Research

Effect of magnesium treatment and glucose levels on delayed cerebral ischemia in

patients with subarachnoid hemorrhage: a substudy of the Magnesium in
Aneurysmal Subarachnoid Haemorrhage trial (MASH-II)
Jolien F. Leijenaar1, Sanne M. Dorhout Mees1*, Ale Algra1,2, Walter M. van den Bergh3, and
Gabriel J. E. Rinkel1 on behalf of the MASH-II Study Group
Background Magnesium treatment did not improve outcome
in patients with aneurysmal subarachnoid haemorrhage in the
Magnesium in Aneurysmal Subarachnoid Haemorrhage II trial.
We hypothesized that high glucose levels may have offset a
potential beneficial effect to prevent delayed cerebral ischemia. We investigated if magnesium treatment led to less
delayed cerebral ischemia and if glucose levels interacted with
magnesium treatment in the Magnesium in Aneurysmal Subarachnoid Haemorrhage II trial.
Aim To investigate the effect of magnesium treatment on
occurrence of delayed cerebral ischemia and the interaction
between glucose levels and magnesium treatment in subarachnoid hemorrhage patients.
Methods The Magnesium in Aneurysmal Subarachnoid Haemorrhage was a phase III randomized placebo-controlled trial
assessing the effect of magnesium sulphate on clinical
outcome in aneurysmal subarachnoid hemorrhage patients.
For the current study, we included only the patients admitted
to the University Medical Centre-Utrecht. We calculated hazard
ratios for occurrence of delayed cerebral ischemia in patients
treated with magnesium vs. placebo for the entire study population, and separately in the subgroups of patients with high
and low mean fasting and mean daily glucose levels until
onset of delayed cerebral ischemia. We used the cross-product
of magnesium and glucose in the regression analysis to evaluate whether an interaction between magnesium and glucose
existed.
Results We included 616 patients: 307 received magnesium
and 309 placebo; 156 patients had delayed cerebral ischemia.
Hazard ratio for magnesium on occurrence of delayed cerebral
ischemia was 10 (95% confidence interval: 0714). Results

Correspondence: Sanne M. Dorhout Mees*, Department of Neurology

and Neurosurgery, Rudolf Magnus Institute of Neuroscience, University
Medical Centre Utrecht, Heidelberglaan 100 3584 CX, Utrecht 3508,
The Netherlands.
E-mail: s.m.dorhoutmees@umcutrecht.nl
1
Department of Neurology and Neurosurgery, Rudolf Magnus Institute of
Neuroscience, University Medical Centre Utrecht, Utrecht, The
Netherlands
2
Julius Centre for Health Sciences and Primary Care, University Medical
Centre Utrecht, Utrecht, The Netherlands
3
Department of Critical Care, University Medical Centre Groningen, University of Groningen, Groningen, The Netherlands
Received: 20 February 2015; Accepted: 30 June 2015
Conflict of interest: None declared.
Trial registration: MASH-II is registered with controlled-trials.com
(ISRCTN 68742385) and the EU Clinical Trials Register (EudraCT 2006003523-36)
Funding: The MASH-II study was funded by the Netherlands Heart Foundation (grant number 2005BO16).
DOI: 10.1111/ijs.12621

108

Vol 10, October 2015, 108112

were similar in patients with low or high fasting or daily

glucose levels. We found no interactions between magnesium
treatment and high fasting (P = 054) and daily glucose
(P = 060).
Conclusions Magnesium treatment did not reduce the risk of
delayed cerebral ischemia in patients with aneurysmal subarachnoid hemorrhage, nor was there an interaction with
glucose levels. It is therefore unlikely that glucose levels
explain the failure of magnesium to prevent delayed cerebral
ischemia and poor outcome after aneurysmal subarachnoid
hemorrhage.
Key words: DCI, delayed cerebral ischemia, glucose, magnesium, SAH,
subarachnoid hemorrhage

Introduction
Delayed cerebral ischemia (DCI) is an important complication
after aneurysmal subarachnoid hemorrhage (aSAH). In the phase
II Magnesium in Aneurysmal Subarachnoid Haemorrhage
(MASH)-I study and Cochrane meta-analysis of 2007, a trend
toward less DCI and poor outcome was observed in aSAH
patients treated with magnesium vs. placebo (1,2). However, in
the subsequent phase III study (MASH-II) and the updated metaanalysis in 2012, magnesium treatment had no effect on clinical
outcome (3). Data on DCI were not analyzed, leaving the question
unanswered whether magnesium reduces the risk of DCI.
The relation between magnesium treatment and DCI might be
more complex than previously thought. In the IMAGES trial, a
randomized controlled trial that studied the effect of magnesium
treatment on outcome after ischemic stroke, a potential interaction was found between high glucose levels and magnesium treatment, leading to more infarct growth in hyperglycemic patients
when treated with magnesium compared with hyperglycemic
patients treated with placebo (4). In aSAH patients, high blood
glucose levels are also a risk factor for the development of DCI
(57). Thus, high blood glucose levels might lead to more DCI in
magnesium-treated patients in MASH-II, thereby offsetting a
possible beneficial effect of magnesium therapy on the occurrence
of DCI.
First, we studied whether the occurrence of DCI differed
between the magnesium and placebo group in MASH-II patients
and second, whether magnesium effects on DCI differed between
patients with low or high glucose levels.

Methods
Patient population
We analyzed data of patients who participated in the MASH-II
trial, a phase III multicenter randomized placebo-controlled trial,
2015 World Stroke Organization

Research

J. F. Leijenaar et al.
and who were admitted to the University Medical Centre Utrecht
(UMCU). Of the 1204 patients included in MASH-II, 630 were
included in the UMCU. For the current study, we excluded 13
patients because of a diagnosis other than aSAH and one patient
because of absence of glucose levels as this patient was transferred
immediately to another hospital. Inclusion criteria for MASH-II
were: admission within four-days of the hemorrhage, aSAH confirmed by the presence of extravasated blood in the basal cisterns
by brain computed tomography (CT), or if CT was negative by
xanthochromia of cerebrospinal fluid in combination with a cerebral aneurysm. Exclusion criteria were age younger than 18 years,
renal failure (defined as serum creatinine concentration of more
than 150 mol/l), a body weight less than 50 kg, or imminent
death (3).
We recorded age, gender, and clinical condition on admission
according to the World Federation of Neurosurgical Surgeons
(WFNS) grading scale. The WFNS scale was dichotomized into
good (WFNS 13) and poor (WFNS 45) clinical condition (8).
Procedures
Data on glucose were collected for the period in which study
medication was administered, i.e., until 20 days after hemorrhage
onset, or until hospital discharge or death if it occurred within 20
days of onset of aSAH. In the original MASH-II trial, the measurement of glucose was not included, but as part of daily practice
glucose measurements were done at least once every day between
6:00 and 8:00 am (fasting glucose) during medium or intensive
care admission, duration depending on the duration of the hospital stay. Most patients however had several glucose measurements per day. In this study, we recorded for each patient a fasting
glucose (measured between 6:00 and 8:00 am) and a mean glucose
level (mean of all available glucose levels) per day as more than
one glucose level per day was often available.
We then calculated for each patient a mean of all available
fasting glucose levels until the occurrence of DCI. For patients
without DCI, the mean fasting glucose values were calculated
until the median day of DCI onset in the patients who did develop
DCI, in order to make a comparison. The same approach was used
with the mean daily glucose levels.

fasting glucose levels separately, and for patients with high and
low daily glucose levels separately. HRs were also calculated for
the occurrence of cerebral infarction for both the mean fasting
and the mean daily glucose independently. To study whether an
interaction existed between magnesium treatment and glucose,
the cross-product of magnesium and either mean fasting glucose
or mean daily glucose was entered into the Cox regression analyses. If the P value of the cross-product was <005, we considered a
statistically significant interaction between treatment modality
and glucose to be present. Because of the explanatory research
questions, all analyses were performed as on-treatment analyses
and adjusted for clinical condition on admission, age, and gender.
During the MASH-II trial, glucose levels were regulated strictly
in the intensive care unit (ICU) unit with insulin treatment. To see
whether insulin influenced the occurrence of DCI, we also performed on-treatment analyses adjusted for insulin.

Results
Of the 616 included patients, 307 (498%) received magnesium
treatment and 309 (502%) placebo (Fig. 1). Baseline characteristics were comparable between the magnesium and placebo groups
(Table 1).
The adjusted HR for the risk of DCI in patients treated with
magnesium vs. those treated with placebo was 10 (95% CI 07
14). Adjusted HRs for occurrence of DCI for magnesium vs.
placebo in patients with high mean fasting (HR 11, 95% CI
0717) vs. low mean fasting (HR 10, 95% CI 0616) and high
daily glucose level (HR 10, 95% CI 0716) vs. low daily glucose
level (HR 09, 95% CI 0515) were similar (Table 2). There were
no significant interactions between magnesium and mean fasting

Outcomes
The outcome of the study was DCI, either clinically (defined as
the occurrence of focal neurological impairment, or a decrease of
at least 2 points on the Glasgow Coma Scale, present for at least
one-hour, not apparent immediately after aneurysm occlusion,
and not explained by other causes), or radiological [defined as the
presence of new cerebral infarction on repeated CT or magnetic
resonance (MR) scan, which could not be attributed to other
causes such as surgical clipping or endovascular treatment] (9).
Data-analysis
First, we calculated a hazard ratio (HR) with 95% confidence
interval (95% CI) with Cox regression analysis for the occurrence
of DCI for magnesium treatment compared with placebo.
Second, we dichotomized mean fasting glucose and mean daily
glucose levels at their median values. We then again calculated
HRs for the occurrence of DCI for magnesium treatment compared with placebo, but now for patients with high and low
2015 World Stroke Organization

Fig. 1 Flowchart of patient inclusion. aSAH, aneurysmal subarachnoid

haemorrhage; MASH-II, Magnesium for Aneurysmal Subarachnoid Haemorrhage, UMC, University Medical Centre.

Vol 10, October 2015, 108112

109

Research

J. F. Leijenaar et al.

Table 1 Baseline and disease characteristics

Mean age (SD; years)

Female (%)
World Federation of Neurological Surgeons subarachnoid hemorrhage grade 4 (%)
Median time to receive study medication from symptom onset (IQR; h)
Aneurysm treatment (%)
Coiling
Clipping
None
Mean fasting glucose from study medication until DCI (SD; mmol/l)
Mean daily glucose until DCI (SD; mmol/l)
DCI (%)

Magnesium (n = 307)

Placebo (n = 309)

562 (135)
219 (713)
71 (231)
26 (1954)

579 (131)
222 (718)
72 (233)
30 (2053)

153 (498)
118 (384)
36 (117)
69 (13)
70 (13)
81 (264)

144 (466)
124 (401)
41 (133)
67 (12)
68 (12)
75 (243)

DCI, delayed cerebral ischemia; IQR, interquartile range; SD, standard deviation.

Table 2 HRs for DCI for magnesium compared with placebo in patients with high fasting/daily glucose and low daily glucose

High fasting glucose

Placebo
Magnesium
Low fasting glucose
Placebo
Magnesium
Interaction fasting glucose and magnesium
High daily glucose
Placebo
Magnesium
Low daily glucose
Placebo
Magnesium
Interaction daily glucose and magnesium

n/N (%)

Crude HR (95% CI)

Adjusted HR (95% CI)*

41/147 (279)
50/165 (303)

Ref.
11 (0717)

Ref.
11 (0717)

34/162 (210)
31/142 (218)

Ref.
09 (0514)
P = 048

Ref.
10 (0616)
P = 054

43/147 (293)
52/161 (323)

Ref.
10 (0716)

Ref.
10 (0716)

32/162 (198)
29/146 (199)

Ref.
09 (0514)
P = 062

Ref.
09 (0515)
P = 060

*Adjusted for age, gender, and World Federation of Neurological Surgeons. CI, confidence interval; DCI, delayed cerebral ischemia; HR, hazard ratio.

glucose (P = 054) or magnesium and mean daily glucose

(P = 060) for the occurrence of DCI after adjustment for baseline
characteristics.
Similar risk estimates were found for the relationships between
fasting glucose and daily glucose with regard to occurrence of
cerebral infarction {adjusted HRs for high mean fasting [HR 09,
95% CI 0615] vs. low mean fasting [HR 09, 95% CI 0515]
and high daily glucose level 09 [0514] vs. low daily glucose level
[08 (0514)]}. There were no statistically significant interactions between magnesium and mean fasting glucose (P = 066) or
magnesium and mean daily glucose (P = 085) for the occurrence
of cerebral infarction after adjustment for baseline characteristics.
In the UMCU patient cohort, 24/616 (39%) were treated with
insulin. Additional adjustment for insulin use did not influence
the HR estimates importantly (data not shown).

Discussion
Our study shows that magnesium treatment did not change the
risk of DCI after aSAH, and that the absence of an effect was not
explained by interaction with glucose levels.

110

Vol 10, October 2015, 108112

The lack of effect of magnesium treatment on occurrence of

DCI can explain the neutral effect of the MASH-II study on
clinical outcome. In contrast with the IMAGES substudy, we
found no interaction between magnesium therapy and high
glucose levels, neither in the analyses with DCI nor in those with
cerebral infarction as outcome measure. There are a couple of
differences between these two trials that could explain the absence
of interaction in our study. First, high glucose levels could have
different effects on ischemic brain tissue in patients with primary
brain infarcts than in patients with secondary brain infarcts after
a subarachnoid hemorrhage. However, in both diseases a clear
relationship between glucose levels and ischemic events has been
shown. Second, the IMAGES trial used the admission glucose for
their analyses (4,10). In our study, we used mean fasting and mean
daily glucose during the study treatment, because study treatment
was not started yet when admission glucose was measured,
whereas in IMAGES start of study treatment was much earlier
after the ictus. For this reason, we thought that an interaction
between admission glucose and study treatment was less relevant
in MASH-II. However, we did perform the same analysis using
admission glucose levels only, and also this analysis found no
interaction (P = 049). A third explanation is that the finding of
2015 World Stroke Organization

Research

J. F. Leijenaar et al.
the interaction in the post hoc study of IMAGES was a chance
finding.
Our study has a couple of strengths: our cohort was much
larger than the sub-analysis cohort of IMAGES wherein the relationship between glucose and infarct growth was investigated
(616 vs. 90 patients). Patients were randomized to receive either
magnesium or placebo therapy, and although we had to use a
retrospective design, the data used in our study were collected
prospectively at that time according to our local aSAH protocol. A
limitation of the study might be that only the UMCU patients of
the MASH-II trial were included, but we have no reason to think
results would be different in patients of the other centers, as
baseline and outcome characteristics were similar in the UMCU
cohort compared with the whole MASH-II cohort. We did not
exclude patients known with diabetes, because diabetes is not a
risk factor for aSAH, in contrast to patients with ischemic stroke,
so we do not expect a high percentage of patients with diabetes
(11). Additional adjustment for insulin use did not influence the
HRs for the occurrence of DCI; therefore, it is unlikely that any
interaction between blood glucose levels and magnesium was
masked by insulin use. We do not have information on how many
patients were on continuous gastric tube feeds.
Several trials (1214) indicated that combined treatment with
mild hypothermia and magnesium had a synergistic neuroprotective effect. In our study population, hypothermia was not
employed, thus we could not assess a potential synergistic effect of
hypothermia and magnesium treatment in our study. Hyperthermia and fever were treated with medication and cooling blankets
according to the institutional protocols.
In conclusion, magnesium treatment did not influence the
occurrence of DCI in MASH-II. Glucose levels did not influence
the relation between magnesium treatment and occurrence of
DCI, and are unlikely to explain the failure of magnesium to
prevent DCI.

Contributions
Jolien F. Leijenaar: Acquisition of data, drafting/revising manuscript for content including medical writing of content, analysis/
interpretation of data.
Sanne M. Dorhout Mees: Interpretation of data, critical revision
of the manuscript for important intellectual content.
Ale Algra: Analysis/interpretation of data, critical revision of
the manuscript for important intellectual content.
Walter M. van den Bergh: Critical revision of the manuscript for
important intellectual content.
Gabriel J. E. Rinkel: Critical revision of the manuscript for
important intellectual content.

MASH-2 study group

Executive committee: S. M. Dorhout Mees, W. M. van den Bergh,
G. J. E. Rinkel, A. Algra, M. van Buuren (University Medical
Centre Utrecht, Utrecht, The Netherlands). Steering committee: R.
Al-Shahi Salman (Western General Hospital, Edinburgh, UK), J.
Boiten, (Medical Centre Haaglanden, the Hague, The Nether 2015 World Stroke Organization

lands), F. van Kooten (Erasmus Medical Centre Rotterdam, Rotterdam, The Netherlands), H. Kuijsten (St Elisabeth Hospital,
Tilburg, The Netherlands), P. M. Lavados (Institute of Neurosurgery, Universidad de Chile and Clinica Alemana, Universidad del
Desarrollo, Santiago, Chile), R. J. van Oostenbrugge (Maastricht
University Medical Centre, Maastricht, The Netherlands), W. P.
Vandertop (Neurosurgical Centre Amsterdam, Academic Medical
Centre and VU University Medical Centre, Amsterdam, The
Netherlands). Data monitoring committee: J. G. van der Bom
(Chair; Leiden University Medical Centre, Leiden, The Netherlands), W. P. Th. M. Mali (University Medical Centre Utrecht,
Utrecht, The Netherlands), P. M. Rothwell, R. S. C. Kerr (University of Oxford and Oxford Radcliffe Hospitals, Oxford, UK). Participating centers (local investigators who included at least 10
patients): University Medical Centre Utrecht, Utrecht, The Netherlands (630 patients; S. M. Dorhout Mees, W. M. van den Bergh,
A. Algra, G. J. E. Rinkel, R. Kleinloog, J. W. Dankbaar, C. S. Gathier,
D. J. Nieuwkamp, M. J. A. Luitse, S. Achterberg); Neurosurgical
Centre Amsterdam, Amsterdam, The Netherlands (150 patients;
W. P. Vandertop); Erasmus Medical Centre Rotterdam, Rotterdam, The Netherlands (104 patients; F. van Kooten); St Elisabeth
hospital Tilburg, Tilburg, The Netherlands (91 patients; H. Kuijsten, G. Roks); Medical Centre Haaglanden, the Hague, The Netherlands (81 patients; J. Boiten, J. Kerklaan, J. P. W. Dennessen);
Maastricht University Medical Centre, Maastricht, The Netherlands (68 patients; R. J. van Oostenbrugge); Western General Hospital, Edinburgh, UK (59 patients; R. Al-Shahi Salman),; Institute
of Neurosurgery, Universidad de Chile, Santiago, Chile (21
patients; P. M. Lavados, V. V. Olavarria).

References
1 van den Bergh WM, Algra A, van Kooten F et al. Magnesium sulfate in
aneurysmal subarachnoid hemorrhage: a randomized controlled trial.
Stroke 2005; 36:10115.
2 Dorhout Mees SM, Rinkel GJ, Feigin VL et al. Calcium antagonists for
aneurysmal subarachnoid haemorrhage. Cochrane Database Syst Rev
2007; (3):CD000277.
3 Dorhout Mees SM, Algra A, Vandertop WP et al. Magnesium for
aneurysmal subarachnoid haemorrhage (MASH-2): a randomised
placebo-controlled trial. Lancet 2012; 380:449.
4 Kidwell CS, Lees KR, Muir KW et al. Results of the MRI substudy of
the intravenous magnesium efficacy in stroke trial. Stroke 2009;
40:17049.
5 Kruyt ND, Roos YW, Dorhout Mees SM et al. High mean fasting
glucose levels independently predict poor outcome and delayed cerebral ischaemia after aneurysmal subarachnoid haemorrhage. J Neurol
Neurosurg Psychiatry 2008; 79:13825.
6 Barletta JF, Figueroa BE, DeShane R, Blau SA, McAllen KJ. High
glucose variability increases cerebral infarction in patients with spontaneous subarachnoid hemorrhage. J Crit Care 2013; 28:798803.
7 Badjatia N, Topcuoglu MA, Buonanno FS et al. Relationship between
hyperglycemia and symptomatic vasospasm after subarachnoid hemorrhage. Crit Care Med 2005; 33:16039.
8 Report of World Federation of Neurological Surgeons Committee on
a Universal Subarachnoid Hemorrhage Grading Scale. J Neurosurg
1988; 68:9856.
9 Vergouwen MD, Vermeulen M, van Gijn J et al. Definition of delayed
cerebral ischemia after aneurysmal subarachnoid hemorrhage as an
outcome event in clinical trials and observational studies: proposal of
a multidisciplinary research group. Stroke 2010; 41:23915.
Vol 10, October 2015, 108112

111

Research
10 Muir KW, Lees KR, Ford I, Davis S, Intravenous Magnesium Efficacy
in Stroke (IMAGES) Study Investigators. Magnesium for acute stroke
(Intravenous Magnesium Efficacy in Stroke trial): randomised controlled trial. Lancet 2004; 363:43945.
11 Cui R, Iso H, Yamagishi K et al. Diabetes mellitus and risk of stroke
and its subtypes among Japanese: the Japan public health center study.
Stroke 2011; 42:26114.
12 Meloni BP, Campbell K, Zhu H, Knuckey NW. In search of clinical
neuroprotection after brain ischemia: the case for mild hypothermia
(35C) and magnesium. Stroke 2009; 40:223640.

112

Vol 10, October 2015, 108112

J. F. Leijenaar et al.
13 Campbell K, Meloni BP, Zhu H, Knuckey NW. Magnesium treatment
and spontaneous mild hypothermia after transient focal cerebral ischemia in the rat. Brain Res Bull 2008; 77:3202.
14 Campbell K, Meloni BP, Knuckey NW. Combined magnesium and
mild hypothermia (35 degrees C) treatment reduces infarct volumes
after permanent middle cerebral artery occlusion in the rat at 2 and 4,
but not 6 h. Brain Res 2008; 1230:25864.

2015 World Stroke Organization

Copyright of International Journal of Stroke is the property of Wiley-Blackwell and its

content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.