Magnesium in subarachnoid haemorrhage Essay

Subarachnoid haemorrhage (SAH) caused by a ruptured aneurysm accounts for only 5% of strokes, but occurs at a fairly young age and carries a worse prognosis. Delayed cerebral ischemia is an important cause of death and dependence after aneurysmal SAH. The current mainstay of preventing delayed cerebral ischemia is nimodipine and maintenance of normovolemia, but even with this strategy delayed cerebral ischemia occurs in a considerable proportion of patients. Magnesium is an inexpensive, easily available neuroprotective agent that reduces cerebral vasospasm and infarct volume after experimental SAH. In a meta-analysis of all randomized clinical trials, magnesium shows a tendency to reduce the occurrence of delayed cerebral ischemia and poor outcome after SAH, but the question if magnesium is advantageous in SAH patients is still in abeyance. Currently a large phase III trial aiming for 1200 patients is being conducted that will hopefully provide definite evidence whether magnesium treatment is beneficial in subarachnoid haemorrhage patients.


Subarachnoid haemorrhage
Delayed cerebral ischemia
Magnesium and cerebral ischemia: opportunities and concerns
Hypomagnesemia in subarachnoid haemorrhage
Magnesium treatment in subarachnoid haemorrhage

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now


Subarachnoid haemorrhage


Subarachnoid haemorrhage (SAH) caused by a ruptured aneurysm accounts for only 5% of strokes, but occurs at a fairly young age and carries a worse prognosis than other types of stroke (van Gijn et at., 2007). The cardinal feature is a history of unusually severe headache that started suddenly, but patients frequent deteriorate into unconsciousness shortly after onset. On admission two-thirds of all patients have depressed consciousness, of whom half are in coma (Brilstra et at., 2000). The school of Hippocrates already described the clinical picture of a SAH: “When persons in good health are suddenly seized with pains in the head, and straightway are laid down speechless, and breathe with stertor, they die in seven days, unless fever comes on” (Hippocrates et at., 1527; Clarke, 1963). Attila the Hun is suspected to have had a SAH as he died suddenly after severe headache developed during sexual intercourse (Babinski, 1893).

The incidence of SAH in most populations is 6-7 per 100000 person-years (Linn et at., 1996). About one in eight patients die before reaching the hospital. The in-hospital case fatality is about one-third (Nieuwkamp et at., 2009). Of patients who survive the SAH approximately one-third remain dependent. Because of the young age  SAH occurs, and its poor prognosis, the loss of productive life years from SAH is as large as that from ischemic stroke, the most frequent subtype of stroke.

The presence of an aneurysm was first demonstrated by an angiograph of the carotid artery by Moniz in 1933 (Moniz, 1933), 6 years after he had introduced this technique (Moniz, 1927). Aneurysms arise at sites of arterial branching, usually at the base of the brain, either on the circle of Willis itself or at a nearby branching point. Most intracranial aneurysms will never rupture. Risk factors for subarachnoid haemorrhage are hypertension, smoking, and excessive alcohol intake, all of which more-or-less double the risk (Feigin et a!., 2005). Diagnosis is made by CT brain scanning. Recurrent bleeding is the most imminent danger; a first aim is therefore occlusion of the aneurysm. Endovascular obliteration by means of platinum spirals (coiling) is the preferred mode of treatment, but some patients require a direct neurosurgical approach (clipping) (Molyneux et al., 2002).

Delayed cerebral ischemia

For those patients who survive the first 24 hours after the haemorrhage, delayed cerebral ischemia (DCl) is consistently the leading cause of death and disability, adversely affecting more than one in five of all patients who have suffered SAH and survived (Dorsch, 1995; Roos et al., 2000).

Unlike thrombo-embolic stroke, cerebral ischemia after SAH has a gradual onset and often involves more than the territory of a single cerebral artery or one of its branches. The clinical manifestations evolve gradually, over several hours, and consist of hemispheric focal deficits, a reduction in the level of consciousness, or both. It is mainly a diagnosis of exclusion, when clinical deterioration occurs and hydrocephalus, recurrent bleeding, hypoxia, and metabolic abnormalities have been ruled out. The peak frequency of cerebral ischemia is from 5*14 days after SAH. The time course for delayed cerebral ischemia parallels that of angiographic vasospasm, but arterial narrowing – a complex process in Itself – is neither a necessary nor a sufficient condition. Although about 70% of patients may develop arterial narrowing, only 20-30% will manifest neurological deficits (Weir et al., 1999; Rabinsteln ef al., 2004).

Nimodipine is a dlhydropyrldlne calcium channel blocker that selectively blocks the voltage-dependent L-type calcium channel. Nimodipine is accepted as beneficial in preventing DCl and subsequent poor outcome when given orally 6 times (60 mg) for 21 days after the haemorrhage (Dorhout Mees et al., 2007b). The initial thought was that nimodipine could prevent or ameliorate cerebral vasospasm, but nimodipine does not influence large vessel diameter and its pivotal method of action must be explained otherwise. Cerebral vasospasm, although often used as a synonym, apparently is a too simplistic explanation for delayed cerebral ischemia.

Research, however, has focused on the mechanisms Involved in the development of cerebral vasospasm as a means by which cerebral ischemia can be prevented. The precise under- lying pathogenic mechanisms remain obscure, but it seems that endothelial mechanisms provide the most prominent contribution to this process, and there is growing evidence that the constituents of a subarachnoid blood clot, especially oxyhaemoglobin, a product of haemoglobin breakdown, seem to be the principle initiating factor (Pluta et al., 1998; Dumont ef al., 2003; Kolias et al., 2009). The current mainstay of preventing and treating DCl include neuroprotection with nimodipine and maintenance of normovolemia, but even with this strategy DCl still occurs in up to 30% of patients and improvement in clinical outcome has been modest (Hop et ol., 1997; Brilstra et al., 2000; Roos et al., 2000; Dorhout Mees et al., 2008). Reducing the consequences of ischemia through neuroprotection may therefore improve outcome after SAH. DCl after SAH is a unique pathophysiological process because the timing and development of potential cerebral ischemia can be predicted. The interval of 4 or more days between the bleeding and the onset of ischemia provides an opportunity for preventive treatment.

Magnesium and cerebral ischemia: opportunities and concerns

Magnesium is a popular drug for acute treatment of ischemia because of its ease of use, low cost, and favourable side-effect profile. In several experimental models of cerebral ischemia, a significant neuroprotective effect of magnesium has been demonstrated with reported infarct reduction of 25*61% (Izumi et al., 1991; Marinov et al., 1996; Muir, 1998). Putative modes of action include inhibition of the presynaptic release of the excitatory amino acid glutamate, and blockade of the postsynaptic NMOA glutamate receptor and voltage dependent calcium channels (van den Bergh et al., 2004b).

Magnesium reduces the production of endothelin and completely attenuates the vasoconstrictive effect of endothelin, possibly by inhibiting calcium channel-mediated, smooth-musde contraction (Kemp et al., 1999; Berthon et al., 2003). Furthermore, magnesium Increases cardiac contractility, which may improve cerebral perfusion in dysautoregulated brain tissue. However, acute ischemic stroke has not the quality of cerebral vasospasm-associated OCl in SAH that occurs mostly more than 4 days after the Initial bleeding, allowing the administration of a treatment before ischemia develops. Delayed treatment might explain why magnesium was ineffective in ischemic stroke in the IMAGES study (Muir et al., 2004). Again a precedent exists for nimodipine, which was ineffective when tested in ischemic stroke.

A study of acute administration of magnesium in stroke (FAST- MAG) is underway {Saver et al., 2004). Another concern about the suitability of magnesium to protect the ischemic brain is that regulation of cerebrospinal fluid magnesium is largely maintained following acute brain injury and limits the brain bioavailability of magnesium sulphate. Current dose regimens may only marginally increase concentrations of magnesium In CSF in brain-injured humans (McKee et al., 200S), with this modest increase occurring in the range of 10-19%. However, experimental evidence suggests that this modest elevation may be sufficient for neuroprotection.

In a study using, lP-MR spectroscopy (MRS) to investigate intracellular brain free magnesium in 37 patients with good-grade aneurysmal SAH, patients were randomized for magnesium therapy and compared (Wong et al., 2009). Magnesium treatment after aneurysmal SAH produced a small (+13%; mean difference 0.018 ± 0.007 mM), but significant elevation of intracellular free magnesium. In a randomized controlled trial where magnesium therapy was compared with nimodipine, CSF magnesium concentration was non-statistically higher in the magnesium group (1.19 ± 0.20 vs 1.14 ± 0.21; p= 0.26) (Schmid- Elsaesser et al., 2006). The negative effect of magnesium on blood pressure may also have an Impact on results.

Currently, vasopressors are often used to prevent hypotension In SAH patients, and DCl is often treated with induced hypertension with the aim to increase cerebral perfusion. However, the lowering effect of magnesium on blood pressure is limited and far less than that of nimodipine. Finally, there Is a dose related association between magnesium sulphate administration and hypocalcaemla, but symptomatic hypocalcaemia is rare and It does not result In an increased risk of poor outcome (van den Bergh er al., 2008b). There is no basis for routine supplementation of calcium In patients with hypocalcaemia when treated with magnesium. Of course, supplementation might be considered In Individual patients to prevent isolated clinical symptoms of hypocalcaemia such as muscle weakness or cardiac arrhythmias.

Hypomagnesemia in subarachnoid haemorrhage

In observational clinical studies, the vast majority of SAH patients develop hypomagnesaemia at some time within three weeks after the haemorrhage (Warren et al., 1993; van den Bergh er al., 2003). Low magnesium serum levels are frequently present at admission and are thus most likely caused by intracellular shift of magnesium ions. Intracellular Mg2′ levels are indeed increased in SAH. However, 90% of the intracellular Mg2’ is complexed with ATP, and the increase of intracellular Mg2′ during ischemia may also be the result of the release of Mg2’ from this complex. ATP binds with Mg2’ with an associate constant of 4, while binding affinity with ADP is about 2 times smaller. The cytosolic and mitochondrial Mg2’ concentration will increase in cells with a poor energy state and less ATP (Jung et al., 1990; Saris et al., 2000; Wong et al., 2009). The increase of intracellular Mg2’ is even less than might be expected from ATP utilization, probably because of a disappearance of Mg2’ by binding to other cell components. Hypomagnesaemia that developed later on in the course might also be caused by renal loss as very often is the case in critically ill patients.

The diminished availability and subsequent decreased extracellular Mg2’ after SAH results in significantly increased intracellular free Ca2’ in cerebral vascular muscle cells. This may cause cerebral microvascularconstriction, followed by a prolnflammatory response, inducing vascular smooth muscle, endothelial and neuronal cell damage (Barbour ef al., 2002). Furthermore, hypomagnesaemia results in a reduced endothelial NO release by which means hypomagnesaemia can induce vasoconstriction (Pearson er al., 1998; Shechter, 2000). All the above gives rise to the concept of a causal relationship between the decreased availability of magnesium and the development of vasospasm-associated DCl.

When testing this hypothesis in 107 prospectlvely studied consecutive patients, we found that hypomagnesaemia during day 4 and 10 after the haemorrhage increased the risk for DCl three fold (van den Bergh et ol., 2003). Apart from DCl, in a study in 62 patients, more than one-third had hypomagnesaemia at admission which was independently related to cardiac dysrhythmias, especially a long PR interval and a shorter QTc interval (van den Bergh et al., 2004a).

This link with cardiac complications might be another reason why hypomagnesaemia may be related to poor outcome (van der Bilt et a!., 2009).

Magnesium treatment in subarachnoid haemorrhage

Animal studies

Magnesium has a vasorelaxing effect in oxyhaemoglobin-induced vasospasm and it ameliorates cerebral vasospasm in experimental SAH (Miura, 1988; Ram et ol., 1991; Pyne et ol., 2001), although no effect of magnesium sulphate on the angiographic diameter of large cerebral vessels in a monkey model of SAH could be noted (Macdonald et ol., 2004). Neither does magnesium infusion decrease middle cerebral artery blood flow velocities in patients with DCI (van den Bergh, unpublished data). Magnesium induces a dose-dependent vasodilatation, reduces cerebrovascular tone.

Increases CBF and protects the metabolism (Seelig et ol., 1983; Altura and Altura, 1984; Torregrosa et ol., 1994) However, in a human study in 6 patients, magnesium sulphate infusion did not increase cerebral blood volume and cerebral blood flow compared with pre- treatment levels, or compared with 6 control patients (Wong et oi, 2010a).

We have demonstrated In an experimental model that the duration of Ischemic depolarizations after SAH is substantially reduced after pre- treatment with magnesium sulphate (van den Bergh et ol., 2002). Magnesium also postpones anoxic depolarization (van der Hel et ol., 1998).

This maintenance of the membrane potential may, at least partly, explain the neuroprotective properties of magnesium in SAH. We have shown In a rat model of SAH that pre-treatment with agnesium sulphate reduces acute cerebral lesion volume with more than 60% (van den Bergh et ol., 2002).

Large randomized controlled trial

In 2005, the “Magnesium and Acetylsalicyllc acid In Subarachnoid Haemorrhage” (MASH-l) trial was published, in which 283 patients between November 2000 and January 2004 were randomized to placebo or a continuous infusion of 64 mmol/l/day of magnesium sulphate starting  within 4 days of SAH and continuing 14 days after occlusion of the aneurysm (van den Bergh et ol., 2005).

The results showed a trend towards less cerebral ischemia in patients allocated to magnesium and improved overall outcome, with a 35% reduction in DCI (RR 0.65; 95% Cl 0.40- 1.05) and a 23% reduction in poor outcomes (RR 0.77; 0.54-1.09). The mean magnesium level during treatment was 1.47 ± 0.32 mmol/L, and side effects were mild and sparse (van Norden et ol., 2005). A post-hoc analysis suggested that the treatment effect of magnesium is larger after endovascular occlusion than after neurosurgical clipping, which underlines the conclusion that further trials are not at risk of being underpowered If the proportion of endovascular treated patients increases (van den Bergh er ol., 2009).

In 2010, the results of the Intravenous Magnesium Sulphate in Aneurysmal Subarachnoid Haemorrhage (IMASH) trial were published (Wong et ol., 2010b). IMASH was a phase III randomized, clinical, international multicentre trial that evaluated the effect of magnesium sulphate on the clinical outcome of 327 patients with aneurysmal SAH, of which 90% were recruited in Hong Kong and China and the remaining 10% In Australia from 2002 to 2008. After randomization, 20 mmol of magnesium sulphate was given over 30 minutes, followed by infusion of 80 mmol/day, or equivalent volume of saline, within 48 h after onset of symptoms and continued for up to 14 days from the day of haemorrhage.

Mean time from ictus to start of study drug infusion was 32 i 15 (SO) hours. More than 90% of the patients completed at least 10 days of study drug infusion. The randomization was single-blinded. The study aimed for a plasma magnesium concentration in the treatment group of twice the serum baseline level, but below 2.5 mmol/L. Average serum magnesium concentrations in the treatment group were 1.67 s 0.27 mmol/L compared with 0.91 x 0.16 mmol/L In controls.

The proportion of patients with a 6-month favourable outcome, defined as an extended Glasgow Outcome Scale of 5 to 8, were similar for both treatment groups (OR 1.0; 95% Cl 0.7-1.6). Also in the secondary outcome analyses, which included incidence of clinical vasospasm, Barthel Index, modified Rankin score, modified National Institute of Health Stroke Score, and MCA velocities as measured by transcranial Doppler, there were no significant differences between the 2 groups. Institute of Health Stroke Score, and MCA velocities as measured by transcranial Doppler, there were no significant differences between the 2 groups.

Although the study was underpowered, the results do not support a significant clinical benefit of magnesium therapy in SAH. Possible explanations for the differences in results between MASH-1 and IMASH are the assumed predominant Asian origin of recruited patients and the time window of administration; too late or too short. In the MASH-1 study the median start of treatment was 28 hours, and It was continued for at least 2 weeks. On the other hand, the achieved serum magnesium concentration might have been too high, as in the treatment group higher magnesium concen- trations were associated with a worse outcome (Wong et al., 2010c). Furthermore, the average serum magnesium concentration of 0.91 ± 0.16 mmol/L in the control group was quite high and could have diluted the results.

Is there an optimal serum magnesium concentration for neuroprotection?

In animal models of stroke, evidence was found for a dose-response effect, with optimal serum magnesium concentrations of approximately 1.40 to 1.50 mmol/l (Marinov et ol., 1996; Yang et ol., 2000; Miles eta!., 2001). The mean serum magnesium level of 1.47 mmol/L achieved in the MASH-1 study was close to the proposed optimal serum magnesium level for achieving maximal neuroprotection. In a substudy performed in patients with a serum magnesium concentration a 1.10 mmol/L, we found no linear relationship between serum magnesium levels and risk reduction for DCI (Dorhout Mees et ol., 2007a).

Magnesium sulphate therapy results in a stable risk reduction of DCI over a broad range of achieved serum magnesium concentrations, and strict titration of the dosage therefore does not seem necessary, although concentrations below 1.28 mmol/l could decrease the effect on DCI. In contrast, no such relationship was observed for poor outcome. Risks tended to be higher In patients with magnesium concentrations above the 75th percentile (1.62 mmol/L) with an OR for poor outcome of 4.9 (95% Cl 1.2 to 19.7) in patients with a serum concentration above 1.62 mmol/L compared to the lowest quartile (1.10-1.28 mmol/L).

The aim of the IMASH study to achieve a plasma magnesium concentration of twice the serum baseline level resulted in an average serum magnesium concentration in the treatment group of 1.67 ± 0.27 mmol/L, which might have had a negative effect on clinical outcome. Of course, an optimal neuroprotective concentration of serum magnesium can only be validated if current large trials show that magnesium therapy is effective.

Systematic review and meta-analysis

We assessed In a systematic review whether magnesium decreases the occurrence of secondary ischemia and poor outcome after aneurysmal SAH. We sought to identify all unconfounded clinical controlled trials with magnesium In patients with aneurysmal SAH. Trials were Identified in the Stroke Group Trials Register of the Cochrane Library, PUBMED and MEDLINE. Outcome measures were DCI and poor outcome. An estimate of the treatment effect across trials was calculated with the Mantel-Haenszel method according to the intention-to- treat principle.

Between 2002 and 2010, four non-randomlzed and eight randomized controlled trials on magnesium were published, all but one comparing magnesium therapy with placebo in addition to nlmodiplne. The exception was a small randomized controlled trial in 104 patients that found similar outcomes in patients treated with nlmodiplne versus patients treated with magnesium (Schmld-Elsaesser et ol., 2006). When the four non-randomized studies are combined, 118 patients with magnesium therapy were compared with a historical control group of 138 patients (Chla et ol., 2002; Prevedello et al., 2006; Stlppler ef ol., 2006; Friedlich et ol., 2009). When the results are pooled. It all bolls down to the conclusion that a trend towards less symptomatic vasospasm was found In favour of the treatment group and, if provided, a positive effect with a less extent on outcome.

Seven trials totalling 9S4 patients were eligible for the meta-analyses (Table 1) (Veyna et ol., 2002; van den Bergh et al., 2005; Wong et al., 2006; Muroi et al., 2008; Westermaier et al., 2010; Akdemir et al., 2010; Wong et al., 2010b).

All used magnesium sulphate in addition to nimodipine. The overall relative risk for DCI was 0.86 (95% Cl 0.67-1.11) (Figure 1), and for poor outcome 0.89 (95% Cl 0.78-1.02) (Figure 2).


In patients with SAH, headache typically is severe, usually lasts for 1-2 weeks and often requires treatment with opioids. Opioids, however, have a wide range of potential side effects, which may have a negative effect on patient comfort and functional outcome. Blocking the NMDA receptor is thought to be involved in pain modulation by preventing the induction of central pain sensitization. Several studies have reported that magnesium reduces perioperative pain or analgesic requirements.

Headache relief would be an important additional effect of magnesium because this would reduce the need of opioids with their potentially harmful side effects. In a sub-study of the MASH-1 trial we found that magnesium treatment was associated with less severe headache and less frequent use of opioids (Dorhout Mees et al., 2010). These data imply that intravenous magnesium therapy, besides a supposed beneficial effect on outcome, also provides pain relief for SAH patients.

On-going phase 3 clinical controlled trials

There is currently one on-going phase 3 clinical controlled trial evaluating the effect of magnesium in SAH. The Magnesium in Aneurysmal Subarachnoid Haemorrhage (MASH-II) study is a phase III randomized, clinical, international multicentre trial that studies the effect of magnesium sulphate after aneurysmal SAH [ISRCTN68742385] (van der Bilt et al., 2009).

Study medication (64 mmol/day magnesium sulphate or placebo) is given via continuous infusion within 4 days and until 20 days after the haemorrhage. Outcome is determined with the modified Rankin scale three months after the haemorrhage. Analysis will be according to the intention-to-treat principle. So far, in August 2010, over 1000 patients have been included in 6 Dutch, 1 UK, and 1 Chilean hospital. Based on the results of the MASH-1 study sample size, calculations indicate that 1200 patients are needed to give a statistically significant result (with a = 5% and a power of 80%). We aim to include these patients before the end of 2011. A first interim analysis was performed after 300 patients and the second and last interim analysis was performed after recruitment and follow-up of 750 patients. At both interim analyses, the Data Monitoring Committee gave unanimous advice to continue the study.


In conclusion, magnesium is a promising agent to prevent the occurrence of delayed cerebral ischemia and to improve outcome in patients with SAH. Hence, at present, the evidence is insufficient and larger randomized controlled  trials are needed to determine the safety and efficacy of magnesium sulphate infusion before it can be recommended for patients with aneurysmal SAH. Currently, a large phase III trial is being conducted that will hopefully provide definite evidence whether magnesium treatment is beneficial in SAH patients. In the meantime, it is probably wise to maintain serum magnesium in the high-normal range, e.g. 0.8-1.0 mmol/l, after SAH to prevent the hypomagnesaemia associated neurological and cardiac complications. Magnesium therapy can be considered as a safe alternative for morphine in the treatment of headache. In case magnesium therapy is considered as a remedy to improve outcome, target serum magnesium levels are probably around 1.4 mmol/l and better not exceed 1.6 mmol/l.


I'm James!

Would you like to get a custom essay? How about receiving a customized one?

Check it out