutive patients with a spaceoccupying hemispheric infarct who had been included in the surgical arm of the
Hemicraniectomy After space-occupying Middle cerebral artery (MCA) infarction
with Life-threatening Edema Trial (HAMLET) underwent neuropsychological testing
approximately one year after surgery. Patients were included in HAMLET if the
following inclusion criteria had been met:17 (1) age 18 up to and including 60 years,
(2) clinical deﬁcits consistent with infarction in the territory of the middle cerebral
artery with a score on the National Institutes of Health Stroke Scale (NIHSS) > 15,18
(3) decrease in consciousness to a score of 13 or lower on the Glasgow Coma Scale for
patients without aphasia, or an Eye and Motor score of 9 or lower for patients with
aphasia, (4) signs on CT or MRI of a unilateral infarct in at least 50% of the territory of
the MCA, with or without additional infarction of the territories of the anterior (ACA)
or posterior cerebral artery (PCA) on the same side, (5) inclusion within 96 hours
135
Chapter 10
after the onset of symptoms, and (6) written informed consent by the patient or a legal
representative. Exclusion criteria were: (1) pre-stroke mRS11 ≥ 2, (2) presence of two
ﬁxed dilated pupils, (3) contralateral ischemia or other serious brain lesions, (4) spaceoccupying hemorrhagic transformation of the infarct, deﬁned as ≥ PH2 according to
ECASS criteria,19 (5) life expectancy <3 years, (6) other serious illness that may aﬀect
outcome, (7) known coagulopathy or systemic bleeding disorder, (8) contra-indication
for anesthesia, and (9) pregnancy.
Data collection
Demographic data were collected at the time of admission. Stroke severity was
assessed on admission and before surgical treatment. Infarct side and location were
assessed on CT scans made at randomization and follow-up. Neuropsychological
examination was performed at the University Medical Center Utrecht approximately
one year after surgery. The duration of the test procedure was dependent on the
patient’s performance and varied between one and two hours. The neuropsychological
examination comprised (1) a standardized battery for general cognitive functioning
[Cambridge Cognitive Examination (CAMCOG)]20 and nine tasks tapping the
following cognitive domains: (2) language [Boston Naming Task, Token Task (21-item
short form), Category Fluency (animals)], (3) memory [Rey Auditory Verbal Learning
Task (verbal), Location Learning Task (non-verbal)], (4) visual perception [Judgment
of Line Orientation], (5) neglect [Star Cancellation], and (6) executive functioning
[Ruﬀ Figural Fluency Test and Letter Fluency (N,A)]. This test battery was developed
for screening all major cognitive domains, but without being too extensive, given the
expected severe cognitive deﬁcits. Memory and Executive functioning were assessed
in both a verbal and a non-verbal way in order to overcome possible feasibility
problems with either language (dominant hemisphere dysfunctioning) or non-verbal
visuospatial information (non-dominant hemisphere dysfunctioning).
Outcome measures and analysis
Data are presented in a qualitative way, using the population norms of the
neuropsychological tests, corrected for age and level of education.20 Either test-speciﬁc
cut-oﬀ values were used to describe the level of functioning (CAMCOG, Token Task),
or the raw scores were converted into percentiles of the general population, to adjudge
a level of functioning ranging from impaired (< -2 standard deviations (SD)), below
average (≥ -2SD and < -1SD), average (≥ -1SD and ≤ 1SD), high average (> 1SD and
≤ 2SD), or superior (> 2SD). To compare performance of patients with infarcts in the
dominant and the non-dominant hemisphere, the number of test scores above the level
of ‘impaired’ was calculated. The Mann Whitney U test was used to test diﬀerences
between the groups. P < 0.05 was considered statistically signiﬁcant.
136
Cognitive outcome after decompressive surgery
Table 1 Baseline data.
Patient
number
Sex
Age
Infarct side
(years)
Infarct location
NIHSS at
randomization
mRS 1 year
1
F
49
Left
MCA
27
4
2
M
50
Left
MCA+ACA
22
4
3
F
59
Left
MCA+PCA
26
4
4
M
38
Left
MCA+PCA
23
5
5
M
57
Right
MCA+ACA
24
3
6
M
57
Right
MCA
22
4
7
M
54
Right
MCA
22
4
8
F
44
Right
MCA+ACA
29
4
9
M
59
Right
MCA
26
4
10
F
44
Right
MCA
23
4
F indicates female; M, male; MCA, middle cerebral artery; ACA, anterior cerebral artery; PCA, posterior
cerebral artery; NIHSS, National Institutes of Health Stroke Scale; mRS, modiﬁed Rankin Scale.
Results
As of December 1st, 2006 twelve patients had been eligible for neuropsychological
examination one year after hemicraniectomy. Two patients refused to participate,
because of fear to be confronted with cognitive decline. The mean interval ± SD
between stroke onset and testing was 14.4 ± 1.6 months. Baseline data are summarized
in Table 1. Patient number 9 had an infarct in the right hemisphere with aphasia. This
patient was left-handed.
One patient had an mRS score of 5, indicating severe disability, eight patients
had an mRS score of 4, indicating moderately severe disability and one patient was
functionally independent but needed help with some daily activities (mRS score of
3). Hetero-anamnestic information from patients’ relatives disclosed that, in general,
patients were cognitively impaired. Relatives often mentioned inability to plan and
structure activities interfering with daily life functioning, and problems with impulse
regulation. Moreover, the capacity to adapt to new situations was often impaired,
leading to frequent frustrations.
Test performances are summarized in Table 2. Testing of general cognitive functioning
was severely hampered in the four patients with left hemisphere infarction and in one
patient with right hemisphere infarction as a result of language impairments. These ﬁve
patients met the criteria for global aphasia (severe impairment in both comprehension
and production of language) prohibiting verbal communication. General cognitive
functioning was impaired in three of the ﬁve patients with infarcts in the dominant
hemisphere. Verbal memory was not testable in patients with infarcts in the dominant
137
Chapter 10
Rey Auditory Verbal Learning
Rey Auditory Verbal Learning del
Location Learning Task displ
Location Learning Task learn
Location Learning Task del
Judgment of Line Orientation
Star Cancellation
Figural Fluency
Number of tasks above ‘impaired’
na
avg
avg
na
imp
avg
imp
3
na
na
na
imp
imp
imp
avg
avg
bavg
3
na
na
na
avg
bavg
avg
bavg
avg
na
5
na
na
na
na
na
na
na
na
na
0
avg
avg
nvg
imp
imp
avg
avg
negl
imp
8
imp
na
bavg
navg
imp
imp
imp
imp
negl
imp
2
avg
avg
bavg
navg
imp
imp
havg
imp
negl
imp
7
imp
negl
imp
6
avg
avg
imp
4
negl
imp
7
1
na
na
na
2
na
imp
imp
3
na
na
imp
4
na
na
na
5
avg
sup
sup
6
imp
imp
7
avg
avg
8
imp
avg
avg
avg
bavg
nvg
imp
imp
avg
9
na
imp
imp
na
na
na
imp
bavg
bavg
10
imp
avg
avg
imp
avg
navg
bavg
imp
havg
bavg
Patient number
Token Task
na
Boston Naming Task
na
CAMCOG
Category Fluency
Table 2 Neuropsychological test performance.
na indicates not assessable; del, delayed recall; displ, displacement score; learn, learning index; imp, impaired;
bavg, below average; avg, average; havg, high average; sup, superior; negl, neglect.
Figure Cognitive functioning deﬁned as the
number of test-scores above the level of impaired
for patients with infarcts in the dominant and the
non-dominant hemisphere. Box plots indicate
Medians, Quartiles, and Ranges.
138
Cognitive outcome after decompressive surgery
hemisphere, and below average in four of the ﬁve patients with infarcts in the nondominant hemisphere. Non-verbal memory was impaired in all patients with infarcts
in the non-dominant hemisphere and in three of the patients with infarcts in the
dominant hemisphere. Visual perception was impaired in four patients and below
average in two. All patients with infarcts in the non-dominant hemisphere suﬀered
severe hemispatial neglect. Executive functioning was below average or impaired in all
patients. In general, patients with infarcts in the non-dominant hemisphere tended to
perform better than patients with infarcts in the dominant hemisphere (Figure), but
this diﬀerence did not reach statistical signiﬁcance (P = 0.08).
Discussion
This is the ﬁrst report of neuropsychological examination in unselected patients after
decompressive surgery for space-occupying MCA infarction. Cognitive functioning
was poor in the majority. In addition to the expected severe focal deﬁcits, such as
aphasia and neglect, executive functioning and memory were aﬀected in all patients
resulting in below average or impaired performance.
Several studies have shown a decline in cognitive functioning after all types of
stroke.21,22 Higher cognitive functions are not solely located in a speciﬁc area of the
brain, but are the result of complex and dynamic neural networks which are scattered
throughout the brain.23 For this reason, neuropsychological functioning is highly
dependent on the overall integrity of the brain, and may be impaired even after lacunar
or brainstem infarction,24,25 let alone after large cortical infarcts.
Irrespective of the side of infarction, most patients had a score on the mRS of 4.
Nevertheless, patients with infarction in the dominant hemisphere performed worse
than patients with infarcts in the non-dominant hemisphere, having impairments
on both verbal and non-verbal tasks. Similar ﬁndings have been reported earlier in
a more general stroke population.22 Language probably has a far-reaching impact on
cognitive functioning and should be considered as the means by which most cognitive
functions are articulated.21
Executive functioning, as tested with the Ruﬀ Figural Fluency Test, was below average
or impaired in all patients. This cannot be completely attributed to the presence of
aphasia, amnesia, or a reduction in motor speed. The impaired performance reﬂects
an inﬂexibility in information processing and an impairment in generation of ideas.
In daily life this means that patients may become locked into one strategy and do
not possess the ability to change to another. Relatives of patients with lesions in the
non-dominant as well as relatives of patients with lesions in the dominant hemisphere
have conﬁrmed this observation. According to their relatives, the patients’ inability to
initiate, plan and structure activities interferes with daily life functioning. Moreover,
the capacity to adapt to new situations appears to be impaired to such an amount that
139
Chapter 10
this leads to frequent frustrations.
Next to executive functioning, both verbal and non-verbal memory were impaired
in most patients. Memory and the ability to learn probably deteriorate even more
by disturbances of attention and concentration. In a previous study of cognitive
functioning in patients with large MCA infarction subjected to decompressive surgery,
profound disturbances of attention were found. In this study only patients with rightsided infarction and with a relatively preserved level of functioning were included.12
We performed neuropsychological examination more than one year after the onset
of symptoms. Although further cognitive recovery after the ﬁrst year following stroke
has been described,26 a lack of improvement or even cognitive decline in the second
year after stroke onset has also been reported.27 Patients with aphasia not only tend
to perform worse, but also show less improvement.27,28 Thus, in our patients further
improvement is probably very limited.
The current study has limitations. First, the small sample size precludes deﬁnitive
conclusions on the eﬀects of cognitive and functional outcome. Despite the overall low
performance, two patients (patient 5 and 7) performed average or above average on
most tasks. Neuropsychological evaluation of all patients participating in randomized
trials of decompressive surgery in space-occupying infarction is clearly warranted.
Secondly, because of the small sample size, data were mainly presented in a descriptive
way. Nevertheless, the serious impact of life-threatening space-occupying infarction
on cognition is evident. Thirdly, in patients with infarcts in the dominant hemisphere
language was too poor to be quantiﬁed by means of any of the tasks included in the
test battery. In general, neuropsychological examination is feasible in the majority of
patients with aphasia caused by stroke.14 Although ﬂoor eﬀects prevented quantiﬁcation
in our series, these may still be seen as an indication of the magnitude of the damage
in these patients. Finally, the present study did not include a control group of patients
with space-occupying MCA infarcts that did not undergo surgery. For this reason,
we do not know the impact of decompressive surgery on cognition in this patient
population.
In conclusion, although decompressive surgery increased the probability of a
favorable outcome in patients with space-occupying MCA infarction in a pooled
analysis of ongoing trials (chapter 11), the present results suggest that survivors are
likely to be left with signiﬁcant cognitive impairments. Despite similar functional
dependence as assessed with the mRS, cognitive functioning in patients with
infarction in the dominant hemisphere was worse than in patients with infarction
in the non-dominant hemisphere. Before implementing decompressive surgery as a
standard treatment modality in patients with MCA infarction, who deteriorate as a
result of space-occupying edema, results of cognitive assessments of patients included
in ongoing trials should be awaited. Moreover, the possible cognitive consequences,
especially for patients with a stroke in the dominant hemisphere, should be considered
before a decision to perform surgery is taken.
140
Cognitive outcome after decompressive surgery
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
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 1996;53:309-315.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review.
Stroke 1984;15:492-496.
Heinsius T, Bogousslavsky J, Van Melle G. Large infarcts in the middle cerebral artery territory.
Etiology and outcome patterns. Neurology 1998;50:341-350.
Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction and early brain swelling:
progression and eﬀect of age on outcome. Mayo Clin Proc 1998;73:829-836.
Schwab S, Rieke K, Aschoﬀ A, Albert F, von Kummer R, Hacke W. Hemicraniotomy in space-occupying
hemispheric infarction: useful early intervention or desparate activism? Cerebrovasc Dis 1996;6:325329.
Carter BS, Ogilvy CS, Candia GJ, Rosas HD, Buonanno F. One-year outcome after decompressive
surgery for massive nondominant hemispheric infarction. Neurosurgery 1997;40:1168-1175.
Delashaw JB, Broaddus WC, Kassell NF. Treatment of right hemispheric cerebral infarction by
hemicraniectomy. Stroke 1990;21:874-881.
van Leusen HJ, Tans JT, Wurzer JA. [Hemicraniectomy for treatment of malignant medial cerebral
artery infarction in 3 patients]. Ned Tijdschr Geneeskd 2001;145:639-643.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the
assessment of handicap in stroke patients. Stroke 1988;19:604-607.
Leonhardt G, Wilhelm H, Doerﬂer A. Clinical outcome and neuropsychological deﬁcits after right
decompressive hemicraniectomy in MCA infarction. J Neurol 2002;249:1433-1440.
Tatemichi TK, Desmond DW, Stern Y, Paik M, Sano M, Bagiella E. Cognitive impairment after stroke:
frequency, patterns, and relationship to functional abilities. J Neurol Neurosurg Psychiatry 1994;57:202207.
Nys GM, Van Zandvoort MJ, De Kort PL. The prognostic value of domain-speciﬁc cognitive abilities
in acute ﬁrst-ever stroke. Neurology 2005;64:821-827.
Tatemichi TK, Paik M, Bagiella E, Desmond DW, Pirro M, Hanzawa LK. Dementia after stroke is a
predictor of long-term survival. Stroke 1994;25:1915-1919.
Nys GM, Van Zandvoort MJ, van der Worp HB, de Haan EH, de Kort PL, Jansen BP, Kappelle LJ. Early
cognitive impairment predicts long-term depressive symptoms and quality of life after stroke. J Neurol
Sci 2006;247:149-156.
Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Macleod MR, Kappelle LJ, van der Worp HB.
Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema Trial
(HAMLET). Protocol for a randomised controlled trial of decompressive surgery in space-occupying
hemispheric infarction. Trials 2006 11;7:29.
Goldstein LB, Bertels C, Davis JN. Interrater reliability of the NIH stroke scale. Arch Neurol
1989;46:660-662.
141
Chapter 10
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
142
Larrue V, von Kummer RR, Muller A, Bluhmki E. Risk factors for severe hemorrhagic transformation
in ischemic stroke patients treated with recombinant tissue plasminogen activator: a secondary
analysis of the European-Australasian Acute Stroke Study (ECASS II). Stroke 2001;32:438-441.
Lezak MD, Howieson DB, Loring DW. Neuropsychological assessment, 4 ed. New York: Oxford
University Press, 2004.
Luria A. Higher cortical functions in man. New York: Basic Books, 1966.
Hochstenbach J, Mulder T, van LJ, Donders R, Schoonderwaldt H. Cognitive decline following stroke:
a comprehensive study of cognitive decline following stroke. J Clin Exp Neuropsychol 1998;20:503517.
Hom J, Reitan RM. Generalized cognitive function after stroke. J Clin Exp Neuropsychol 1990;12:644654.
van Zandvoort M, de Haan EH, van Gijn J, Kappelle LJ. Cognitive functioning in patients with a small
infarct in the brainstem. J Int Neuropsychol Soc 2003;9:490-494.
Van Zandvoort MJ, De Haan EH, Kappelle LJ. Chronic cognitive disturbances after a single
supratentorial lacunar infarct. Neuropsychiatry Neuropsychol Behav Neurol 2001;14:98-102.
del Ser T, Barba R, Morin MM. Evolution of cognitive impairment after stroke and risk factors for
delayed progression. Stroke 2005;36:2670-2675.
Hochstenbach JB, den Otter R, Mulder TW. Cognitive recovery after stroke: a 2-year follow-up. Arch
Phys Med Rehabil 2003;84:1499-1504.
Nys GM, Van Zandvoort MJ, De Kort PL, Jansen BP, van der Worp HB, Kappelle LJ, de Haan EH.
Domain-speciﬁc cognitive recovery after ﬁrst-ever stroke: a follow-up study of 111 cases. J Int
Neuropsychol Soc 2005 Nov;11:795-806.
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
Chapter 11
Early decompressive surgery in spaceoccupying hemispheric infarction:
a pooled analysis of three randomized
controlled trials
Katayoun Vahedi,* Jeannette Hofmeijer,* Eric Juettler,* Eric Vicaut, Bernard George, Ale Algra, G. Johan
Amelink, Peter Schmiedeck, Stefan Schwab, Peter M. Rothwell, Marie-Germaine Bousser, H. Bart van der
Worp, and Werner Hacke for the DECIMAL, DESTINY, and HAMLET investigators
*These authors contributed equally
Based on Lancet Neurology 2007; 6:215-222
143
Chapter 11
Summary
Space-occupying middle cerebral artery (MCA) infarction has a mortality rate of
about 80%. Non-randomized studies have suggested that decompressive surgery
reduces mortality without increasing the number of severely disabled survivors. To
obtain suﬃcient data to reliably estimate the eﬀects of decompressive surgery as
soon as possible, data from three European randomized controlled trials (DECIMAL,
DESTINY, and HAMLET) were pooled. The trials were ongoing when the pooled
analysis was planned.
Individual data of patients between 18 and 60 years of age with space-occupying
MCA infarction included in one of the three trials and treated within 48 hours after
stroke onset were pooled. The protocol was designed prospectively when the trials
were still recruiting and outcomes were deﬁned blind to the results of the individual
trials. The primary outcome measure was the score on the modiﬁed Rankin Scale
(mRS) at one year dichotomized between favorable (0-4) and unfavorable (5 and
death). Secondary outcome measures included a dichotomization of the mRS between
0-3 and 4 to death and case fatality rate at one year. Data analysis was performed by an
independent data monitoring committee. Absolute risk reductions (ARR), odds ratios,
and 95% conﬁdence intervals (95% CI) were calculated for the speciﬁed outcomes in
each trial and then pooled by the Mantel-Haenszel method. P < 0.05 was considered
statistically signiﬁcant.
Ninety-three patients were included. After decompressive surgery, more patients
had an mRS≤4 (75% vs 24%; pooled ARR [95%CI]: 51%[34-69]), an mRS≤3 (43% vs 21%;
ARR: 23%[5-41]), and survived (78% vs 29%; ARR: 50%[33-67]) indicating numbers
needed to treat of 2 for survival with an mRS≤4, 4 for survival with an mRS≤3, and 2
for survival irrespective of outcome. The eﬀect of surgery was highly consistent across
the three trials.
In patients with space-occupying MCA infarction, decompressive surgery reduces
case fatality by 50% with an mRS≤3 at one year in 55% of the survivors. The decision
to perform decompressive surgery should be made on an individual basis in every
patient.
144
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
L
ife-threatening space-occupying brain edema occurs in 1 to 10% of patients
with a supratentorial infarct, and usually manifests itself between the second
and the ﬁfth day after stroke onset.1-3 However, up to one third of the patients
may have neurological deterioration within 24 hrs of symptom onset.4 The prognosis
of these space-occupying or ‘malignant’ MCA infarcts is poor, with case fatality rates
in intensive care-based series of nearly 80%.5,6 No medical therapy has been proven effective.7 Diﬀerent predictors of fatal brain edema formation have been identiﬁed, such
as major early CT hypodensity involving >50% of the MCA territory and involvement
of additional vascular territories besides that of the MCA.8 However, to date no single
prognostic factor with suﬃcient prognostic value has been identiﬁed.
Prospective, non-randomized studies have suggested that decompressive surgery,
consisting of a hemicraniectomy and duraplasty, reduces mortality in patients with
space-occupying MCA infarction without increasing the number of severely disabled
survivors.9-12 However, evidence from randomized trials is lacking. While most
clinicians agree that the procedure is probably life-saving, no convincing data are
available regarding functional outcome of survivors.
The eﬀect of decompressive surgery on functional outcome in patients with
space-occupying MCA infarction has been studied in three European randomized
controlled trials: the French DECIMAL [DEcompressive Craniectomy In MALignant
middle cerebral artery infarcts], the German DESTINY [DEcompressive Surgery for
the Treatment of malignant INfarction of the middle cerebral arterY], and the Dutch
HAMLET [Hemicraniectomy After Middle cerebral artery infarction with Lifethreatening Edema Trial].13 Two of these trials interrupted recruitment early in 2006:
DECIMAL because of slow recruitment and a statistically signiﬁcant diﬀerence in
mortality between the treatment groups favoring surgery, and DESTINY because a
pre-deﬁned sequential analysis showed a statistically signiﬁcant beneﬁt of surgery on
mortality. HAMLET is ongoing.
The three trials have a similar design and share the same primary outcome measure,
i.e. favorable versus unfavorable functional outcome as determined by the score on
the modiﬁed Rankin Scale (mRS).14 A collaborative protocol for a pooled analysis of
individual patient data from the three trials was planned before the interruption of the
ﬁrst two trials. The principal aim of this pooled analysis was to obtain suﬃcient data to
reliably estimate the eﬀects of decompressive surgery as soon as possible so as to avoid
unnecessary (and unethical) continuation of randomization in the individual trials.
Methods
Design
We combined individual patient data from DECIMAL (NCT00190203), DESTINY
(ISRCTN01258591), and HAMLET (ISRCTN94237756), which are multi-center,
145
Chapter 11
randomized, controlled clinical trials assessing the eﬀect of decompressive surgery in
patients with space-occupying MCA infarction. When the pooled analysis was planned,
the trials were still ongoing and there was no knowledge on outcome data except for
mortality rates in DECIMAL and DESTINY. At the time of the analysis, DECIMAL and
DESTINY had been interrupted, while HAMLET was still ongoing. Randomization,
treatment, and outcome assessment were performed according to the individual study
protocols that were approved by the relevant institutional review boards. Informed
consent was obtained from the patients or the patients’ legal representatives.
Trial characteristics
DECIMAL was designed to include a maximum of 60 patients, 30 in each group.
Recruitment stopped after inclusion of 38 patients in March 2006 because of slow
enrollment, a signiﬁcant diﬀerence in mortality favoring decompressive surgery, and
the opportunity of a pooled analysis with DESTINY and HAMLET. DESTINY aimed
to include a maximum of 68 patients. Recruitment was interrupted in February 2006
after a planned interim analysis including 32 patients showed a statistically signiﬁcant
beneﬁt of surgery on 30-day mortality, and was stopped deﬁnitively after a revised
sample size projection indicated that 188 patients would be required to show a
statistically signiﬁcant diﬀerence with regard to the primary endpoint (mRS 0 to 3
versus 4 to death at six months). HAMLET aims to include 112 patients.13
Eligibility criteria for the pooled analysis
The inclusion and exclusion criteria of the three trials were largely similar. The main
diﬀerences included: a longer interval from stroke onset to treatment start allowed
in HAMLET (96 h) than in DECIMAL (30 h) and DESTINY (36 h). For the pooled
analysis, a maximum time window from stroke onset to randomization of 45h (i.e. 48h
to treatment) was adopted. Neuro-imaging criteria diﬀered between the three trials
and could therefore not be included in the pooled analysis. These criteria were a DWI
infarct volume more than 145 cm3 in DECIMAL, brain CT ischemic changes involving
more than 2/3 of the MCA territory and including the basal ganglia in DESTINY and
brain CT ischemic changes involving at least 2/3 of the MCA territory with spaceoccupying edema in HAMLET. Patients included before 1 November 2005 in any of
the three trials and fulﬁlling the prospectively deﬁned eligibility criteria listed in Table
1 were included in this pooled analysis.
Randomization and treatment
Patients were randomized to either decompressive surgery or conservative treatment
in all trials. In DECIMAL, patients were centrally randomized in blocks of four using a
pre-established randomization list. In DESTINY, randomization was done according
to a central computer generated randomization list for each participating center.
146
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
Table1 Eligibility criteria for the pooled analysis.
Inclusion criteria
Age 18 - 60 years.
Clinical deﬁcits suggestive of infarction in the territory of the middle cerebral artery (MCA) with a score on
the National Institutes of Health Stroke Scale (NIHSS) > 15.
Decrease in the level of consciousness to a score of 1 or greater on item 1a of the NIHSS.
Signs on CT of an infarct of at least 50% of the MCA territory, with or without additional infarction in the
territory of the anterior (ACA) or posterior cerebral artery (PCA) on the same side, or DWI infarct volume >
145 cm3.
Inclusion within 45 hours after onset of symptoms.
Written informed consent by the patient or a legal representative.
Exclusion criteria
Pre-stroke score on the mRS ≥ 2.
Two ﬁxed dilated pupils.
Contralateral ischemia or other brain lesion that may aﬀect outcome.
Space-occupying hemorrhagic transformation of the infarct.
Life expectancy < 3 years.
Other serious illness that may aﬀect outcome.
Known coagulopathy or systemic bleeding disorder.
Contra-indication for anesthesia.
Pregnancy.
In HAMLET, randomization was performed centrally, with a computerized algorithm
in which an element of chance was added to the treatment decision of minimization
(chapter 7).
Decompressive surgery consisted of creating a large bone ﬂap and duraplasty. In
summary, a large (reversed) question mark-shaped skin incision based at the ear was
made. A bone ﬂap with a diameter of at least 12 cm (always including the frontal,
temporal, and parietal bones) was removed. Additional temporal bone was removed so
that the ﬂoor of the middle cerebral fossa could be reached. The dura was opened and
a dural patch, consisting of pericranium or a commercially available dura substitute,
was inserted and secured to enlarge the intradural space. To prevent epidural bleeding,
dural tacking sutures were used when considered necessary. The temporal muscle and
the skin ﬂap were then re-approximated and sutured. Infarcted brain tissue was not
resected. In surviving patients, cranioplasty was performed after at least six weeks,
using the stored bone ﬂap or acrylate. After surgery, patients were transferred to
an intensive care unit, but anti-edema therapy was usually not necessary.15 In the
147
Chapter 11
conservative group, patients received best medical treatment, based on published
guidelines for the management of acute ischemic stroke and space-occupying brain
edema .16-18
Data collection
A broad range of baseline characteristics and outcome measures was collected in
the individual trials. For the pooled analysis the following pre-speciﬁed baseline
characteristics were used: age, sex, time between stroke onset and randomization,
medical history, physical examination (blood pressure, body temperature), presence
of aphasia, and score on the National Institutes of Health Stroke Scale (NIHSS) at
randomization.19
Outcome measures
The trials used largely similar outcome measures. In the pooled analysis, the primary
outcome measure was the score on the mRS at one year dichotomized between
‘favorable’ (mRS 0 to 4) and ’unfavorable’ (mRS 5 and death). Secondary analyses
included a dichotomization of the mRS in which ‘favorable’ was deﬁned as a score of
0 to 3 and ‘unfavorable’ as a score of 4 to death, and case fatality at one year. The mRS
measures functional outcome after stroke.14 The scores range from 0 to 6, 0 indicating
no symptoms at all, 1 indicating no signiﬁcant disability despite symptoms, being able
to carry out all usual duties and activities, 2 indicating slight disability, being unable to
carry out all previous activities, but able to look after own aﬀairs without assistance,
3 indicating moderate disability, requiring some help, but being able to walk without
assistance, 4 indicating moderately severe disability, being unable to walk without
assistance and unable to attend to own bodily needs without assistance, 5 indicating
severe disability, being bedridden, incontinent and requiring constant nursing care
and attention, and 6 indicating death. In DECIMAL outcomes were assessed by a
neurologist blinded to treatment allocation, in DESTINY outcome was assessed
unblinded, and in HAMLET the score on the mRS was determined independently by
three blinded investigators, based on a narrative written by an unblinded independent
study nurse, if necessary followed by a consensus meeting.
Statistical analyses
Data analysis was performed according to a pre-speciﬁed protocol, by an independent
Data Monitoring Committee. The distributions of the mRS, with death scored as 6,
were compared between the treatment groups with the Mann-Whitney U test. To
assess the eﬀect of surgical treatment, absolute risk reductions (ARR), odds ratios (OR),
and 95% conﬁdence intervals (95% CI) were calculated for the speciﬁed outcomes in
each trial and then pooled by the Mantel-Haenszel method. Heterogeneity of ARR and
OR between trials was determined by the Breslow-Day test. The inﬂuence of baseline
148
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
diﬀerences between the treatment groups was assessed by the comparison of crude
and adjusted ORs. Data were also analyzed in a cumulative logit model.20 These ORs
represent the odds of obtaining higher rather than lower mRS scores after surgical
treatment compared with conservative treatment. Subgroup analyses were performed
according to age (dichotomized at 50 years), timing of randomization (dichotomized
at 24 hours), and presence of aphasia. Analyses were performed on an intention-totreat basis. The SPSS software package was used for all analyses and P < 0.05 was
considered statistically signiﬁcant.
Results
All patients randomized in DECIMAL (38 patients) and DESTINY (32 patients) and
28 patients randomized in HAMLET were eligible for the pooled analysis. From
HAMLET, 29 of a total of 57 patients were excluded because they were randomized
after 45 hours from stroke onset or were included after November 1st 2005. For all
other patients, there were no missing data on primary or secondary outcome measures.
Thus, 93 patients were included, of whom 51 were randomized to decompressive
surgery and 42 to conservative treatment. There was one cross-over in DESTINY
from conservative treatment to decompressive surgery. There were no cross-overs in
the other trials. Of two patients from DESTINY the primary outcome measure was
assessed at ten months.
Baseline characteristics
Baseline characteristics are shown in Table 2. Treatment groups within the individual
trials were broadly similar. There were two minor diﬀerences: in DESTINY, the
conservatively treated group had a higher NIHSS score than the surgically treated
group, and in DECIMAL, mean systolic blood pressure was higher in the conservatively
treated group than in the surgically treated group. There were baseline diﬀerences
between the three trials: Time to randomization in DESTINY and HAMLET was
signiﬁcantly longer than in DECIMAL. Time to randomization in HAMLET was
also longer than in DESTINY. NIHSS scores in DESTINY and DECIMAL were
lower than in HAMLET. Body temperature was lower in DESTINY and DECIMAL
than in HAMLET. History of TIA or stroke was more frequent in HAMLET than in
DESTINY, and a history of ischemic heart disease was more common in DESTINY
than in DECIMAL.
Outcome
Figure 1 shows the distributions of the scores on the mRS after 12 months according
to randomized treatment. The diﬀerence in distributions of the scores on the mRS
between the two treatment groups was highly signiﬁcant (p < 0.001; Figure 1).
149
150
DECIMAL
9 (45%)
2/19 (11%)
0/19 (0%)
0/19 (0%)
6/19 (32%)
0/19 (0%)
9/18 (50%)
36.9 (0.5)
135.5 (19.3)
75.8 (15.6)
12 (60%)
21.5 (18-25)
History of TIA/stroke
Ischemic heart disease
Atrial ﬁbrillation
Hypertension
Diabetes
Current smoker
Temperature (mean [SD])
Systolic BP (mean [SD])
Diastolic BP (mean [SD])
Aphasia
NIHSS (median [IQR])
Hours to randomization (median [IQR]) 16 (11-21)
43.4 (9.7)
Male sex
16 (11-20)
21.5 (18-26)
11 (61%)
83.4 (15.1)
154.0 (25.1)
37.0 (0.6)
4/16 (25%)
4 (22%)
5 (28%)
1 (6%)
1 (6%)
0 (0%)
9 (50%)
43.4 (7.3)
Surgery (n=20) Conservative (n=18)
Age (mean [SD])
Table 2 Baseline characteristics.
16 (11-21)
21.5 (18-25)
23 (61%)
79.4 (15.6)
144.2 (23.8)
37.0 (0.6)
13/34 (38%)
5/37 (14%)
11/37 (30%)
1/37 (3%)
1/37 (3%)
2/37 (5%)
18 (47%)
43.4 (8.5)
Total (n=38)
24 (18-29)
21 (19.5-23)
10 (59%)
74.7 (15.6)
139.4 (16.6)
37.0 (0.5)
7 (41%)
2 (12%)
9 (53%)
3 (18%)
3 (18%)
0 (0%)
8 (47%)
43.2 (9.7)
23 (17-33)
24 (22-26)
11 (73%)
72.3 (11.2)
133.3 (14.2)
37.2 (0.6)
5 (33%)
3 (20%)
7 (47%)
3 (20%)
4 (27%)
0 (0%)
7 (47%)
46.1 (8.4)
24 (18-29)
22 (20.3-24)
21 (66%)
73.6 (13.5)
136.6 (15.6)
37.1 (0.5)
12 (38%)
5 (16%)
16 (50%)
6 (19%)
7 (22%)
0 (0%)
15 (47%)
44.6 (9.1)
Surgery (n=17) Conservative (n=15) Total (n=32)
DESTINY
Chapter 11
HAMLET
9 (64%)
3 (21%)
2 (14%)
4 (29%)
4 (29%)
2 (14%)
5/12 (42%)
37.5 (0.6)
147.5 (27.4)
74.6 (16.5)
6 (43%)
23 (21.8-27)
Male sex
History of TIA/stroke
Ischemic heart disease
Atrial ﬁbrillation
Hypertension
Diabetes
Current smoker
Temperature (mean [SD])
Systolic BP (mean [SD])
Diastolic BP (mean [SD])
Aphasia
NIHSS (median [IQR])
30 (20-41)
27 (22.5-32)
3 (33%)
76.3 (21.9)
148.1 (23.0)
37.5 (0.5)
4 (44%)
0 (0%)
1 (11%)
0 (0%)
0 (0%)
2/8 (25%)
3 (33%)
43.0 (12.6)
30 (24-40)
24 (22-28)
9 (39%)
75.3 (18.3)
147.7 (25.2)
37.5 (0.5)
9/21 (43%)
2 (9%)
5 (22%)
4 (17%)
2 (9%)
5 (22%)
12 (52%)
48.2 (9.9)
Total (n=23)
a
0.90b
0.77
b
1.00
0.13a
0.014
0.55
a
0.17
0.18
1.00
0.49
0.49
0.49
b
0.65b
0.0033
0.47
0.63a
0.28
a
0.34
a
0.73
0.65
1.00
1.00
0.68
1.00
1.00
0.39a
0.98a
1.00
DESTINY
0.91b
0.13b
1.00
0.83a
0.96a
0.96a
1.00
0.50
0.61
0.13
0.50
1.00
0.21
0.086a
HAMLET
Comparison of treatment groups
DECIMAL
Data are number (%) umless otherwise indicated. Group comparisons are Fisher’s exact test unless otherwise indicated. SD indicates standard deviation; IQR,
interquartile range; a, t-test or ANOVA as appropriate; b, Mann-Withney U or Kruskal-Wallis test as appropriate.
Hours to randomization (median [IQR]) 31 (23-39)
51.6 (6.1)
Surgery (n=14) Conservative (n=9)
Age (mean [SD])
Table 2 – Continued –
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
151
Chapter 11
Figure 1 Distributions of the scores on the mRS and death after 12 months for patients treated with or without
surgical decompression.
Signiﬁcantly fewer patients had an unfavorable outcome deﬁned as an mRS score of
5 or death at 12 months after surgery than after conservative treatment (13/51 (25%)
patients in the surgical group vs 32/42 (76%) in the conservative group; pooled ARR
[95%CI]: 51.2% [33.9-68.5]; Figure 2a). Signiﬁcantly fewer patients had a mRS > 3 at
12 months after surgical treatment than after conservative treatment (29/51 (57%) in
the surgical group vs 33/42 (79%) in the conservative group; pooled ARR [95%CI]:
22.7% [4.6-40.9]; Figure 2b). The survival rate at 12 months was higher after surgical
treatment than after conservative treatment (40/51 (78%) in the surgical group vs
12/42 (29%) in the conservative group; pooled ARR [95%CI]: 50.3% [33.3-67.4]; Figure
2c). The results of our analyses remained essentially the same after adjustment for
baseline incomparabilities.
With regard to all three outcome measures, there was no statistically signiﬁcant
heterogeneity between the three trials. If baseline diﬀerences between the treatment
groups were taken into account, the reduction in ORs remained essentially the same
for all three analyses. The resulting numbers needed to treat [95%CI] for the three
outcomes are 2 [1.5-3] for the prevention of mRS 5 or death, 4 [2-22] for the prevention
of mRS 4 to death and 2 [1.5-3] for survival irrespective of outcome.
Surgery was beneﬁcial (p < 0.01) in all pre-deﬁned subgroups [age (above and
below 50 years), presence of aphasia, and time to randomization (above and below 24
hours)], as deﬁned by mRS ≤ 4 at 12 months, with no statistically signiﬁcant subgrouptreatment eﬀect interactions (Figure 3).
152
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
Figure 2 Absolute Risk Reductions and Odds Ratios for unfavorable outcome at 12 months deﬁned as mRS >
4 (a), mRS > 3 (b), and death (c).
Outcome / Patients
Conservative
Surgery
ARR (%)
95% CI
OR
95% CI
0.10
0.15
0.05
0.02-0.43
0.03-0.73
0.00-0.54
0.10
0.04-0.27
0.29
0.41
0.31
0.07-1.18
0.09-1.81
0.03-3.38
0.33
0.13-0.86
0.10
0.19
0.03
0.02-0.43
0.04-0.94
0.00-0.39
0.10
0.04-0.27
(a) mRS > 4 at 12 months
DECIMAL
DESTINY
HAMLET
14 / 18
10 / 15
8/9
5 / 20
4 / 17
4 / 14
52.8
43.1
60.3
25.8-79.8
11.9-74.4
29.0-91.6
TOTAL
32 / 42
13 / 51
51.2
33.9-68.5
-10 0
Significance: p < 0.0001
Heterogeneity: p = 0.74
10
20
30 40
50
60 70
80
(b) mRS > 3 at 12 months
DECIMAL
DESTINY
HAMLET
14 / 18
11 / 15
8/9
10 / 20
9 / 17
10 / 14
27.8
20.4
17.5
-1.4-56.9
-12.2-53.0
-13.9-48.8
TOTAL
33 / 42
29 / 51
22.7
4.6-40.9
-10 0
Significance: p = 0.014
Heterogeneity: p = 0.89
10
20
30 40
50
60 70
80
(c) Death at 12 months
DECIMAL
DESTINY
HAMLET
14 / 18
8 / 15
8/9
5 / 20
3 / 17
3 / 14
52.8
35.7
67.5
25.8-79.8
4.6-66.8
37.7-97.2
TOTAL
30 / 42
11 / 51
50.3
33.3-67.4
Significance: p < 0.0001
Heterogeneity: p = 0.34
-10 0
10
20
30 40
50
60 70
80
% Absolute Risk Reduction (95% CI)
Discussion
This pooled analysis of randomized trials conﬁrms suggestions from uncontrolled
studies that decompressive surgery performed within 48 hours of stroke onset reduces
mortality and increases the number of patients with a favorable functional outcome
after space-occupying hemispheric infarction.9,10,12 The numbers needed to treat for
survival, survival with an mRS of 4 or lower, and survival with an mRS of 3 or lower
are 2, 2, and 4 respectively.
The present study is the ﬁrst in the ﬁeld of stroke in which a pooled analysis of individual
patient data from three independent randomized trials was planned while these trials
were still ongoing. This has the obvious advantage of being able to keep the number
of patients included to a minimum and to report the results several years earlier than
would have been possible based on the individual trials alone. The (expected) results
of this pooled analysis have led to premature termination of DESTINY. DECIMAL had
153
Chapter 11
Figure 3 Subgroup analyses of outcome according to age, timing of randomization, and presence of aphasia.
Outcome / Patients
ARR (%)
95% CI
OR
95% CI
3 / 16
1 / 13
4/6
54.6
52.3
16.7
25.1-84.0
18.7-86.0
-31.4-64.8
0.08
0.06
0.40
0.02-0.46
0.01-0.61
0.03-6.18
8 / 35
46.9
26.7-67.0
0.10
0.03-0.35
0.14
0.75
0.01
0.00-4.47
0.03-17.51
0.00-0.51
0.13
0.02-0.76
0.08
0.06
2.00
0.02-0.42
0.00-0.82
0.05-78.25
0.12
0.04-0.43
0.33
0.33
0.02
0.01-16.80
0.04-2.87
0.00-0.42
0.13
0.03-0.54
0.06
0.13
0.03
0.00-0.82
0.01-2.18
0.00-0.74
0.06
0.01-0.31
0.11
0.14
0.25
0.02-0.78
0.02-1.03
0.01-4.73
0.14
0.04-0.50
Conservative
Surgery
Age < 50 years
DECIMAL
DESTINY
HAMLET
11 / 15
6 / 10
5/6
TOTAL
22 / 31
-20 -10 0
Significance: p < 0.0001
Heterogeneity: p = 0.39
Age 50 years
DECIMAL
DESTINY
HAMLET
TOTAL
3/3
4/5
3/3
2/4
3/4
0/8
37.5
5.0
81.9
-17.0-92.0
-50.0-60.0
46.2-117.6
10 / 11
5 / 16
44.5
17.0-72.1
-20 -10 0
Significance: p = 0.0015
Heterogeneity: p = 0.044
Time to randomization < 24 hours
13 / 16
DECIMAL
7/8
DESTINY
HAMLET
1/2
TOTAL
21 / 26
5 / 19
2/7
2/3
54.9
58.9
-16.7
27.4-82.5
18.4-99.5
-104.1-70.8
9 / 29
49.7
27.6-71.9
-20 -10 0
Significance: p < 0.0001
Heterogeneity: p = 0.28
Time to randomization 24 hours
1/2
0/1
DECIMAL
3/7
2 / 10
DESTINY
7/7
2 / 11
HAMLET
25.0
22.9
72.9
-57.5-107.5
-21.4-67.1
44.5-101.4
TOTAL
46.9
22.3-71.4
11 / 16
4 / 22
-20 -10 0
Significance: p = 0.0002
Heterogeneity: p = 0.099
No aphasia
DECIMAL
DESTINY
HAMLET
TOTAL
5/7
3/4
6/6
1/8
2/7
2/8
58.9
46.4
65.1
18.4-99.5
-7.6-100.5
30.1-100.0
14 / 17
5 / 23
58.2
34.1-82.3
-20 -10 0
Significance: p < 0.0001
Heterogeneity: p = 0.85
Aphasia
DECIMAL
DESTINY
HAMLET
9 / 11
7 / 11
2/3
4 / 12
2 / 10
2/6
48.5
43.6
33.3
13.4-83.6
5.9-81.4
-32.0-98.7
TOTAL
18 / 25
8 / 28
44.2
20.2-68.1
Significance: p = 0.0003
Heterogeneity: p = 0.92
154
-20 -10 0
10 20
10 20
10 20
10 20
10 20
10 20
30
30
30
30
30
30
40 50
40 50
40 50
40 50
40 50
40 50
60
60
60
60
60
60
70 80
70 80
70 80
70 80
70 80
70 80
% Absolute Risk Reduction (95% CI)
90
90
90
90
90
90
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
been terminated after the diﬀerence in mortality had become signiﬁcant and the data
monitoring committee had recommended stopping enrollment
Although a score on the mRS of ≤ 3 is generally accepted as a favorable outcome in
stroke research, an mRS ≤ 4 after 12 months was chosen as the primary outcome in this
pooled analysis. Given that survival with no or only slight disability after large MCA
infarction is rare, the primary aim of the study was to assess whether decompressive
surgery reduced mortality without an increase in the number of severely disabled
survivors (mRS score of 5). However, decompressive surgery did in fact result in a
statistically signiﬁcant increase in survival with an mRS ≤ 3 after 12 months.
The present study shows that after decompressive surgery, the probability of survival
increases from 28% to nearly 80% and the probability of survival with an mRS of ≤
3 almost doubles. However, the probability of surviving in a condition requiring
assistance from others (mRS of 4) increases more than ten times, although very severe
disability (mRS of 5) is not increased. The choice of performing decompressive surgery
in an individual patient with space-occupying hemispheric infarction will therefore
depend on the willingness to accept survival with moderate disability. Information
on quality of life of survivors is essential for guiding such decisions. Previous studies
on quality of life after decompressive surgery for space-occupying infarction have
reported divergent results.21-23 Even patients with aphasia may improve signiﬁcantly.24
Information on quality of life will be provided in the separate publications of the
trials.
In the three trials under study, patients were excluded if they were older than 55
or 60 years of age. The results can probably not be generalized to patients who are
older. In a systematic review of uncontrolled studies on decompressive surgery, 80%
of the patients older than 50 years were dead or remained severely disabled compared
with 32% of the patients ≤ 50 years.12 Moreover, quality of life may remain impaired
especially in older patients.22,23
Data from a large uncontrolled series have suggested that outcome is substantially
improved if treatment is initiated within 24 hours of stroke onset as compared with
longer time windows for treatment.10 In the above-mentioned systematic review,
the timing of surgery did not aﬀect outcome. Similar observations were made in a
recent series of patients, in which the mean interval from stroke onset to surgery was
47 hours.25 In the present study, given the limited patient numbers, no diﬀerence in
outcome was found between patients treated on the ﬁrst and those treated on the
second day. In the majority of patients clinical sings of herniation appear after 2 days
of stroke onset.5 Whether decompressive surgery is also beneﬁcial if performed after
the ﬁrst 48 hours, is currently tested in HAMLET.13
The present study has limitations. First, as a result of slightly diﬀerent eligibility
criteria between the individual trials there were diﬀerences in baseline characteristics
of the patients included. However, there was no statistically signiﬁcant heterogeneity
between the trials in the eﬀect of surgery on any of the outcome measures. Second,
155
Chapter 11
as with most surgical trials, the nature of the treatment under study prevented a
fully blinded outcome assessment. Although observer bias cannot be excluded, the
consistency of results across the three trials, of which two used some form of blinding,
argues against any major bias. Third, subgroup analyses on expected prognostic
factors, such as age and the interval between the onset of symptoms and treatment
were not powered to show quantitative diﬀerences in treatment eﬀect between groups.
However, surgery was signiﬁcantly beneﬁcial in all subgroups, suggesting that there
are unlikely to be any qualitative subgroup-treatment eﬀect interactions (i.e. harm
in one group and beneﬁt in another). Fourth, there were baseline incomparibilities
between the treatment groups. However, analyses adjusted for baseline diﬀerences
provided essentially the same results.
In conclusion, decompressive surgery increases the probability of survival without
increasing the number of very severely disabled survivors. Still, the decision to perform
decompressive surgery should be made on an individual basis in every patient.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
156
Shaw C, Alvord Jr E, Berry E. Swelling of the brain following ischemic infarction with arterial occlusion.
Arch Neurol 1959;1:161-77.
Frank JI. Large hemispheric infarction, deterioration, and intracranial pressure. Neurology
1995;45:1286-1290.
Sakai K, Iwahashi K, Terada K, Gohda Y, Sakurai M, Matsumoto Y. Outcome after external
decompression for massive cerebral infarction. Neurol Med Chir Tokyo 1998;38:131-135.
Qureshi AI, Suarez JI, Yahia AM. Timing of neurologic deterioration in massive middle cerebral artery
infarction: a multicenter review. Crit Care Med 2003;31:272-277.
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 1996;53:309-315.
Berrouschot J, Sterker M, Bettin S, Koster J, Schneider D. Mortality of space-occupying (‘malignant’)
middle cerebral artery infarction under conservative intensive care. Intensive Care Med 1998;24:620623.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral infarction. Crit
Care Med 2003;31:617-625.
Kasner SE, Demchuk AM, Berrouschot J. Predictors of fatal brain edema in massive hemispheric
ischemic stroke. Stroke 2001;32:2117-2123.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Morley NC, Berge E, Cruz-Flores S, Whittle IR. Surgical decompression for cerebral edema in acute
ischaemic stroke. Cochrane Database Syst Rev 2002;CD003435.
Gupta R, Connolly ES, Mayer S, Elkind MS. Hemicraniectomy for massive middle cerebral artery
territory infarction: a systematic review. Stroke 2004;35:539-543.
Decompressive surgery in MCA infarction: pooled analysis of 3 trials
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Macleod MR, Kappelle LJ, van der Worp HB.
Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema Trial
(HAMLET). Protocol for a randomised controlled trial of decompressive surgery in space-occupying
hemispheric infarction. Trials 2006;7:29.
van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van GJ. Interobserver agreement for the
assessment of handicap in stroke patients. Stroke 1988;19:604-607.
Wirtz CR, Steiner T, Aschoﬀ A. Hemicraniectomy with dural augmentation in medically uncontrollable
hemispheric infarction. Neurosurg Focus 1997;2:E3.
Hacke W, Kaste M, Bogousslavsky J. European Stroke Initiative Recommendations for Stroke
Management-update 2003. Cerebrovasc Dis 2003;16:311-337.
Adams HP, Jr., Adams RJ, Brott T. Guidelines for the early management of patients with ischemic stroke:
A scientiﬁc statement from the Stroke Council of the American Stroke Association. Stroke
2003;34:1056-1083.
Bereczki D, Liu M, do PG, Fekete I. Mannitol for acute stroke (Cochrane Review). Cochrane Database
Syst Rev 2001 Jan;1:CD001153.
Brott T, Adams HP, Jr., Olinger CP. Measurements of acute cerebral infarction: a clinical examination
scale. Stroke 1989;20:864-870.
McCullagh P. Regression Models for Ordinal Data. Journal of the Royal Statistical Society, Series B
1980;42:109-42.
Woertgen C, Erban P, Rothoerl RD, Bein T, Horn M, Brawanski A. Quality of life after decompressive
craniectomy in patients suﬀering from supratentorial brain ischemia. Acta Neurochir (Wien )
2004;146:691-695.
Foerch C, Lang JM, Krause J. Functional impairment, disability, and quality of life outcome after
decompressive hemicraniectomy in malignant middle cerebral artery infarction. J Neurosurg
2004;101:248-254.
Curry WT, Jr., Sethi MK, Ogilvy CS, Carter BS. Factors associated with outcome after hemicraniectomy
for large middle cerebral artery territory infarction. Neurosurgery 2005;56:681-692.
Kastrau F, Wolter M, Huber W, Block F. Recovery from aphasia after hemicraniectomy for infarction
of the speech-dominant hemisphere. Stroke 2005;36:825-829.
Rabinstein AA, Mueller-Kronast N, Maramattom BV. Factors predicting prognosis after decompressive
hemicraniectomy for hemispheric infarction. Neurology 2006;67:891-893.
157
Discussion
Chapter 12
General discussion
159
Chapter 12
T
his thesis describes the preparation and preliminary results of a randomized
clinical trial to study the eﬀect of decompressive surgery in patients with
middle cerebral artery (MCA) infarction who deteriorate as a result of spaceoccupying brain edema. The rationale of surgical decompression is to accommodate shift of brain tissue and to normalize intracranial pressure, thereby preserving
cerebral blood ﬂow and preventing secondary damage. This treatment has been
proposed in the past, since patients with a hemispheric infarct and massive edema
have a poor prognosis, with a case fatality rate of about 80%, despite maximal medical
therapy in intensive care-based series.1,2 Moreover, there is no medical treatment of
proven eﬃcacy (chapter 2).
Case reports and small patient series have suggested a substantial reduction in death
rate after surgical decompression.3-7 As decompressive surgery may reduce brain
tissue shifts and intracranial pressure, but is not expected to reduce tissue damage in
the infarct itself, most patients will be left with a severe neurological deﬁcit.8 Given
this deﬁcit, many neurologists are reluctant to recommend hemicraniectomy as a lifesaving treatment in patients who deteriorate as a result of edema formation. In 2001,
an interview among Dutch neurologists disclosed that this intervention had been
performed in only 3 of 15 hospitals with neurosurgical facilities, and only sporadically.
However, other experts think decompressive surgery should be implemented as a
standard treatment modality, because uncontrolled studies showed a large reduction
in case fatality, without increasing the amount of severely disabled survivors.6,7 In
agreement with the current standard of evidence-based medicine (EBM), we therefor
initiated a randomized controlled trial (RCT).9
Expert opinion and evidence-based medicine
EBM has replaced pathophysiological reasoning and expert opinion as the basis of
clinical decision making.9 EBM relies on systematic, reproducible, unbiased clinical
observation, and marks pathophysiological reasoning and expert opinion as insuﬃcient,
or even inaccurate, in guiding deﬁnite clinical decisions. The two pillars of EBM are
RCTs and meta-analyses,9 which provide the best (Level 1) evidence. Observational
studies, such as cohort and case-control studies are assigned a level of evidence of
2. Level 3, the lowest level, is assigned to consensus guidelines, expert opinion, and
common clinical practice.9
The term EBM seems to imply that clinical medicine which is not based on RCTs or
meta-analyses, is not based on evidence. Before the concept of EBM was introduced,
experienced authority, manifested as the ‘expert opinion’, was assumed to be the
highest source of knowledge in clinical medicine. Collecting knowledge depended on
accumulating observations. The more observations were made (the greater the clinical
experience), the more conﬁdently knowledge could be extrapolated.10 This way of
collecting knowledge may be referred to as ‘inductive logic’, which was well accepted
160
Discussion
according to research traditions after the Enlightenment.
The role of expert opinion in the context of EBM has been largely ignored. EBM
assumes a reality, which is only approximately knowable, based on statistical
probability. The basic mode of investigation involves hypothesis testing. Therefore, it
would be tempting to describe EBM as based on ‘deductive logic’. The major problem
with this interpretation is that the hypotheses tested in EBM are not axiomas, but
ideas formulated by clinical experts, and thereby the result of induction. Moreover,
if diﬀerent RCTs on the same treatment would show conﬂicting results, and metaanalyses are not conclusive, consensus on the implementation of the treatment under
study still has to be obtained from experts in the ﬁeld. The logic involved in the process
of implementing results of RCTs is inductive as well. The greater the accumulated
evidence, the greater the conﬁdence in the conclusion. Although EBM is built on the
premise that expert opinion is not the best source for solving medical problems, it
relies on expert opinion in its ﬁrst stages (hypothesis formulation) and in its ﬁnal
stages (accepting the accumulated evidence as suﬃcient). Moreover, access to medical
journals for the introduction of new treatment strategies may be easier for authors
with authority in the ﬁeld than for relative newcomers, which may create publication
bias by expert opinion.
Until very recently, the decision to perform hemicraniectomy in patients with
space-occupying MCA infarction was based on ‘expert opinion’, extrapolated
from uncontrolled patient series. Most uncontrolled and controlled trials on
hemicraniectomy in patients with space-occupying infarction have been conducted by
the same experts with belief in the beneﬁt of the treatment. This inﬂuences the a priori
probability of conﬁrming the hypothesis that surgical decompression is of beneﬁt in
patients with space-occupying MCA infarction; thereby this belief also inﬂuences
interpretation and presentation of trial data.
The individual patient
In an RCT groups of patients are analyzed, whereas in clinical practice the individual
patient is the unit of observation. To implement the results of RCTs into clinical
practice in a very strict way, each patient should be reduced to a set of prognostic
factors and subsequently compared with similar patients treated in the trial. For RCTs
with reasonably objective outcome measures (for example a new stroke), weighing
the available evidence may be relatively easy. However, for trials with less objective
outcome measures, such as functional outcome and quality of life, advantages and
disadvantages are a matter of personal judgment. Interpretations become important
in the process of implementing scientiﬁc evidence in patient care and therefore it is
tempting to standardize treatment decisions.
Scientiﬁc reasoning is an essential part of clinical decision making. However,
161
Chapter 12
clinical medicine not only consists of science, but also of humanitariarism.11 Scientiﬁc
thinking is based on observation and deduction, whereas humanitarian thinking is
based on empathy and ethical norms. Treatments that can be given are not always
the ones that ought to be given. Clinical decisions are not solely based on the results
of scientiﬁc research but also on value judgments. In a prospective study among nonselected consecutive patients admitted in a general hospital, approximately 25 percent
of the treatment decisions was inﬂuenced by ethical problems or value judgments.12
Value judgments become more important if subjective measures, such as quality of
life, are concerned.
Ethical norms vary between doctors, patients, and societies, and so do ideas about
quality of life. Despite the lack of Level 1 evidence, almost every neurologist believes
that decompressive surgery reduces the risk of death in patients with space-occupying
MCA infarction. However, since a substantial proportion of the surviving patients is
left severely disabled, the answers to the question whether or not this treatment is
‘good’ are divergent. Some experts believe it is unethical to withhold decompressive
surgery because of the reduction in death rate and the reputedly modest proportion
of severely disabled survivors. Others think the neurological deﬁcit in survivors is so
severe that keeping patients alive is unethical. This disagreement is probably not only
the result of diﬀerent doctors seeing patients with diﬀerent infarcts. Personal ideas
about the value and quality of life with a severe neurological deﬁcit also play a role. An
RCT may help to assess handicaps of patients surviving after space-occupying MCA
infarction with or without hemicraniectomy in a systematic fashion. However, it will
not solve the divergence of opinion about the value of life of survivors.
Still, the results of HAMLET and other RCTs described in this thesis will undoubtedly
play an important role in the decision whether or not to perform hemicraniectomy in
a future patient with a life-threatening MCA infarct. Utilitarian considerations play an
important role in the process of these kinds of clinical decisions. These considerations
require not only the valuation of the outcomes of the two treatment options, but also
assessment of the probability that either of the outcomes will occur. The estimation
of the probability of diﬀerent outcomes requires empirical evidence; this aspect can
certainly be provided by an RCT.11
Animal experiments
Experimental stroke research has contributed much to the understanding of ischemic
brain damage, but its contribution to the current management of stroke patients is
poor.13 More than 500 neuroprotective agents have been shown to reduce damage in
models of cerebral ischemia,14 but the only treatment for patients with acute ischemic
stroke that has been proven eﬀective in RCTs is trombolysis.15 Remarkably, the
thrombolysis trials were not based on animal experiments.
162
Discussion
There are several explanations for the discrepancies between experimental and
clinical studies. Functional outcome is the most relevant outcome measure in clinical
trials. The relationship between infarct volume and functional outcome of patients
is marginal at best,16,17 let alone the relationship between infarct volume in animals
and functional outcome in patients. Extrapolation from animal models to the clinic
becomes even more diﬃcult as the severity of the clinical condition increases.18
Therefore, stroke studies in animals are largely explanatory, and should be extrapolated
to the clinic only with great care.
Although animal research mainly is intended for explanatory purposes, its predictive
validity is strongly dependent on the choice of the animal model. Previous studies on
the temporal evolution of cerebral infarction in rats, and studies on hemicraniectomy
in rats with space-occupying infarction, have been performed with models of variable
or moderate focal cerebral ischemia and may therefore not reﬂect the syndrome
of space-occupying infarction.19-22 In chapter 3 a more adequate model of spaceoccupying infarction is presented, with consistently large infarct volumes, substantial
midline shift, and high death rates, reﬂecting this clinical syndrome in man.
This model was used to study the course of MRI indices of cerebral perfusion and
tissue damage before (chapters 4) and after (chapter 5) decompressive surgery. It was
shown that in the presence of space-occupying edema, perfusion values were reduced
not only in the core of the infarct, but also in ‘healthy’ tissue surrounding the infarct.
After hemicraniectomy, cerebral blood ﬂow increased in the cerebral tissue directly
surrounding the infarct, potentially protecting this tissue from secondary ischemic
damage. This is in line with the hypothesis formulated to explain a potential beneﬁt
of decompressive surgery on functional outcome7 and implies that patients without
clinical signs of herniation might beneﬁt from this treatment as well. However, despite
a pseudo-normalization of apparent diﬀusion coeﬃcient (ADC) values within the
infarct core, histological examination revealed irreversible damage of brain tissue in
the complete territory of the MCA, indicating severe permanent damage.
Ethical aspects of animal experiments
The recognition of the intrinsic value of animals is one of the main aspects of the
legislation of animal experiments. The European Science Foundation acknowledges,
among other rules, that infringement of the intrinsic value is acceptable only if the
consequences for humanity of not performing the experiments are more important
than the inconvenience caused to the animal (www.esf.org). Since most experimental
research has an explanatory nature, the importance for mankind of many animal
experiments can be doubted. A critical reappraisal of the acceptability of these
experiments may reveal inconsistencies in the handling of this rule.
Acknowledgement of the intrinsic value of animals leads to equality of animals and
163
Chapter 12
humans in many moral respects. In case of approximately moral equality, criteria used
for the appraisal of experiments on animals and patients should be similar. However,
in general, ethical committees are much more rigid in permitting clinical trials with
patients than in permitting experiments on animals. Also, methodological ﬂaws have
been readily accepted in experimental research, whereas statistical rules of clinical
trials are very strict.23 Moreover, it is almost impossible to publish negative results of
animal experiments, whereas in clinical trials the ﬁnding that a treatment modality
does not change outcome measures is often judged almost equally valuable as the
opposite result. Publication bias caused by the impossibility to publish negative results
undoubtedly results in unnecessarily repetition of animal experiments and unjustiﬁed
initiation of clinical trials.23,24
Treatment of space-occupying MCA infarction
Decompressive surgery has to be considered in every patient with a large hemispheric
infarct deteriorating as a result of brain edema formation in the ﬁrst 48 h after stroke
onset, because in the pooled interim analysis of individual patient data of three recent
randomized controlled clinical trials a favorable outcome was signiﬁcantly more likely
after hemicraniectomy than after conservative treatment (chapter 11).
Several considerations have to be taken into account. Although favorable outcome
was more likely after hemicraniectomy, a large proportion of surviving patients was
left with severe disability. In the primary analysis, favorable outcome was deﬁned as a
score on the modiﬁed Rankin Scale (mRS)25 of 4 or better, which means that patients
could be ‘moderately disabled’. By deﬁnition, these patients were unable to walk
independently (with or without a cane or walker) and were unable to attend to their
own bodily needs without assistance. The choice of performing decompressive surgery
in an individual patient with space-occupying hemispheric infarction will therefore
depend on the willingness to accept survival with a moderate to severe disability.
Neuropsychological examination in a cohort of patients after decompressive
surgery for space-occupying hemispheric infarction showed severe cognitive deﬁcits
in most patients (chapter 10). In patients with aphasia, language was compromised to
such an extent that communication was very diﬃcult or even impossible. Apart from
severe focal deﬁcits general cognitive decline was found in most patients. Cognitive
impairment after stroke has been reported in several studies,26 and may limit the
patient’s ability to achieve life satisfaction.27 Furthermore, neuropsychiatric disorders,
such as depression, occur in 30% to 40% of stroke survivors.28 Not only motor function,
but also depression, cognitive impairment, and incontinence were identiﬁed as major
determinants of health-related quality of life after stroke.29 Information on quality of
life of survivors is essential for treatment decisions in patients with space-occupying
infarction and will be provided in future publications of the trials.
164
Discussion
In controlled intervention studies on hemicraniectomy in patients with spaceoccupying infarction such as HAMLET30 and other trials, feasibility considerations play
a role. Recruiting patients is diﬃcult and laborious. For ethical reasons, the number
of patients to be randomized should be kept to a minimum and if hemicraniectomy is
of beneﬁt for the group as a whole, trial results should be made public to implement
this treatment in clinical practice. However, since impairments may diﬀer between
diﬀerent subgroups of patients, e.g. those with aphasia, those of old age, and those
with concomitant infarction in the posterior or anterior cerebral artery territory, it will
remain questionable whether hemicraniectomy should be introduced as a standard
treatment modality for every patient with impending herniation after cerebral
infarction. Early termination of clinical trials may lead to exaggerated treatment
eﬀects, inﬂuencing decision making from the level of the individual patient to the level
of international guidelines.31 Early termination may also prevent suﬃcient subgroup
analyses and carries the risk that surgery will be performed on the basis of ‘level 1’
evidence in patients who will not truly beneﬁt. Subgroup analyses should at least focus
on age, interval to treatment, presence of aphasia, and involvement of other vascular
territories besides that of the MCA.
With the projected number of 112 inclusions, HAMLET is a relatively small trial.
To prevent loss of power without introducing selection bias, an alternative method
of randomizing patients is used (chapter 7). Because hemicraniectomy is supposed
to reduce case fatality to a large extent, randomized treatment allocation may be
considered unethical and unacceptable by the patients’ legal representatives. However,
structured interviews showed no diﬀerences in trial appreciation between HAMLET
participants and participants in a less dramatic trial, which investigates the eﬀect of
paracetamol in stroke patients (chapter 9).
The results of the pooled analysis play an important role in the decision to perform
early hemicraniectomy in patients with life-threatening MCA infarction. However,
weighing the pros and cons will to a large extent remain an individual judgement.
Moreover, quality of life and depression are important secondary outcome measures
in the separate clinical trials; the full reports should be awaited before the treatment
is implemented as standard. For individual patients, the ‘best clinical evidence’
derived from RCTs should be combined with individual expertise and personal value
judgments. In case of the decision to withhold surgical treatment because of severe
neurological damage, the indication for medical treatment can be doubted for the
same reason.
Future research
Decompressive hemicraniectomy has to be considered on an individual basis in
patients with a large hemispheric infarct who deteriorate as a result of progressive
165
Chapter 12
brain edema. Future research should focus on patient selection. It remains unclear
which patients with MCA infarction are at risk for developing life-threatening edema
(chapter 6) and studies on this topic are necessary. It is unlikely that trials will be
repeated to answer the question which patients beneﬁt most from hemicraniectomy,
so data on subgroups should be largely extrapolated from ongoing RCTs, especially
with regard to age, interval to treatment, presence of aphasia, and involvement of
other vascular territories besides that of the MCA.
Timing of the intervention is another subject of interest. Hemicraniectomy not only
prevents death by herniation, but it may also prevent secondary ischemic damage by
improving cerebral perfusion (chapter 5). This suggests that patients without clinical
signs of herniation might beneﬁt as well. Clinical trials on hemicraniectomy in patients
with large MCA infarction at risk for massive edema formation may therefore be a
next step. Ultimately, with the help of suﬃciently sensitive and speciﬁc prognostic
factors it should become possible to inform individual patients on their prognosis with
or without hemicraniectomy.
Conclusions
•
•
•
•
A pooled analysis of individual patient data of ongoing randomized trials has
demonstrated that decompressive surgery performed within 48 hours after the
onset of symptoms in patients with space-occupying hemispheric infarction leads
to a reduction in case fatality and an increased probability of a favorable outcome.
Still, many patients are left severely disabled and it remains tempting to extrapolate
these negative examples to future patients. Early termination of ongoing trials
should be prevented, because subgroup analyses are necessary.
Although the decision to perform decompressive surgery in patients with spaceoccupying hemispheric infarction involves ethical and subjective considerations,
RCTs oﬀer important information on outcome of these patients. For individual
patients, the ‘best clinical evidence’ derived from RCTs should be judged in the
context of individual values and desires.
Future research should focus on prognostic factors to investigate which patients
with MCA infarction are at risk for developing life-threatening edema and which
patients beneﬁt most from hemicraniectomy.
Experimental evidence supports the notion that hemicraniectomy not only prevents
death by herniation, but may also prevent secondary ischemic damage, by improving
cerebral perfusion. This suggests that patients with MCA infarction at risk of
massive edema formation, but without life-threatening brain shift, may beneﬁt from
this treatment as well.
166
Discussion
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
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 1996;53:309-315.
Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction and early brain swelling:
progression and eﬀect of age on outcome. Mayo Clin Proc 1998;73:829-836.
Schwab S, Rieke K, Aschoﬀ A, Albert F, von Kummer R, Hacke W. Hemicraniotomy in space-occupying
hemispheric infarction: useful early intervention or desparate activism? Cerebrovasc Dis 1996;6:325329.
Carter BS, Ogilvy CS, Candia GJ, Rosas HD, Buonanno F. One-year outcome after decompressive
surgery for massive nondominant hemispheric infarction. Neurosurgery 1997;40:1168-1175.
Koh MS, Goh KY, Tung MY, Chan C. Is decompressive craniectomy for acute cerebral infarction of
any beneﬁt? Surg Neurol 2000;53:225-230.
Rieke K, Schwab S, Krieger D. Decompressive surgery in space-occupying hemispheric infarction:
results of an open, prospective trial. Crit Care Med 1995;23:1576-1587.
Schwab S, Steiner T, Aschoﬀ A. Early hemicraniectomy in patients with complete middle cerebral
artery infarction. Stroke 1998;29:1888-1893.
Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. Classiﬁcation and natural history of clinically
identiﬁable subtypes of cerebral infarction. Lancet 1991;337:1521-1526.
Evidence-Based Medicine Working Group. Evidence-based medicine. A new approach to teaching
the practice of medicine. JAMA 1992;268:2420-2425.
Kulkarni AV. The challenges of evidence-based medicine: a philosophical perspective. Med Health
Care Philos 2005;8:255-260.
Wulﬀ H, Gotzsche P. Medicine and the Humanities. In: Wulﬀ H, Gotzsche P, eds. Rational Diagnosis
and Treatment, third ed. Oxford: Blackwell Science, 2000:147-174.
Kollemorten I, Strandberg C, Thomsen BM. Ethical aspects of clinical decision-making. J Med Ethics
1981;7:67-69.
van der Worp HB. Treatment of acute ischaemic stroke with antioxidants. The gap between laboratory
and practice. Utrecht University, doctorate thesis. 1999
O’Collins VE, Macleod MR, Donnan GA, Horky LL, van der Worp BH, Howells DW. 1,026
experimental treatments in acute stroke. Ann Neurol 2006;59:467-477.
Tissue plasminogen activator for acute ischemic stroke. The National Institute of Neurological
Disorders and Stroke rt-PA Stroke Study Group. N Engl J Med 1995;333:1581-1587.
Saver JL, Johnston KC, Homer D. Infarct volume as a surrogate or auxiliary outcome measure in
ischemic stroke clinical trials. The RANTTAS Investigators. Stroke 1999;30:293-298.
Lyden PD, Zweiﬂer R, Mahdavi Z, Lonzo L. A rapid, reliable, and valid method for measuring infarct
and brain compartment volumes from computed tomographic scans. Stroke 1994;25:2421-2428.
Beynen AC, Kort WJ. Animal models. In: van Zutphen LFM, Baumans V, Beynen AC, eds. Handboek
proefdierkunde, 3 ed. Maarssen: Elsevier gezondheidszorg, 2001:202-211.
Doerﬂer A, Forsting M, Reith W. Decompressive craniectomy in a rat model of “malignant” cerebral
hemispheric stroke: experimental support for an aggressive therapeutic approach. J Neurosurg
1996;85:853-859.
Forsting M, Reith W, Schabitz WR. Decompressive craniectomy for cerebral infarction. An
experimental study in rats. Stroke 1995;26:259-264.
167
Chapter 12
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
168
Engelhorn T, Doerﬂer A, Kastrup A. Decompressive craniectomy, reperfusion, or a combination for
early treatment of acute “malignant” cerebral hemispheric stroke in rats? Potential mechanisms
studied by MRI. Stroke 1999;30:1456-1463.
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 2002;33:617-622.
van der Worp HB, de Haan P, Morrema E, Kalkman CJ. Methodological quality of animal studies on
neuroprotection in focal cerebral ischaemia. J Neurol 2005;252:1108-1114.
Macleod MR, O’Collins T, Howells DW, Donnan GA. Pooling of animal experimental data reveals
inﬂuence of study design and publication bias. Stroke 2004;35:1203-1208.
van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the
assessment of handicap in stroke patients. Stroke 1988;19:604-607.
Sachdev PS, Brodaty H, Valenzuela MJ, Lorentz LM, Koschera A. Progression of cognitive impairment
in stroke patients. Neurology 2004;63:1618-1623.
Clarke P, Marshall V, Black SE, Colantonio A. Well-being after stroke in Canadian seniors: ﬁndings
from the Canadian Study of Health and Aging. Stroke 2002;33:1016-1021.
Ramasubbu R, Patten SB. Eﬀect of depression on stroke morbidity and mortality. Can J Psychiatry
2003;48:250-257.
Haacke C, Althaus A, Spottke A, Siebert U, Back T, Dodel R. Long-term outcome after stroke:
evaluating health-related quality of life using utility measurements. Stroke 2006;37:193-198.
Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Macleod MR, Kappelle LJ, van der Worp HB; the
HAMLET investigators. Hemicraniectomy After Middle cerebral artery infarction with Lifethreatening Edema Trial (HAMLET). Protocol for a randomised controlled trial of decompressive
surgery in space-occupying hemispheric infarction. Trials 2006;7:29.
Altman DG, Schulz KF, Moher D. The revised CONSORT statement for reporting randomized trials:
explanation and elaboration. Ann Intern Med 2001;134:663-694.
Appendix
Appendix
169
Appendix
The HAMLET investigators
Executive Committee - A. Algra, G.J. Amelink, J. van Gijn, J. Hofmeijer (trial
coordinator), L.J. Kappelle, M.R. Macleod (UK national coordinator), and H.B. van
der Worp (principal investigator). The Steering Committee is constituted of the
principal investigators of each actively randomizing centre (S.F.T.M. de Bruijn, G.J.
Luijckx, R. van Oostenbrugge, J. Stam, and J.Th.J. Tans) and of the members of the
executive committee. Data Monitoring Committee - Y. van der Graaf , P.J. Koudstaal,
and A.I.R. Maas. Advisory Committee - G.W. van Dijk, W. Hacke (Chairman), C.J.
Kalkman, C.A.F. Tulleken, and C.A.C. Wijman. Research nurse - M. van Buuren
Participating centers and investigators (local principal investigators marked with *) University Medical Centre Utrecht: A. Algra, G.J. Amelink, J. van Gijn, J. Hofmeijer,
L.J. Kappelle, H.B. van der Worp, Academic Medical Centre, Amsterdam: J. Stam,* G.J.
Bouma, W.P. Vandertop, University Medical Centre Groningen: G.J. Luijckx,* J.J.A.
Mooij, J.D.M. Metzemakers, Academic Hospital Maastricht: R. van Oostenbrugge,*
J. Dings, Medical Centre Haaglanden, The Hague: J.Th.J. Tans,* J.A.L. Würzer, J.
Boiten Haga Hospital, The Hague: S.F.T.M. de Bruijn,* C.F.E. Hoﬀmann.
170
Summary
Summary
171
Summary
Patients with a hemispheric infarct and massive space-occupying edema formation
have a poor prognosis. Mortality rates of about 80% have been described, despite
maximal medical therapy on an intensive care unit.
The aim of the research presented in this thesis was to investigate the beneﬁt of
decompressive surgery in patients with middle cerebral artery (MCA) infarction who
deteriorate as a result of space-occupying brain edema. In part 1, experimental studies
on this subject are described. In part 2, preparations for and preliminary results of the
ongoing randomized Hemicraniectomy After MCA infarction with Life-threatening
Edema Trial (HAMLET) are presented.
Chapter 2 describes the diﬀerent treatment modalities for patients with MCA
infarction, deteriorating as a result of edema formation. Various treatment strategies
have been proposed to limit brain tissue shifts and to reduce intracranial pressure,
such as osmotic therapy, hyperventilation, and sedation with barbiturates. However,
there is no evidence of eﬃcacy of any of these therapies, and several reports suggest
that these are ineﬀective or even detrimental. It is concluded that there is no speciﬁc
treatment modality of proven eﬃcacy for patients with space-occupying hemispheric
infarction.
In chapter 3 a rat model of space-occupying cerebral infarction is presented, obtained
by permanent intra-luminal MCA occlusion in Fisher rats. The time courses of ischemic
damage and cerebral perfusion were studied by means of magnetic resonance imaging
(MRI). At 24 hours of ischemia, midline shift peaked and perfusion of the tissue
surrounding the infarct was maximally decreased, suggesting that this surrounding
tissue was at risk for secondary damage as a result of hypoperfusion. It is concluded
that therapies to reduce edema and to improve perfusion may be most eﬃcacious if
started before 24 hours of ischemia in this model.
Studies of cerebral perfusion were performed in the above-mentioned model with
two diﬀerent MRI techniques (chapter 4). The commonly used dynamic susceptibility
contrast (DSC)-MRI technique, which is dependent on intravenously injected contrast
agent, was compared with non-invasive perfusion-weighted MRI (Flow-sensitive
Alternating Inversion Recovery (FAIR)). FAIR correlated with DSC-MRI if perfusion
was preserved (r2 = 0.719, P < 0.001), but provided lower perfusion values in areas with
severe hypoperfusion (mean FAIR-aCBF ipsi- versus contralateral ratio and mean
DSC-rCBF ipsi- versus contralateral ratio 0.39 ± 0.14 and 0.50 ± 0.14 respectively,
P = 0.008). These areas were not only found within the infarct, but also in vascular
territories other than that of the MCA in the same hemisphere. These territories are at
risk for secondary damage and may be saved by eﬀective treatment.
In rats with space-occupying MCA infarction, perfusion improved in brain tissue
surrounding the infarct after surgical decompression (25 ± 9 ml/min/100g tissue before
172
Summary
and 38 ± 9 ml/min/100g tissue after surgical decompression, P = 0.035), indicating
that secondary damage may be prevented by decompression (chapter 5). Apparent
Diﬀusion Coeﬃcient values seemed to normalize within the infarct, but, according
to histological examination, this was a sham eﬀect. Thus, secondary damage in tissue
surrounding the infarct may be prevented by decompressive surgery, but the infarct
itself remains permanently damaged.
In chapter 6 the results of a meta-analysis of predictors of fatal edema formation in
patients with MCA infarction are presented. It is concluded that infarct size is the
major determinant. Unfortunately, single clinical and radiological variables, such
as age, body temperature, infarct size, involvement of other vascular territories, or
internal carotid artery occlusion lack suﬃcient predictive value to be useful in clinical
practice.
In chapter 7 a new method for randomizing patients to one of two treatments in a
clinical trial is described. The method is based on minimization, to which an element
of chance is added. The probability of allocating one of two treatments depends on the
actual imbalance between the groups at the time of randomization. In trial simulations
with diﬀerent numbers of patients and prognostic factors, lower levels of imbalance
between treatment groups and bias by predictability were obtained than with pure
randomization or minimization, respectively, as well as compared with allocation
according to a biased coin principle. The method may be of use in relatively small
trials, when loss of power as a result of imbalance between treatment groups becomes
important.
The study protocol of HAMLET is presented in chapter 8. HAMLET is a multi-centre,
randomized clinical trial aimed to compare the eﬃcacy of decompressive surgery
with best medical treatment on functional outcome in patients with supratentorial
infarction with life-threatening space-occupying edema. One hundred and twelve
patients will be included, aged between 18 and 60 years, with a large hemispheric
infarct with space-occupying edema that leads to a decrease in consciousness. The
primary outcome measure is functional outcome, as determined by the score on the
modiﬁed Rankin Scale, at one year. HAMLET started recruiting in September 2002.
As of January 1st, 58 patients had been included.
Because of the life-threatening situation at the time of study inclusion and the
suggested large reduction in mortality by surgical decompression, randomized
treatment allocation in HAMLET may be considered unethical and unacceptable. Legal
representatives of patients included in HAMLET were interviewed to study how they
had experienced the informed consent procedure and which trial information they
could recall after one year. The results are described in chapter 9. Most representatives
considered trial participation and randomization acceptable, despite the serious
173
Summary
illness of their relatives. Recall of trial information was better in representatives of
HAMLET patients than in relatives of patients with less severe strokes participating in
the Paracetamol (Acetaminophen) In Stroke (PAIS) study.
The results of neuropsychological testing in an unselected cohort of 10 patients with
space-occupying MCA infarction one year after decompressive surgery are presented
in chapter 10. In addition to the expected severe focal deﬁcits, such as aphasia and
neglect, global cognitive decline was found in the majority of patients. In patients with
infarcts in the dominant hemisphere, language was compromised to such an extent
that communication was very diﬃcult or even impossible. Despite similar scores on
the modiﬁed Rankin Scale, cognitive functioning in patients with an infarct in the
dominant hemisphere was worse than in patients with an infarct in the non-dominant
hemisphere.
To obtain suﬃcient data to reliably estimate the eﬀects of early decompressive surgery
in patients with space-occupying MCA infarction as soon as possible, individual data of
patients from three ongoing European randomized controlled trials treated within 48
hours after stroke onset were pooled. The results of the analysis of the pooled data are
summarized in chapter 11. Ninety-three patients were included. After decompressive
surgery, more patients had an mRS≤4 (75% vs 24%; pooled absolute risk reduction
(ARR) [95% conﬁdence interval (95%CI)]: 51% [34-69]), an mRS≤3 (43% vs 21%;
ARR: 23% [5-41]), and survived (78% vs 29%; ARR: 50% [33-67]). The eﬀect of surgery
was highly consistent across the three trials. Thus, decompressive surgery reduced
mortality and increased the probability of a good functional outcome. However, the
probability of surviving in a condition requiring assistance from others also increased
substantially. It is concluded that the choice of performing decompressive surgery in
an individual patient with space-occupying hemispheric infarction will depend on
the willingness to accept survival with moderate disability. Information on quality of
life of survivors is essential for guiding such decisions and will be provided in future
publications of the trials.
The implications of the studies reported in this thesis are discussed in chapter 12. It
is concluded that, although the decision to perform decompressive surgery in patients
with space-occupying hemispheric infarction comprises ethical and subjective aspects,
randomized clinical trials oﬀer important information on outcome of these patients.
For individual patients, the ‘best clinical evidence’ derived from randomized controlled
trials should be judged in the context of individual values and desires. Early termination
of ongoing trials should be prevented, because subgroup analyses are necessary, at least
with regard to age, interval to treatment, presence of aphasia, and involvement of other
vascular territories besides that of the MCA. Future research should focus on prognostic
factors to investigate which patients with MCA infarction are at risk of developing lifethreatening edema and which will beneﬁt from hemicraniectomy most.
174
Samenvatting
Samenvatting
175
Samenvatting
Grote infarcten in het stroomgebied van de arteria cerebri media (ACM) kunnen
gepaard gaan met oedeemvorming, die in ernstige gevallen kan leiden tot cerebrale
inklemming. Patiënten met een dergelijk ruimte-innemend herseninfarct hebben
hierdoor een slechte prognose; in prospectief onderzoek was de sterfte ongeveer
80%, ondanks maximale conservatieve therapie. Overlevenden zijn meestal ernstig
gehandicapt.
Omdat het eﬀect van conservatieve therapie waarschijnlijk gering is, adviseren
sommige neurologen en neurochirurgen om bij patiënten met een ruimte-innemend
herseninfarct een chirurgische decompressie uit te voeren door een groot deel van
het schedeldak en het harde hersenvlies weg te nemen. Door de operatie kan het
beschadigde hersenweefsel buiten de schedel zwellen, waardoor binnen de schedel
weefselverschuivingen en compressie van gezond hersenweefsel worden voorkomen.
Het doel van deze behandeling is de cerebrale doorbloeding te verbeteren en secundaire
hersenschade te voorkomen.
De meeste in de literatuur beschreven ervaringen met chirurgische decompressie
bij patiënten met ruimte-innemende herseninfarcten betreﬀen relatief kleine groepen
patiënten, met in het algemeen gunstige resultaten. In twee grotere prospectieve
onderzoeken was de sterfte bij conservatief behandelde patiënten ongeveer 78% en bij
patiënten die geopereerd waren respectievelijk 34 en 16%. Gesuggereerd werd dat de
kans op ernstige neurologische restverschijnselen niet toenam en dat de kans op een
goed herstel beter was naarmate de operatie eerder plaatsvond. Deze twee prospectieve
onderzoeken suggereren een groot gunstig eﬀect van chirurgische behandeling. De
onderzoeken kenden echter zoveel methodologische tekortkomingen dat de resultaten
niet direct geëxtrapoleerd kunnen worden naar de praktijk. Ook is onduidelijk hoe het
functionele herstel van de overlevende patiënten is.
In dit proefschrift wordt onderzoek beschreven naar het eﬀect van chirurgische
decompressie bij patiënten met een herseninfarct in het stroomgebied van de ACM
die neurologische achteruit gaan als gevolg van hersenoedeem. Deel 1 bestaat uit
dierexperimenteel onderzoek en deel 2 uit de voorbereidingen en de eerste resultaten
van een lopend multi-centrisch gerandomiseerd klinisch onderzoek naar het eﬀect
van chirurgische decompressie op het functionele herstel van patiënten met een ACM
infarct die achteruit gaan als gevolg van oedeem: de Hemicraniectomy After MCA
infarction with Life-threatening Edema Trial (HAMLET).
In hoofdstuk 2 wordt samengevat welke behandelingsmogelijkheden bestaan voor
patiënten met een ruimte-innemend herseninfarct. Er zijn diverse behandelingen
beschreven om oedeemvorming en intracraniële drukverhoging tegen te gaan.
Deze bestaan onder meer uit toediening van mannitol, glycerol, corticosteroïden
of furosemide, hyperventilatie en sedatie met barbituraten. Van geen van deze
behandelingen is in klinische trials een gunstig eﬀect aangetoond. Er wordt door
sommige onderzoekers zelfs gesuggereerd dat deze weinig eﬀectief of schadelijk zijn.
176
Samenvatting
Er bestaat geen bewezen eﬀectieve behandeling voor patiënten met een ACM infarct
die neurologisch verslechteren als gevolg van hersenoedeem.
In hoofdstuk 3 wordt een rattenmodel beschreven van het ruimte-innemend
herseninfarct. De infarcten zijn geïnduceerd door intraluminale occlusie van de ACM
bij Fisher ratten. Het beloop van de weefselschade en de doorbloeding zijn bestudeerd
met behulp van Magnetic Resonance Imaging (MRI). Na 24 uur was de zwelling
maximaal en was de doorbloeding van nog niet geïnfarceerd hersenweefsel rond het
infarct het sterkst verminderd. Dit suggereert dat het nog gezonde hersenweefsel rond
het infarct is bedreigd en dat therapieën om dit weefsel te behouden in dit model
binnen 24 uur na ontstaan van het infarct gestart moeten worden.
De doorbloeding van de hersenen in het model is uitgebreid bestudeerd (hoofdstuk
4). Daarvoor is gebruik gemaakt van twee MRI technieken, namelijk de standaard
techniek, waarbij contrastvloeistof wordt ingespoten (dynamic susceptibility contrast
magnetic resonance imaging (DSC-MRI)), en een nieuwere techniek, zonder contrast
vloeistof (ﬂow-sensitive alternating inversion recovery (FAIR)). De doorbloeding met
FAIR kwam grotendeels overeen met die volgens DSC-MRI, behalve in gebieden waar
de doorbloeding ernstig was afgenomen; hierin werd met FAIR een lagere doorbloeding
gemeten. Gebieden met een afgenomen doorbloeding werden niet alleen gevonden in
het infarct zelf, maar ook in hersenweefsel om het infarct heen. Voor dit hersenweefsel
bestaat het gevaar van secundaire beschadiging, die voorkomen zou kunnen worden
door de juiste behandeling.
In het hierboven beschreven model verbeterde de doorbloeding van hersenweefsel
rond het infarct na chirurgische decompressie (hoofdstuk 5). Dit betekent dat deze
behandeling secundaire schade zou kunnen voorkomen. Er was ook een verbetering
van de ‘Apparent Diﬀusion Coeﬃcient’ in het infarct zelf, die een maat is voor
beschadiging van hersenweefsel. Bij microscopisch onderzoek bleek deze verbetering
echter niet te berusten op daadwerkelijk herstel, maar waarschijnlijk op een toename
van oedeem. De schade in de herseninfarcten zelf, met de bijbehorende ernstige
neurologische uitval, werd door chirurgische decompressie dus niet verbeterd.
De resultaten van een meta-analyse van in de literatuur beschreven voorspellers van
levensbedreigend hersenoedeem bij patiënten met ACM infarcten zijn beschreven in
hoofdstuk 6. Infarctgrootte blijkt de belangrijkste determinant te zijn. Helaas werd
er geen klinische of radiologische variabele gevonden die levensbedreigend oedeem
voorspelt in een vroeg stadium, voordat patiënten daadwerkelijk achteruit gaan.
In hoofstuk 7 is een nieuwe methode beschreven om één van twee behandelingen
toe te wijzen in klinische trials. De methode is gebaseerd op minimisatie, waarbij
de behandeling wordt toegewezen louter op grond van de toewijzingen bij eerdere
patiënten, om ongelijkheid tussen de behandelingsgroepen zoveel mogelijk te
177
Samenvatting
beperken. Bij de nieuwe methode wordt een kanselement toegevoegd aan de
toewijzingsbeslissing op grond van minimisatie. De grootte van deze kans hangt af van
de ongelijkheid tussen de behandelingsgroepen. Met simulatietrials is aangetoond dat
met deze nieuwe methode selectiebias kleiner is dan bij minimisatie en de ongelijkheid
tussen de behandelingsgroepen kleiner dan bij pure randomisatie. Minder ongelijkheid
tussen de behandelingsgroepen zonder grote toename van selectie bias kan in trials
met weinig te includeren patiënten van belang kan zijn.
In Hoofdstuk 8 staat het onderzoeksprotocol van HAMLET. HAMLET is een open
gerandomiseerde klinische trial, waarin bij 112 patiënten met een ruimte-innemend
supratentorieel herseninfarct het herstel na chirurgische decompressie vergeleken
wordt met dat na conservatieve behandeling. Patiënten tot en met 60 jaar met een
ruimte-innemend infarct in het stroomgebied van de ACM kunnen binnen 96 h na
het ontstaan van de symptomen worden geïncludeerd, als de behandelend neuroloog
de indruk heeft dat er achteruitgang bestaat op basis van oedeemvorming. Bij deze
patiënten moeten ernstige neurologische uitval, geleidelijke achteruitgang van het
bewustzijn en ruimte-innemend oedeem op de CT- of MRI-scan van de hersenen
bestaan. De primaire uitkomstmaat is het functionele herstel na 1 jaar volgens de
gemodiﬁceerde Rankin Schaal (mRS). Daarnaast zal onder meer de kwaliteit van
leven onderzocht worden. De eerste patiënt werd geïncludeerd in september 2002 en
1 januari 2007 waren 58 patiënten geïncludeerd.
De vraag om toestemming voor deelname aan HAMLET wordt gesteld aan wettelijke
vertegenwoordigers van patiënten, omdat de patiënten zelf door een gedaald bewustzijn
niet wilsbekwaam zijn. De vraag om toestemming zou onacceptabel gevonden kunnen
worden door wettelijke vertegenwoordigers, gezien de levensbedreigende toestand en
het vermoede grote eﬀect van chirurgische decompressie op de kans van overleven.
Wettelijke vertegenwoordigers die toestemming gaven voor deelname aan HAMLET
van een familielid werden geïnterviewd, met als doel te onderzoeken hoe de vraag om
toestemming beleefd was en hoeveel informatie onthouden was (hoofdstuk 9). De meeste
wettelijke vertegenwoordigers vonden de vraag om toestemming acceptabel. Er was na
een jaar meer onderzoeksinformatie onthouden door wettelijke vertegenwoordigers
die toestemming gaven voor HAMLET dan door wettelijke vertegenwoordigers die
toestemming gaven voor een trial waarin een minder invasieve behandeling werd
onderzocht, namelijk het Paracetamol (Acetaminophen) In Stroke (PAIS) onderzoek.
In hoofdstuk 10 staan de resultaten van neuropsychologisch onderzoek van 10 patiënten
met een ruimte-innemend ACM infarct een jaar na chirurgische decompressie. Naast
de ernstige focale cognitieve uitvalsverschijnselen, die verwacht waren op grond van
de plaats en de grootte van het herseninfarct, werd bij 8 van de 10 patiënten globaal
cognitief verval vastgesteld. Bij patiënten met een infarct in de dominante hemisfeer
was de taalfunctie zo slecht dat communicatie vrijwel niet mogelijk was. Patiënten met
178
Samenvatting
een infarct in de dominante hemisfeer deden het bij neuropsychologisch onderzoek
slechter dan patiënten met een infarct in de niet-dominante hemisfeer, ondanks gelijke
uitkomst scores volgens de mRS. Voordat chirurgische decompressie wordt ingevoerd
als standaard behandeling bij patiënten met een ruimte-innemend herseninfarct,
moeten de resultaten van neuropsychologisch onderzoek van HAMLET worden
afgewacht.
Om het eﬀect van vroege chirurgische decompressie (binnen 48 uur na ontstaan van
het herseninfarct) bij patiënten met een ruimte-innemend herseninfarct eerder te
kunnen beoordelen dan op basis van kleine individuele trials mogelijk zou zijn werden
data van drie lopende gerandomiseerde Europese onderzoeken samengevoegd. De
resultaten van de analyse van deze data staan in hoofdstuk 11. Er werden 93 patiënten
geïncludeerd. Na chirurgische decompressie hadden meer patiënten een mRS≤4 (75%
vs 24%; gepoolde absolute risico reductie (ARR) [95% betrouwbaarheids interval
(95%CI)]: 51% [34-69]), een mRS≤3 (43% vs 21%; ARR: 23% [5-41]) en overleefden
meer patiënten (78% vs 29%; ARR: 50% [33-67]) dan na conservatieve behandeling.
Dit betekent ‘numbers needed to treat’ van 2 voor overleven met een mRS≤4, 4 voor
overleven met een mRS≤3 en 2 voor overleven ongeacht de uitkomst. Dat houdt in dat
chirurgische decompressie de mortaliteit reduceert met 50%, met een mRS≤3 bij 55%
van de overlevenden. Het eﬀect van chirurgische decompressie was consistent tussen
de drie onderzoeken.
De eindconclusies van het onderzoek worden bediscussieerd in hoofdstuk 12. De
beslissing om chirurgische decompressie te verrichten moet hoge mate individueel
genomen worden. Toch zijn gerandomiseerde onderzoeken van belang. De informatie
uit het onderzoek over de kans op herstel zal voor iedere patiënt in het licht van
persoonlijke ideeën over kwaliteit van leven beoordeeld moeten worden. Voortijdige
onderbreking van trials moet voorkomen worden, mede omdat subgroepanalyses
gewenst zijn, om te bepalen of bijvoorbeeld leeftijd, afasie, tijdstip van operatie en
infarcering van andere stroomgebieden naast dat van de ACM van belang zijn voor
het functionele herstel. Toekomstig onderzoek moet zich richten op prognostische
factoren, om erachter te komen welke patiënten met een ACM infarct het risico lopen
op levensbedreigend oedeem en welke patiënten het meest gebaat zijn bij chirurgische
decompressie.
179
hoofdstuk 2
Dankwoord
Dankwoord
181
Dankwoord
Velen wil ik dankzeggen voor hun bijdragen aan dit proefschrift.
Dr. H.B. van der Worp, beste Bart, je stond aan de basis van dit onderzoek. Je
handtekening staat in ieder hoofdstuk van dit proefschrift. Ieder idee, experiment of
artikel werd beter als het door jouw handen ging. Ik heb onze samenwerking plezierig
en leerzaam gevonden. Je relativeringsvermogen was een steun bij frustraties.
Aan de HAMLET promotie tour, in jouw mintgroene Volvo 360, bewaar ik goede
herinneringen. Je buitensporige controle behoefte en je onstuitbare neiging dezelfde
ﬂauwe grappen steeds te herhalen zijn je vergeven.
Prof. Dr. L.J. Kappelle, beste Jaap, jij bent een voorbeeld voor me op wetenschappelijk,
klinisch-neurologisch en (niet in de laatste plaats) menselijk vlak. Bedankt voor al je
hulp.
Prof. Dr. A. Algra, beste Ale, ik heb grote bewondering voor de prettige manier waarop
jij leermeester bent voor velen. Je methodologische bijdragen aan HAMLET en aan de
gepoolde analyse waren onmisbaar. Enorm bedankt.
Dr. G.J. Amelink, beste Hans, naast de nachtelijke operaties ad hoc waren je
ongenuanceerde relativerende relazen onontbeerlijk. Bedankt.
Prof. Dr. J. van Gijn, het is een eer door u opgeleid te zijn tot neuroloog. Ik dank u in
het bijzonder voor uw bijdrage aan mijn manuscripten, waardoor geschreven staat wat
bedoeld wordt.
Prof. Dr. K. Nicolay, beste Klaas, bedankt voor je hartelijke ontvangst en leerzame
begeleiding op de afdeling ‘experimentele in vivo NMR’ .
Marrit van Buuren, bedankt voor je bijdragen aan de data verzameling en de patiënten
follow up voor HAMLET, die groter zijn dan menigeen denkt.
Patiënten en hun familieleden die aan HAMLET hebben deelgenomen dank ik zeer
voor alle inspanningen.
Prof. Dr. J. Stam, Dr. G.J. Bouma, Prof. Dr. W.P. Vandertop, Dr. G.J. Luijckx, Prof. Dr.
J.J.A. Mooij, Dr. J.D.M. Metzemakers, Dr. R. van Oostenbrugge, Prof. Dr. J. Dings, Dr.
J.Th.J. Tans, Dr. J.A.L. Würzer, Dr. S.F.T.M. de Bruijn, Dr. C.F.E. Hoﬀmann en alle andere
neurologen, neurochirurgen en intensivisten uit alle deelnemende ziekenhuizen die
een bijdrage hebben geleverd aan HAMLET ben ik zeer erkentelijk. Alle Nederlandse
neurologen die patiënten voor deelname aan HAMLET naar één van de deelnemende
182
Dankwoord
ziekenhuizen hebben verwezen en verpleegkundigen die zich voor HAMLET hebben
ingezet bedank ik hiervoor.
Prof. Dr. Y. van der Graaf, Dr. A.I.R. Maas en Prof. Dr. P.J. Koudstaal, bedankt voor uw
inspanningen als Data Monitoring and Safety Committee.
Malcolm Macleod, thank you for your eﬀorts for HAMLET in the UK. Prof. Dr. W.
Hacke, dear Werner, thank you for giving advice in the early stages of the trial and
for chairing the advisory committee. The DECIMAL and DESTINY investigators
and Peter Rothwell, thank you for sharing and analyzing data in a pooled analysis of
decompressive surgery trials.
Mijn collega’s in de kliniek bedank ik voor hun amicale collegialiteit. De vriendschap
met Floor, de vele kopjes koﬃe en thee met Ynte en Marieke, de humoristische kleine
vakanties voor en na congressen met Walter, Geert-Jan, Bart, Floor, Ynte, Marieke, en
Wouter droegen in hoge mate bij aan plezier op de werkvloer. Gert van Dijk, bedankt
voor je steun. De zeiltochtjes behoren tot de hoogtepunten van mijn assistententijd.
Dop Bär maakte mijn eerste schreden op experimenteel onderzoeksgebied mogelijk.
In het laboratorium van de afdeling ‘experimentele in vivo NMR’ werd ik met raad
en daad bijgestaan door Gerard van Vliet, Wouter Veldhuis en Janneke Schepers.
Bedankt!
Ivonne en Esther, paranimfen en makkers vanaf de eerste dag, wat ben ik blij met
jullie!
Mijn ouders bedank ik voor hun onvoorwaardelijke steun en vertrouwen, mijn broer
voor zijn vriendschap en mijn zus, mijn beste vriendin en ware paranimf, voor alles
wat ze is en mij laat zijn.
Tenslotte Erwin, bedankt dat je het mogelijk maakt ‘werk’ en ‘thuis’ probleemloos te
combineren. Bedankt voor je geduld en al het goede dat wij samen delen.
183
Dankwoord
184
Curriculum Vitae
Curriculum Vitae
185
Curriculum Vitae
Jeannette Hofmeijer was born on December 20th, 1971. She ﬁnished secondary
education in 1991 (Corderius College, Amersfoort). She studied medicine and
philosophy at the Radboud University Nijmegen from 1991 to 1999. As a student, she
studied neuropsychological functioning of patients with myotonic dystrophy (prof.
dr. B.G.M. van Engelen). In 1999 she did internships in Tanzania at the Sengerema
District Designated Hospital (dr. M. J. Voeten). After obtaining her medical degree
(cum laude), she started working as a resident at the Department of Neurology of the
University Medical Center Utrecht (prof. dr. J. van Gijn) in 2000. In 2001 she worked
at the department of Experimental In Vivo NMR of the Utrecht University, where
she investigated tissue damage and perfusion in an animal model of space-occupying
cerebral infarction by means of MRI (prof. dr. K. Nicolay). From 2002 she coordinates
HAMLET (Hemicraniectomy After MCA-infarction with Life-threatening Edema
Trial), sponsored by the Netherlands Heart Foundation. In February 2007 she qualiﬁed
as a neurologist and started as a fellow at the Intensive Care Center of the University
Medical Center Utrecht (prof. dr. J. Keseçioglu). She lives together with Erwin Blezer.
They have a daughter, Fee, and are expecting a second baby in June 2007.
186
Publications
Publications
187
Publications
Hofmeijer J, Klijn CJ, Kappelle LJ, van Huﬀelen AC, van Gijn J. Collateral circulation via
the ophthalmic artery or leptomeningeal vessels is associated with impaired cerebral
vasoreactivity in patients with symptomatic carotid artery occlusion. Cerebrovasc Dis
2002;14:22-26.
Hoogendam A, Hofmeijer J, Frijns CJ, Heeringa M, Schouten T, Jansen PA. [Severe
parkinsonism due to metoclopramide in a patient with polypharmacy]. Ned Tijdschr
Geneeskd 2002;146:175-177.
Hofmeijer J, van der Worp HB, Amelink GJ, Algra A, van Gijn J, Kappelle LJ. [Surgical
decompression in space-occupying cerebral infarction; notiﬁcation of a randomized
trial]. Ned Tijdschr Geneeskd 2003;147:2594-2596.
Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying cerebral
infarction. Crit Care Med 2003;31:617-625.
Hofmeijer J, Veldhuis WB, Schepers J, Nicolay K, Kappelle LJ, Bär PR, van der Worp
HB. The time course of ischemic damage and cerebral perfusion in a rat model of
space-occupying cerebral infarction. Brain Res 2004;1013:74-82.
Hofmeijer J, Schepers J, Veldhuis WB, Nicolay K, Kappelle LJ, Bär PR, van der Worp
HB. Delayed decompressive surgery increases apparent diﬀusion coeﬃcient and
improves peri-infarct perfusion in rats with space-occupying cerebral infarction.
Stroke 2004;35:1476-1481.
Hofmeijer J, Schepers J, van der Worp HB, Kappelle LJ, Nicolay K. Perfusion MRI with
FAIR and DSC-MRI. A qualitative comparison in rats with permanent middle cerebral
artery occlusion. NMR Biomed 2005;18(6):390-4.
Hofmeijer J, Amelink GJ, Algra A, van Gijn J, Macleod MR, Kappelle LJ, van der Worp
HB. Hemicraniectomy after middle cerebral artery infarction with life-threatening
Edema trial (HAMLET). Protocol for a randomised controlled trial of decompressive
surgery in space-occupying hemispheric infarction. Trials 2006;7:29.
Hofmeijer J, van der Worp HB, Kappelle LJ. Complications of acute ischemic stroke
and their management. In: Adams-HP J, ed. Handbook of cerebrovascular diseases, 2
ed. New York: Marcel Dekker, 2006:183-204.
188
Publications
Vahedi K,* Hofmeijer J,* Juettler E,* Vicaut E, George B, Algra A, Amelink GJ,
Schmiedeck P, Schwab S, Rothwell PM, Bousser MG, van der Worp HB, Hacke W for
the DECIMAL, DESTINY, and HAMLET investigators. Early decompressive surgery
in malignant middle cerebral artery infarction: a pooled analysis of three randomized
controlled trials. Lancet Neurol 2007;6:215-222.
Hofmeijer J, Anema PC van der Tweel I. A new algorithm for sequential allocation of
two treatments in small clinical trials. J Clin Epidemiol Under revision.
Hofmeijer J, Amelink GJ, den Hertog H, Algra A, Kappelle LJ, van der Worp HB, on
behalf of the HAMLET and the PAIS investigators. Appreciation of the informed
consent procedure in a randomized trial of decompressive surgery for space-occupying
hemispheric infarction. JNNP Under revision.
Hofmeijer J, Algra A, Kappelle LJ, van der Worp HB. Predictors of life-threatening
brain edema in middle cerebral artery infarction. A systematic review. Submitted.
Hofmeijer J, van der Worp HB, Amelink GJ, Kappelle LJ, Nys GMS, van Zandvoort
MJE. Long-term cognitive outcome after decompressive surgery for space-occupying
hemispheric infarction. In preparation.
189