Magnesium sulfate in the management of patients with aneurysmal subarachnoid hemorrhage: a randomized, placebo-controlled, dose-adapted trial.

Abstract

Recent studies suggest that high-dose MgSO4 therapy is safe and reduces the incidence of DIND and subsequent poor outcome after SAH. We intended to assess the safety and efficacy of high-dose MgSO4 therapy after SAH as means to prevent DIND and to evaluate the impact on clinical outcome.
This was a prospective, randomized, single-blind, placebo-controlled study. The MgSO4 infusion was adjusted every 12 hours until day 12 according to the target serum Mg2+ level. The occurrence of DIND, secondary infarction, side effects, and the outcome after 3 and 12 months were assessed.
Fifty-eight patients were randomized; 27 received placebo and 31 MgSO4. The difference in occurrence of DIND and secondary infarction was not significant. The intention-to-treat analysis revealed a trend toward better outcome (P = .083) after 3 months. On-treatment analysis showed a significantly better outcome after 3 months (P = .017) and a trend toward better outcome after 1 year (P = .083). Significantly more often hypotension (P = .040) and hypocalcemia (P = .005) occurred as side effects in the treatment group. In 16 patients (52%), the MgSO4 therapy had to be stopped before day 12 because of side effects. No predictive factor leading to termination was found in a postrandomization analysis.
High-dose MgSO4 therapy might be efficient as a prophylactic adjacent therapy after SAH to reduce the risk for poor outcome. Nevertheless, because of the high frequency of the side effects, patients should be observed in an intensive or intermediate care setting.

Full text

Page 1

Vascular
Magnesium sulfate in the management of patients with
aneurysmal subarachnoid hemorrhage: a randomized,
placebo-controlled, dose-adapted trial
Carl Muroi, MD⁎, Andrej Terzic, MD, Mathias Fortunati, MD,
Yasuhiro Yonekawa, MD, Emanuela Keller, MD
Department of Neurosurgery, University Hospital Zurich, Zurich CH-8091, Switzerland
Received 6 March 2007; accepted 8 July 2007
AbstractBackground: Recent studies suggest that high-dose MgSO4 therapy is safe and reduces the
incidence of DIND and subsequent poor outcome after SAH. We intended to assess the safety and
efficacy of high-dose MgSO4therapy after SAH as means to prevent DIND and to evaluate the
impact on clinical outcome.
Methods: This was a prospective, randomized, single-blind, placebo-controlled study. The MgSO4
infusion was adjusted every 12 hours until day 12 according to the target serum Mg2+level.
The occurrence of DIND, secondary infarction, side effects, and the outcome after 3 and 12 months
were assessed.
Results: Fifty-eight patients were randomized; 27 received placebo and 31 MgSO4. The difference
in occurrence of DIND and secondary infarction was not significant. The intention-to-treat analysis
revealed a trend toward better outcome (P = .083) after 3 months. On-treatment analysis showed a
significantly better outcome after 3 months (P = .017) and a trend toward better outcome after 1 year
(P = .083). Significantly more often hypotension (P = .040) and hypocalcemia (P = .005) occurred as
side effects in the treatment group. In 16 patients (52%), the MgSO4therapy had to be stopped
before day 12 because of side effects. No predictive factor leading to termination was found in a
postrandomization analysis.
Conclusions: High-dose MgSO4therapy might be efficient as a prophylactic adjacent therapy after
SAH to reduce the risk for poor outcome. Nevertheless, because of the high frequency of the side
effects, patients should be observed in an intensive or intermediate care setting.
© 2008 Elsevier Inc. All rights reserved.
Keywords:Subarachnoid hemorrhage; Vasospasm; Magnesium sulfate; Randomized placebo-controlled dose-adapted trial
1. Introduction
With early aneurysm clipping, DIND due to cerebral
vasospasm has become the most common cause of death
and disability after aneurysmal SAH [6]. Recent studies
suggest that high-dose MgSO4therapy is safe and reduces
the incidence of delayed cerebral ischemia and subsequent
poor outcome after SAH (Table 1) [4,5,13,14,17-21,
23,24]. However, optimal dosages are under discussion
and results are not yet definitive. The objectives of the
present monocenter, prospective, randomized, single-blind,
Available online at www.sciencedirect.com
Surgical Neurology 69 (2008) 33–39
www.surgicalneurology-online.com
Abbreviations: Ca2+, calcium; CT, computed tomography; DIND,
delayed ischemic neurologic deficit; ECG, electrocardiogram; GOS,
Glasgow Outcome Scale; MCA, middle cerebral artery; Mg2+, magnesium;
MgSO4, magnesium sulfate; SAH, subarachnoid hemorrhage; TCD,
transcranial Doppler sonography; Vmean, mean blood flow velocity;
WFNS, World Federation of Neurological Surgeons.
⁎Corresponding author. Tel.: +41 44 255 2660; fax: +41 44 255 4505.
E-mail address: carl.muroi@usz.ch (C. Muroi).
0090-3019/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.surneu.2007.07.015

Page 2

Table 1
Comparison of previously published pilot studies with continuous MgSO4applications in patients with SAH
Study DesignnMgSO4appl.Dose and target level MonitoringEnd points Positive resultsComedicationSide effects
Boet and
Mee [4]
Monocenter,
uncontrolled
10Start within
5 d, for 10 d
Bolus: 20 mmol/20 min, cont.: 80 mmol/d
dose-adapted, target [Mg2+]: 2× baseline or
2.0-2.5 mmol/L
Bolus: 25 mmol/30 min, cont.: 144 mmol/d
dose-adapted, target [Mg2+]:
1.6-2.3 mmol/L
BP, ECG,
[Mg2+]
CVS: TCD and
clinical, GOS
after 3 mo
CVS: TCD and
clinical, GOS
after 3 mo
CVS: TCD (n = 5),
GOS 4-5: n = 8
Nimodipine No serious side effects,
transient flushed feeling
Veyna et al
[21]
Monocenter,
randomized,
placebo-controlled,
single-blind
Monocenter,
historical control
group
Monocenter,
uncontrolled,
dose finding study
40 Start within
3 d, for 10 d
BP, ECG,
[Mg2+], [Ca2+]
Trend in patients with
H&H II for better GOS
Nimodipine, phenytoin,
Ca-gluconate if [Ca2+]
b0.9 mmol/L
No serious side effects,
[Ca2+] lower in treatment
group, supplemented
Chia et al [5]
23Start within 5 d,
until discharge
from ICU
Start within
2 d, for 14 d
Bolus: none, cont.: 24-52 mmol/d
dose-adapted, target [Mg2+]: 1.0-1.5 mmol/L
BP, ECG,
[Mg2+]
CVS: angio,
outcome scale
(1-4) after 3 mo
DCI, rebleeding,
GOS
Significantly less angio
CVS
NimodipineNo serious side effects
Van den Bergh
et al [17]
14 Group A—bolus: 16 mmol, cont.:
16 mmol/d, group B—bolus: none, cont.:
30 mmol/d, group C—bolus: none, cont.:
64 mmol/d
Bolus: none, cont.: 64 mmol/d
BP, ECG,
[Mg2+] every
other day
Optimal dosage:
64 mmol/24h, no DCI
and all GOS 4-5 in
group C
See side effects
NimodipineNo serious side effects,
transient flushed feeling
in group A
Van Norden
et al [20]
Monocenter,
randomized,
placebo-controlled,
double-blind
Multicenter,
randomized,
placebo-controlled,
double-blind
94 Start within
4 d, for
14-18 d
BP, ECG,
[Mg2+] every
other day
Side effects and
termination
Nimodipine Termination due to
hypermagnesiemia (n = 3),
hypotension (n = 1),
renal failure (n = 1)
Termination due to
hypotension (n = 1),
bradycardia and atrial
fibrillation (n = 1),
hypermagnesemia (n = 1)
No serious side effects
Van den Bergh
et al [18]
283Start within
4 d, for
14-18 d
Bolus: none, cont.: 64 mmol/dBP, ECG,
[Mg2+] not
obligatory
DCI, new
hypodensity in CT
with DCI, mRS
after 3 mo
Risk reduction for DCI:
34%, risk reduction for
poor outcome: 23%
Nimodipine
Yahia et al [24] Monocenter,
uncontrolled
19Start within
3 d, for 10 d
Bolus: none, cont.: 97 mmol/d
dose-adapted, target [Mg2+]:
1.5 -4.0 mmol/L
BP, ECG,
[Mg2+],
[glucose],
[phenytoin]
BP, ECG,
[Mg2+]
CVS: TCD, angio,
clinical, new
hypodensity in CT,
GOS after 1 mo
CVS: TCD, clinical,
GOS and BI
after 6 mo
CVS: TCD n = 5, angio
n = 9, clinical n = 2,
CT: no ischemic infarction,
GOS 4-5: n = 18
Trend to less TCD and
clinical CVS, duration
of TCD CVS
significantly shorter
Patients with CVS
required less
hospitalization time
if they received mg2+
Nimodipine,
phenytoine
Wong et al [23]Monocenter,
randomized,
placebo-controlled,
double-blind
Monocenter,
controlled:
standard therapy +
MgSO4vs standard
therapy only
Monocenter,
randomized,
controlled: MgSO4
vs nimodipine
Monocenter,
historical control
group
60Start within
2 d, for up
to 14 d
Bolus: 20 mmol/20 min, cont.: 80 mmol/d
dose-adapted, target [Mg2+]: 2× baseline,
b2.5 mmol/L
Nimodipine,
valproate
No serious side effects
Prevedello
et al [13]
72Start at time
of admission
until discharge
from ICU
Bolus: 20 mmol/20 min, cont.: 100 mmol/d
dose-adapted, target [Mg2+]: 2× baseline,
b2.6 mmol/L
BP, ECG,
[Mg2+]
CVS: TCD, clinical
Nimodipine No serious side effects
Schmid-
Elsaesser
et al [14]
130 Start within
4 d, for 7 d PO,
then 7 d IV
Bolus: 0.4 mmol/kg body weight in
30 min, cont.: 1.2 mmol/kg body
weight per day
BP, ECG,
[Mg2+]
CVS: TCD, angio,
clinical, new
hypodensity in CT,
GOS after 1 y
CVS: clinical and
new hypodensity in
CT or angio, TCD
and CT perfusion,
GOS, mRS, BI, and
SF-36 after 3 mo
No difference –No serious side
effects
Stippler et al
[16]
76Start after
verification
of SAH, for
12 d
Bolus: none, cont.: 100 mmol/d ICU setting
(not described in
detail) [Mg2+]
Significantly less CVS,
trend to better outcome
based on mRS
Nimodipine No serious side effects
Appl. indicates application; cont., continuous infusion; [Mg2+], serum Mg2+level; [Ca2+], serum Ca2+level; BP, blood pressure; CVS, cerebral vasospasm; DCI, delayed cerebral ischemia; angio, angiographic; H&H, Hunt and Hess; mRS,
modified Rankin scale, BI, Barthel index; SF-36, 36-item short-form health survey subscale; ICU, intensive care unit, IV, intravenous; PO, per oral.
34
C. Muroi et al. / Surgical Neurology 69 (2008) 33–39

Page 3

placebo-controlled study were (1) to substantiate the
efficacy of MgSO4 in preventing the occurrence of
DIND and secondary ischemic infarction; (2) to evaluate
the impact on clinical outcome; and (3) to assess the
safety and side effects of MgSO4 infusion at a given
dosage and concomitant with other medication.
2. Methods
2.1. Patient population
The prospective, randomized, single-blind, and placebo-
controlled pilot study, approved by the ethics committee of
the University of Zurich, was carried out from December
2001 until November 2004 at the Department of Neurosur-
gery, University Hospital Zurich (Zurich, Switzerland).
Patients were included if they were older than 17 years, if
they were presenting within 3 days after SAH, and if
informed consent was available. The diagnosis of aneur-
ysmal SAH was based on CT and angiography studies. The
severity of SAH was assessed by the WFNS scale and the
Fisher scale. Exclusion criteria were pregnancy, history of
allergy, renal and neuromuscular diseases, heart disease,
hypotension (systolic blood pressure b110 mm Hg, unre-
sponsive to intravenous administration of fluids and/or
catecholamines in low dosage), or bradycardia (heart rate
b55 /min). All patients were treated in the neurocritical care
unit and were kept under close observation with continuous
invasive monitoring of the arterial blood pressure, ECG
monitoring, and arterial oxygen saturation until at least day
12 after ictus. Mg2+levels were measured every 12 hours,
and arterial blood gas analysis with ionized Ca2+levels was
measured every 4 hours. Patients at random were treated
with MgSO4(treatment group) or with the carrier substance
Ringer's lactate (placebo group).
2.2. Administration of MgSO4
At the day of admission, patients of the treatment
group received MgSO4 with a bolus of 16 mmol in a
150-mL solution of Ringer's lactate administered over
15 minutes, followed by a continuous, intravenous infusion
of 64 mmol/d. To maintain the serum Mg2+at the level of
twice the baseline, with a maximum of 2.0 mmol/L until day
12 after SAH, subsequent dosage adjustments of the MgSO4
infusion were made every 12 hours. Patients of the placebo
group were treated with a corresponding fluid amount of
Ringer's lactate. The administration of MgSO4 was
interrupted if bradycardia (heart rate b45 /min) occurred
and entirely stopped in case of hemodynamic relevance and
need of atropine. In case of hypotension (systolic blood
pressure b110 mm Hg), the application of MgSO4 was
interrupted and entirely stopped if the amount of nor-/
epinephrine had to be doubled or exceeded 100 μL/min and/
or a second catecholamine was needed to maintain the
blood pressure above 110 mm Hg. Further indications to
end the MgSO4therapy were atrioventricular conduction
disturbances, asystole (N2 seconds), respiratory failure
(suspected of being MgSO4related), oliguria, and/or severe
fluid and electrolyte imbalance.
2.3. Structured treatment
All patients were managed according to a standardized
treatment protocol for aneurysmal SAH and vasospasm [9].
The bleeding aneurysms were occluded with early surgery or
coiling within 3 days. Patients received prophylactic anti-
epileptic treatment with phenytoin or valproate and contin-
uous intravenous infusion of nimodipine 0.5 to 2 mg/h, if no
contraindication existed. Daily TCD measurements were
performed between day 4 and 12 by a blinded neuroradiol-
ogist. In patients with ultrasonographic evidence of vasos-
pasm and/or clinically suspected vasospasm, therapy with
hypertension, hypervolemia, and hemodilution (triple-H
therapy) was initiated [15]. Cerebral angiography with
balloon angioplasty and/or superselective papaverine instil-
lation was performed if triple-H therapy was not effective.
Computed tomographic scans were performed at least once
after aneurysm surgery and before discharge from the
neurocritical care unit.
2.4. Study end points
The incidence of TCD-detected vasospasm, DIND,
occurrence of infarction attributed to vasospasm, and
outcome after 3 months and 1 year were prospectively
analyzed. Vmeanof at least 150 cm/s of the MCA with a
Lindegaard index (ratio between the flow velocities of MCA
and distal segment of the extracranial internal carotid artery)
of more than 3 were regarded as TCD-detected vasospasm.
Vmeanof at least 200 cm/s with a Lindegaard index of more
than 3 were regarded as severe TCD-detected vasospasm [8].
New ischemic lesions in CTscans that could not be attributed
to other causes were considered as vasospasm-induced
infarctions. Delayed ischemic neurologic deficit was defined
as new focal neurologic deficits after exclusion of seizures,
hydrocephalus, electrolyte disturbance, or infection. The
presence of ischemic lesion in CT scans was not mandatory
for DIND. The outcome according to GOS was assessed by a
blinded neurologist after 3 months and 1 year.
2.5. Statistical analysis
Data are given as mean ± SD. Continuous data were
analyzed by the independent t test and nominal variables by
the Fisher exact test, as appropriate. Differences between
groups in severity of hemorrhage according the WFNS and
Fisher scale and outcome according GOS were analyzed by
the Mann-Whitney U test. A P value of less than .05 was
regarded as statistically significant.
3. Results
A total of 58 patients were randomized. Fifty-six patients
(97%) were treated with clipping and 2 patients with coiling.
35 C. Muroi et al. / Surgical Neurology 69 (2008) 33–39

Page 4

Twenty-seven patients were assigned to the placebo group
and 31 patients to the treatment group. Baseline character-
istics did not differ (Table 2). The intention-to-treat analysis
[11] revealed no difference between the baseline Mg2+and
Ca2+levels of the 2 groups, whereas the mean Mg2+levels
during treatment were significantly higher and the mean
Ca2+levels were significantly lower in the treatment group
than in the placebo group (Table 3). In the treatment group,
Mg2+levels during treatment were significantly higher and
Ca2+levels significantly lower than the baseline levels (both
P b .001) (Fig. 1A). Baseline Mg2+levels were 0.77 to
0.78 mmol/L (reference range, 0.70-1.10 mmol/L). The
average MgSO4dosage was 79 (±28) mmol/d (range 38-
174 mmol/d). The overall incidence of TCD-detected
vasospasm did not differ. If only severe TCD-detected
vasospasm was compared, the difference fell short of
significance: 5 (19%) in the placebo and only 1 (3%) in
the treatment group (P = .085). Vasospasm-induced infarc-
tions occurred less often in the treatment group although
without reaching statistical significance: 3 (10%) in the
treatment vs 6 (22%) in the placebo group. A trend toward
better outcome after 3 months was observed in the treatment
group (P = .083). Patients of the treatment group showed
significantly more episodes of hypotension: 3 (11%) in the
placebo vs 15 (48%) in the treatment group (P = .040).
Bradycardia occurred in 1 patient of the placebo and in 2
patients of the treatment group. Hypocalcemia (ionized Ca2+
level b1.0 mmol/L) occurred in 8 of the 31 patients (26%) of
the treatment group, whereas no patient in the placebo group
developed hypocalcemia (P = .005) (Table 3). In all patients
with hypocalcemia, the blood pH was within the reference range between 7.35 and 7.45. An inverse correlation between
serum Mg2+and Ca2+level was detected during treatment
(Spearman r = −0.623, P b.001) (Fig. 1B). An on-treatment
analysis [11] showed a trend toward less vasospasm-induced
infarctions in CT scans compared to the placebo group: 0
(0%) vs 6 (22%) (P = .073). Regarding the outcome, the on-
treatment analysis showed a statistically significant better
outcome after 3 months (P = .017) in patients treated with
MgSO4. A trend toward better outcome compared to the
placebo group was found after 1 year (P = .083).
In 16 patients (52%) of the treatment group, the MgSO4
infusion was discontinued between day 1 and 11 (mean,
5.6 days) (discontinued group): in 12, because of hypoten-
sion; in 2, because of arrhythmias; in 1, after an episode of
respiratory arrest; and in 1, because of myocardial infarction.
The mean Mg2+doses was higher compared with those of
patients who received MgSO4infusion until day 12 after
SAH (continued group) but without statistical significance:
84 (±36) mmol/d in the discontinued vs 74 (±13) mmol/d in
the continued group. The mean Mg2+level in the
discontinued group at time of termination was 1.40 (±0.25)
mmol/L, and during the treatment period, the Mg2+level was
even lower in the discontinued group compared to the
continued group: 1.39 (±0.37) vs 1.48 (±0.19) mmol/L.
Analysis of baseline characteristics and concomitant medi-
cation, that is, phenytoin (N300 mg/d) and nimodipine
Table 2
Baseline characteristics
Placebo
(n = 27)
Treatment
(n = 31)
Sex (female-male)
Mean age (±SD, y)
History of hypertension (n [%])
Concomitant medication (n [%])
Phenytoine
Nimodipine
WFNS grade (n [%])
I
II
III
IV
V
Fisher grade (n [%])
II
III
IV
Mean Mg2+baseline level (±SD, mmol/L) 0.78 (±0.07) NS 0.77 (±0.09)
Mean Ca2+baseline level (±SD, mmol/L) 1.15 (±0.09) NS 1.16 (±0.05)
MgSO4dose
Mean (±SD, mmol/d)
Range (mmol/d)
8:19
52.1 (±11.3) NS 53.6 (±14.4)
13 (48)NS
NS 7:24
9 (29)
25 (81)
21 (78)
NS
NS
28 (90)
18 (58)
9 (34)
3 (11)
3 (11)
6 (22)
6 (22)
NS 9 (29)
4 (13)
1 (3)
8 (26)
9 (29)
3 (11)
11 (41)
13 (48)
NS 3 (10)
13 (42)
15 (48)
0 (±0) 79 (±28)
38-1740
NS indicates statistically not significant.
Table 3
Treatment and outcome
Placebo
(n = 27)
Treatment
(n = 31)
Mean Mg2+level during treatment
(±SD, mmol/L)
Mean Ca2+level during treatment
(±SD, mmol/L)
Vasospasm (n [%])
TCD VmeanMCA N150 cm/s
TCD VmeanMCA N200 cm/s
New ischemia in CT
DIND
Adverse effects (n [%])
Hypotensiona
Bradycardiaa
Respiratorya
Hypocalcemiaa
Outcome after 3 mo (n [%])
Death
Vegetative
Severely disabled
Moderately disabled
Good recovery
Outcome after 1 y (n [%])
Death
Vegetative
Severely disabled
Moderately disabled
Good recovery
0.85 (±0.09)P b.0011.45 (±0.26)
1.15 (±0.08)P b.0011.06 (±0.06)
10 (37)
5 (19)
6 (22)
4 (15)
NS
P = .085
NS
NS
12 (39)
1 (3)
3 (10)
4 (13)
3 (11)
1 (4)
0 (0)
0 (0)
P = .040
NS
NS
P = .005
15 (48)
2 (6)
1 (3)
8 (26)
4 (15)
2 (7)
8 (30)
7 (26)
6 (22)
P = .0834 (13)
0 (0)
7 (23)
5 (16)
15 (48)
6 (22)
0 (0)
4 (15)
5 (19)
12 (44)
NS4 (13)
0 (0)
5 (16)
4 (13)
18 (58)
aSee text for definitions.
36 C. Muroi et al. / Surgical Neurology 69 (2008) 33–39

Page 5

(N1 g/d), did not show any significant differences compared
to the continued group (Table 4).
4. Discussion
In the present study, the intention-to-treat analysis showed
a trend toward fewer ultrasonographic evidence of severe
vasospasm and significantly better outcome after 3 months,
although the occurrence of hypotension and hypocalcemia
was significantly higher. The on-treatment analysis showed a
trend toward fewer CT-detected ischemia compared with the
analysis for the placebo group. There was a statistically
significant better outcome after 3 months and a trend toward
better outcome after 1 year. Because it was not possible to
predict if the MgSO4application would have to be stopped
before day 12, this is a postrandomization analysis and
therefore contains a certain risk of bias.
In contrast to all other studies, we were faced with
cardiovascular side effects. Based on our criteria for
termination of MgSO4 infusion, the therapy had to be
discontinued in 16 (52%) patients. The MgSO4administra-
tion scheme and target serum Mg2+level in our study did not
differ from previous studies. Postrandomization analysis of
concomitant medication and baseline characteristics showed
no differences, resulting in no predictive factor leading to
termination (Table 4). Although the average MgSO4dose
in patients of the discontinued group was higher than those
of the continued group, the mean Mg2+level was lower in
Fig. 1. A: Mg2+and Ca2+levels of patients of the treatment group shown as box plots. During treatment, mean Mg2+levels were significantly higher and Ca2+
levels were significantly lower compared to the baseline (both P b.001). B: Inverse correlation of Mg2+and Ca2+levels (Spearman r = −0.623, P b.001) shown
as a scattergram.
37C. Muroi et al. / Surgical Neurology 69 (2008) 33–39

Page 6

the discontinued group. The high SD and wider range
might indicate that the dose optimization was more
difficult in these patients. However, the broad range of
MgSO4doses needed to reach the target values indicates
that monitoring of serum Mg2+level is needed for appro-
priate dose optimization.
The fact that the randomization was not double-blind
carries certain risk of bias, although it can be justified
because (1) we intended to perform a dose-adapted study and
(2) assessment of side effects and complications was of
interest. We monitored the flow velocities of the MCA by
TCD. Although TCD is a sensitive tool for screening, the test
is nonspecific for vasospasm. However, in clinical routine, it
is reasonable to use a test with high sensitivity and poor
specificity. Furthermore, TCD remains the only noninvasive
and well-established technique to assess early signs of
vasospasm, and further diagnostic steps and therapy often
depend on its findings. Therefore, we assumed that the
assessment of TCD values in Mg2+therapy is of interest,
although ultrasonographic evidence of vasospasm does not
always lead to DIND. However, in this study, the question of
whether MgSO4reduces the incidence of vasospasm or not
cannot be answered.
Baseline Mg2+levels were at the lower end of the
reference range. This observation supports the findings that
serum Mg2+levels are lower early after SAH, as previously
described [18]. The lower Ca2+level in the intention-to-treat
group in our study is of interest. Twenty-six percent of the
patients treated with MgSO4developed hypocalcemia. An
inverse correlation between serum Mg2+and Ca2+levels was
found during treatment.
To date, 11 pilot studies have been reported to assess the
effectiveness and safety of MgSO4therapy in patients with
SAH (Table 1) [4,5,13,14,17-21,23,24]. The most promising
study was reported by van den Bergh et al [18]. Their
randomized, double-blind, placebo-controlled multicenter
study included a total of 283 patients. Magnesium treatment
reduced the risk of developing new infarctions, detected in
CT scans, by 34%. After 3 months, the reduction of risking
poor outcome (defined as modified Rankin score N3) was
23%. However, the results have to be interpreted with
caution because in this 2 × 2 factorial trial (magnesium and
acetylsalicylic acid in subarachnoid hemorrhage trial), data
on patients receiving aspirin were not disclosed. The dose
regimen was based on a dose finding study [17]. In the
randomized study, monitoring of Mg2+level was not
obligatory, and therefore, no data were available. Seventeen
patients allocated to magnesium treatment discontinued
intervention for the following reasons: no aneurysmal SAH
(4); hypotension (1); phlebitis (2); bradycardia and atrium
fibrillation (1); routine magnesium suppletion (1); hyper-
magnesemia (3); renal failure (1); imminent death (1);
request of patient or family (1); and unreliable registration of
treatment administration (2). All 11 pilot studies have the
following in common: (1) MgSO4application is reported to
be safe and feasible and that no serious side effects were
observed; (2) no strict criteria for termination of MgSO4
infusion were defined prospectively.
Of the above-mentioned studies, only one, performed by
Veyna et al [21], described the phenomenon of hypocalcemia
in which Ca2+was substituted. In some case reports,
hypocalcemia induced by therapeutic hypermagnesemia
and the possible role of parathyroid hormone have been
mentioned [7,10,12]. The question arises whether low Ca2+
levels in patients with SAH should be treated or not. In the
present series, no patient with hypocalcemia showed obvious
symptoms of increased neuromuscular irritability with
characteristic Chvostek and Trousseau signs. However, in
association with hypoparathyroidism, the literature mentions
further infrequent presentations with extrapyramidal chor-
eoathetoid syndromes, parkinsonism, dystonia, hemiballis-
mus, oculogyric crisis, and epileptic seizures [2,3]. Mild
mental status changes such as irritability, perioral numbness,
paresthesias, and muscle cramps are already described as
being related to hypocalcemia [1]. These symptoms might be
confused with secondary neurologic deficits induced by
vasospasm. Ca2+is essential for homeostasis and the
function of multiple organ systems, so Ca2+levels have to
be closely monitored during MgSO4infusion. Particularly, in
patients with SAH, the importance of sustaining hemody-
namic, temperature, and metabolic homeostasis is well
recognized [22].
5. Conclusions
High-dose MgSO4 therapy might be effective as a
prophylactic adjacent therapy in patients with SAH to
Table 4
Baseline characteristics of the continued and discontinued group
Continued
(n = 15)
Discontinued
(n = 16)
Sex (female-male)
Mean age (±SD, y)
History of hypertension (n [%])
Concomitant medication (n [%])
Phenytoine
Nimodipine
SAH grade (n [%])
WFNS grade
I
II
III
IV
V
Fisher grade
II
III
IV
Mean Mg2+baseline level
(±SD, mmol/L)
Mean Ca2+baseline level
(±SD, mmol/L)
3:12
50.9 (±18.3)
4 (27)
NS
NS
NS
4:12
56.1 (±9.3)
5 (31)
13 (87)
9 (60)
NS
NS
15 (94)
9 (56)
4 (27)
2 (13)
0 (0)
4 (27)
5 (33)
NS 5 (31)
2 (13)
1 (6)
4 (25)
4 (25)
2 (14)
5 (33)
8 (53)
0.76 (±0.08)
NS1 (6)
8 (50)
7 (44)
0.77 (±0.10) NS
1.16 (±0.05)NS1.16 (±0.05)
n: number of patients, SD: standard deviation, WFNS: world federation of
neurologic surgeons, NS: statistically not significant.
38C. Muroi et al. / Surgical Neurology 69 (2008) 33–39

Page 7

reduce the risk of poor outcome. High-dose MgSO4, at least
in combination with nimodipine and phenytoin, may be
associated with relevant cardiovascular side effects. There-
fore, patients should be carefully observed in an intensive or
intermediate care setting with closed meshed arterial blood
pressure and ECG monitoring as well as frequent measure-
ments of Ca2+and Mg2+levels.
References
[1] Ariyan CE, Sosa JA. Assessment and management of patients with
abnormal calcium. Crit Care Med 2004;32(Suppl 4):S146-54.
[2] Armelisasso C, Vaccario ML, Pontecorvi A, Mazza S. Tonic-clonic
seizures in a patient with primary hypoparathyroidism: a case report.
Clin EEG Neurosci 2004;35:97-9.
[3] Bindu M, Harinarayana CV. A rare treatable cause of epilepsy—report
of two cases. Eur J Neurol 2006;13:786-8.
[4] Boet R, Mee E. Magnesium sulfate in the management of patients with
Fisher grade 3 subarachnoid hemorrhage: a pilot study. Neurosurgery
2000;47:602-6.
[5] Chia RY, Hughes RS, Morgan MK. Magnesium: a useful adjunct in the
prevention of cerebral vasospasm following aneurysmal subarachnoid
haemorrhage. J Clin Neurosci 2002;9:279-81.
[6] Dorsch NW. Therapeutic approaches to vasospasm in subarachnoid
hemorrhage. Curr Opin Crit Care 2002;8:128-33.
[7] Eisenbud E, LoBue CC. Hypocalcemia after therapeutic use of
magnesium sulfate. Arch Intern Med 1976;136:688-91.
[8] Harrod CG, Bendok BR, Batjer HH. Prediction of cerebral vasospasm
in patients presenting with aneurysmal subarachnoid hemorrhage: a
review. Neurosurgery 2005;56:633-54.
[9] Keller E, Krayenbuhl N, Bjeljac M, Yonekawa Y. Cerebral vasospasm:
results of a structured multimodal treatment. Acta Neurochir (Wien)
2005;94(suppl):65-73.
[10] Koontz SL, Friedman SA, Schwartz ML. Symptomatic hypocalcemia
after tocolytic therapy with magnesium sulfate and nifedipine. Am J
Obstet Gynecol 2004;90:1773-6.
[11] Lewis SC, Warlow CP. How to spot bias and other potential problems
in randomised controlled trials. J Neurol Neurosurg Psychiatry 2004;
75:181-7.
[12] Mayan H, Hourvitz A, Schiff E, Zvi F. Symptomatic hypocalcaemia in
hypermagnesaemia-induced hypoparathyroidism, during magnesium
tocolytic therapy—possible involvement of the calcium-sensing
receptor. Nephrol Dial Transplant 1999;14:1764-6.
[13] Prevedello DM-S, Cordeiro JG, de Morais AL, Saucedo NS, Chen IB,
Araujo JC. Magnesium sulfate: role as possible attenuating factor in
vasospasm morbidity. Surg Neurol 2006;65(Suppl 1):S1:14-21.
[14] Schmid-Elsaesser R, Kunz M, Zausinger S, Prueckner S, Briegel J,
Steiger HJ. Intravenous magnesium versus nimodipine in the treatment
of patients with aneurysmal subarachnoid hemorrhage: a randomized
study. Neurosurgery 2006;58:1054-65.
[15] Sen J, Belli A, Albon H, Morgan L, Petzold A, Kitchen N. Triple-H
therapy in the management of aneurysmal subarachnoid haemorrhage.
Lancet Neurology 2003;2:614-21.
[16] Stippler M, Crago E, Levy EI, Kerr ME, Yonas H, Horowitz MB,
Kassam A. Magnesium infusion for vasospasm prophylaxis after
subarachnoid hemorrhage. J Neurosurg 2006;105:723-9.
[17] van den Bergh WM, Albrecht KW, Berkelbach van der Sprenkel JW,
Rinkel GJ. Magnesium therapy after aneurysmal subarachnoid
hemorrhage a dose-finding study for long term treatment. Acta
Neurochir (Wien) 2003;145:195-9.
[18] van den Bergh WM, Algra A, van Kooten F, Dirven CM, van Gijn J,
Vermeulen M, Rinkel GJ. Magnesium sulfate in aneurysmal subar-
achnoid hemorrhage: a randomized controlled trial. Stroke 2005;36:
1011-105.
[19] van den Bergh WM, Algra A, van der Sprenkel JW, Tulleken CA,
Rinkel GJ. Hypomagnesemia after aneurysmal subarachnoid hemor-
rhage. Neurosurgery 2003;52:276-81.
[20] van Norden AGW, van der Bergh WM, Rinkel GJE. Dose evaluation
for long-term magnesium treatment in aneurysmal subarachnoid
haemorrhage. J Clin Pharm Ther 2005;30:439-42.
[21] Veyna RS, Seyfried D, Burke DG, Zimmerman C, Mlynarek M,
Nichols V, Marrocco A, Thomas AJ, Mitsias PD, Malik GM.
Magnesium sulfate therapy after aneurysmal subarachnoid hemor-
rhage. J Neurosurg 2002;96:510-4.
[22] Wartenberg KE, Mayer SA. Medical complications after subarachnoid
hemorrhage:newstrategiesfor preventionandmanagement.Curr Opin
Crit Care 2006;12:78-84.
[23] Wong GKC, Chan MTV, Boet R, Poon WS, Gin T. Intra-
venous magnesium sulfate after subarachnoid hemorrhage: a prospec-
tive randomized pilot study. J Neurosurg Anesthesiol 2006;18:142-8.
[24] Yahia AM, Kirmani JF, Qureshi AI, Guterman LR, Hopkins LN. The
safety and feasibility of continuous intravenous magnesium sulfate for
prevention of cerebral vasospasm in subarachnoid hemorrhage.
Neurocrit Care 2005;3:16-23.
Commentary
This study adds to at least 11 other studies of magnesium
treatment for SAH. The significant effects ofmagnesium were
to lower blood pressure and serum calcium and to improve
outcome. The improvement in outcome is remarkable in view
oftheverylimitednumberofpatientsstudiedandtheincreased
incidence of hypotension in the treated patients. In larger
studies, such as that of van den Bergh et al [1], no significant
effect of magnesium infusion on outcome was demonstrated
after randomizing over 4 times as many patients as in the
Zurich study. The limitations of the larger study are cited
though, and include lack of monitoring of serum magnesium
levels and inclusion of aspirin treatment in some patients. The
group in Zurich includes highly experienced neurointensivists
aswellasphysicianscapableoftreatingrupturedaneurysms.It
is thus disappointing to see that half the patients treated with
magnesium had to have it discontinued because of perceived
side effects. Because the study was not blinded, the question
arises as to whether the treating physicians were more
concerned about hypotension and arrhythmias in these
patients. A blinded study is ongoing now and will help
determine the efficacy and safety of therapeutic magnesium
infusions for vasospasm and SAH. Magnesium levels are not
being measured in the randomized trials, which may be a
problemifthestudydoesnotshowtheefficacyofmagnesium.
R. Loch Macdonald, MD, PhD
Department of Neurosurgery
University of Toronto
Toronto, Ontario, Canada M5C 3G7
Reference
[1] van den Bergh WM, Algra A, van Kooten F, Dirven CM, van Gijn J,
Vermeulen M, Rinkel GJ. Magnesium sulfate in aneurysmal subar-
achnoid hemorrhage: a randomized controlled trial. Stroke 2005;36:
1011-5.
39C. Muroi et al. / Surgical Neurology 69 (2008) 33–39