Current understanding of migraine pathogenesis
The exact aetiology of migraine still remains to be defined. However, there have been significant advances in our understanding of migraine pathophysiology. The current prevailing theories are based on both neuronal (cerebral cortex, brain stem) and vascular components (the ‘trigeminovascular complex’) in patients who are genetically predisposed to migraine. In these patients, cortical hyperexcitability, characterized by low excitability thresholds and exaggerated response to stimuli, has been proposed in the initiation of an attack 1. During the attack, the trigeminovascular neurons release neurotransmitters, such as calcitonin gene-related peptide, and substance P. This leads to vasodilation, mast cell degranulation and increased vascular permeability, resulting in meningeal neurogenic inflammation 1.
Role of magnesium in migraine pathophysiology
It has been speculated that, during headaches, migraine sufferers excrete excessive amounts of magnesium, as a result of stress, resulting in a transient serum hypomagnesaemia 2. Migraine is also associated with low levels of magnesium in the cerebrospinal fluid 3, and in vivo 31P nuclear magnetic resonance spectroscopy has demonstrated low magnesium in the brain during migraine attacks and interictally in some patients 4,5.
Magnesium deficiency has been associated with neurotransmitter release 6, platelet aggregation 7 and vasoconstriction 8, all of which are relevant features in our understanding of migraine pathophysiology. Hypomagnesaemia also results in the generation and release of substance P, which is believed to act on sensory fibres and produce headache 9.
Magnesium also has a modulatory role on the sensitivity of N-methyl-D-aspartate receptors to glutamate, which plays an important role in the initiation and spreading of cortical depression 10. Animal experimental studies have shown that magnesium can inhibit the cortical spreading depression induced by glutamate 11. Magnesium also regulates the cerebral and peripheral vascular tone by acting like a physiological calcium channel blocker 2. Serotonin receptor activity is altered by changes in levels of ionized magnesium, and vasoconstriction induced by serotonin can be effectively blocked by pretreatment with magnesium 12.
Therefore, magnesium therapy may target various aspects of the neurogenic inflammation that occur during migraine attacks by counteracting vasospasm, inhibiting platelet aggregation, minimizing the formation of inflammatory mediators and stabilizing cell membranes.
There are inconsistent findings about the efficacy of magnesium in the prophylaxis of migraine attacks, although some studies reported that oral 600 mg magnesium is effective in reducing the number of attacks 13,14.
We assessed the evidence from double-blind, randomized controlled trials (RCTs) on the efficacy and tolerability of intravenous magnesium for acute migraine in adults.
Materials and methods
The four-part question presented below was modelled to guide the search.
Population: in adult patients with acute migraine attack.
Intervention: intravenous magnesium.
Comparison: compared with conventional analgesia or placebo.
Outcome: more effective for resolution of headache?
This produced the following basic search strategy:
- 1 AND 2.
The search was refined, expanded and used on the electronically indexed databases, including the Medline, EMBASE and CINAHL, through the NHS Evidence healthcare databases interface. The Cochrane Central Register of Controlled Trials was searched using the terms ‘migraine’, ‘headache’ and ‘magnesium’. Other sources used were the ACP Journal Club, National Research Register and the US Government’s Clinical Trial Database.
To identify unpublished research, we reviewed conference proceedings from major meetings on neurology, headache or emergency medicine from 2006 to 2012. In addition, hand searches were conducted of the references in the retrieved papers and through the electronic citations.
Studies were considered eligible for the meta-analysis if they were double-blind, RCTs of intravenous magnesium given for acute migraine attacks in adults. We included trials only if they were conducted in a setting that indicated headache as an acute episode – emergency department or headache clinic.
Two independent reviewers identified studies for eligibility. Papers considered potentially relevant were selected and the full manuscripts were reviewed for inclusion. Two independent reviewers abstracted information on patients, study design, primary and secondary outcomes, key results and study weaknesses onto specially designed, pretested forms. Disagreements were resolved by consensus.
The primary outcome chosen for the meta-analysis was the percentage of patients with headache relief in acute migraine attacks at 30 min following treatment. There were many variations on the selected studies’ choice of primary outcomes and their endpoints, hence making it difficult to combine the results for other outcomes. For example, to explore the mean difference in visual analogue scale (VAS) score, one would need the mean and SD for each group. These data were missing in some papers.
The secondary outcomes chosen were the percentage of patients reporting side-effects or adverse events, and the percentage of patients requiring rescue analgesia. All statistical analyses were carried out using StatsDirect statistical software (StatsDirect Ltd, Cheshire, UK).
We identified 1203 abstracts, of which nine 15–23 were potentially relevant articles (Tables 1 and 2). Independent review of these articles led to the inclusion of five RCTs for the primary outcome analysis 15–19.
These five RCTs were all double-blinded, and conducted in the emergency departments or acute headache clinic settings. One RCT 17 was compared directly between intravenous magnesium and prochlorperazine, but the others 15,16,18,19 were placebo-controlled.
One RCT 20 was excluded from the primary outcome analysis as it was not double-blinded and lacked the acceptable randomization process. But its data formed part of the adverse events analysis as other selected studies 19 failed to report such findings.
Magnesium versus placebo
Two studies compared intravenous magnesium with placebo. A study by Frank et al. 15 focused on the median difference in 100 mm VAS at 30 min as the primary outcome. The outcome difference was not statistically significant (3-mm median improvement in magnesium group vs. 8-mm median improvement in placebo group, P=0.63). Only interim analysis was carried out because of slow patient recruitment, and the study was underpowered with a possibility of a type II error.
A study by Bigal et al.19 examined the percentage of response to treatment as measured by improvement in pain intensity on the four-point scale (0 – no pain and 3 – severe pain) at 30 and 60 min. The differences between placebo and magnesium were not significant in migraine without aura at 30 min (10% placebo vs. 16.7% MgSO4, P>0.05) and at 60 min (16.7% placebo vs. 33.3% MgSO4 at 60 min, P>0.05), and in migraine with aura at 30 min (3.3% placebo vs. 16.7% MgSO4, P>0.05).
However, magnesium was more effective than placebo for headache relief (50 vs. 13.3%, P<0.05) at 60 min in migraine with aura.
A study by Bigal et al.19 lacked the methodological rigour. Patients were randomized ‘by drawing lots’, and no further details were given regarding their method of randomization. It was not clear whether allocation concealment and intention-to-treat analysis were observed in the study, and their method of follow-up was not clear.
Magnesium versus metoclopramide versus placebo
Cete et al.16 chose the primary endpoint as the VAS difference in pain at 30 min.
The changes were significant in all groups [metoclopramide 73 (0 min), 38 (30 min) vs. magnesium 70 (0 min), 39 (30 min) vs. placebo 70 (0 min), 47 (30 min), P<0.0001], but no significant difference was observed between the groups (P=0.619).
A small sample size leads to a risk of type II error.
Magnesium versus prochlorperazine
Ginder et al.17 chose the mean VAS score reduction for headache at 30 min as the primary endpoint. Intravenous prochlorperazine was superior to magnesium in headache relief (47 mm prochlorperazine vs. 24 mm magnesium, P=0.045). However, the authors did not perform the sample size calculation before the study, and a small study sample means that the study is likely to be underpowered.
Magnesium with metoclopramide versus placebo with metoclopramide
Corbo et al.18 chose the differences in initial (0 min) and final (45 min) 100 mm VAS score for headache as the primary outcome. Intravenous placebo with metoclopramide were more effective than intravenous magnesium with metoclopramide in headache relief [71 mm in placebo plus metoclopramide vs. 55 mm in magnesium plus metoclopramide, mean difference 16; 95% confidence interval (CI) −2 to 34]. However, the wide CIs suggest a possible type II error.
Percentage of patients with pain relief after 30 min
A meta-analysis using a random-effects model demonstrated pooled risk difference equal to −0.07 (95% CI=−0.23 to 0.09), which was not statistically significant from 0 (P=0.3894) in the percentage of subjects with pain relief after 30 min. This means that when results of the five RCTs were combined in a random-effects meta-analysis the percentage of patients who experienced reduction in pain after 30 min was 7% lower in the magnesium groups compared with the control groups. Thus no beneficial effect of the treatment was demonstrated. The output from the meta-analysis is shown below.
Percentage of patients reporting side-effects/adverse events
In this analysis, we used the data from Demirkaya et al.20 in place of Bigal et al.19 as Bigal et al. did not report side-effects. The meta-analysis using a random-effects model demonstrated pooled risk difference equal to 0.370 (95% CI=0.06–0.68), which was statistically significant from 0 (P=0.021) in the percentage of patients reporting side-effects or adverse events. This means that when results of the five studies were combined in a random-effects meta-analysis the percentage of patients who experienced side-effects or adverse events was greater in the magnesium groups compared with the control groups by 37%. Thus, those receiving intravenous magnesium experienced significantly more side-effects/adverse events than those in the control groups. The output from the meta-analysis is shown below.
Need for rescue medications
In this analysis, we estimated the numbers in the paper by Bigal et al.19 using the graphs published in the paper. The meta-analysis using a random-effects model demonstrated pooled risk difference equal to −0.021 (95% CI=−0.16 to 0.12), which was not statistically significant from zero (P=0.7679) in the percentage of patients who required rescue medication. The percentage of patients who needed rescue medications was slightly lower in the control groups, but this was not significant. The output from the meta-analysis is shown below.
Five RCTs 15–19 confirmed that intravenous magnesium showed no statistically significant benefit over placebo, metoclopramide or prochlorperazine in the treatment of acute migraine. In fact, in one RCT 18, magnesium appeared to attenuate the effectiveness of metoclopramide when used as an adjunctive therapy. One RCT 19 demonstrated the superior therapeutic efficacy of magnesium over placebo at 60 min post-treatment, but only amongst migraine sufferers with aura.
The meta-analyses failed to demonstrate a beneficial effect of intravenous magnesium in terms of reduction in pain relief in acute migraine in adults, showed no benefit in terms of the need for rescue medication and in fact have shown that patients treated with magnesium were significantly more likely to report side-effects/adverse events.
Heterogeneity of the studies
The studies were of variable quality. Most studies were underpowered to detect primary and secondary outcomes, the study populations were the convenience sampling, and some did not report inclusion and exclusion criteria in sufficient detail.
Variety of comparators
Different studies adopted different comparators, that is magnesium versus metoclopramide versus prochlorperazine versus placebo.
The variety of comparators is not surprising given the lack of consensus about a standard of care for acute migraine. Scientifically sound and clinically relevant guidelines are lacking. In particular, there is insufficient evidence to establish how to select one therapy over another. For example, several studies 24,25 have shown that prochlorperazine intravenously and intramuscularly was significantly superior to metoclopramide in headache relief. However, prochlorperazine does not have a UK marketing authorization for this indication, except for the relief of nausea and vomiting 26.
Variety of outcome measures
Many different primary and secondary outcomes were reported. The most commonly used primary outcome measure was either the change in 100 mm VAS pain score, 15–18 or the verbal numeric rating (0–10) scale 21. The latest International Headache Society guidelines recommend that a four-point scale for the headache pain intensity (0 – no headache; 1 – mild; 2 – moderate; 3 – severe), or alternatively, VAS should be used for all controlled trials of drugs in migraine 27. Many studies that used the VAS score as their primary outcome failed to report SDs, preventing statistical pooling.
Different types of benign headaches
Clinically the distinction between types of benign headaches in the emergency department often blurs, and clinicians have difficulty distinguishing between different types of headaches 28,29. A study by Frank et al.15 included patients with all types of benign headaches, although 93% of patients were diagnosed with migraine. Ginder et al.17 made no attempt to classify the different types of benign headaches enrolled in the study. Cete et al.16, Corbo et al.18, and Bigal et al.19 only included patients who met the International Headache Society criteria for acute migraine.
It is possible that the results may have differed if only patients meeting strict criteria for migraine were included in the meta-analysis.
Only the study by Bigal et al.19 looked at the differential response to intravenous magnesium in migraine with and without aura. They demonstrated that magnesium was more effective than placebo for headache relief, but only at 60 min and only among migraine sufferers with aura. Other studies made no attempt at differentiating between different types of migraine. The treatment response and side-effect profile may be different between acute migraine with and without aura.
Primary outcome endpoints
The primary outcome endpoint for all studies was at 30 min, except for Corbo et al.18, where the outcomes were measured at 15, 30 and 45 min, and Bigal et al.19, where the endpoints were at 30 and 60 min.
The latest International Headache Society guidelines suggest the percentage of patients pain-free at 2 h before any rescue medication as the primary outcome of choice for all controlled trials of drugs in migraine 27. However, they also accept that pain freedom at a point earlier than 2 h could be considered for parenteral (e.g. intravenous, intramuscular, subcutaneous) test drugs 27. It is possible that magnesium may have shown greater effectiveness at 1 or 2 h from the point of treatment. However, previous migraine studies have shown that with the intravenous route of drugs, the maximal absorption and distribution occurs within minutes, and most patients who respond to intravenous treatment do so within this interval 30–32. There is also an ethical argument that the use of a prolonged outcome endpoint at 60 or 120 min in the emergency department is inappropriate, as most patients presenting to the department with headache do so because of its severity and require prompt symptom relief. If magnesium’s onset of analgesia is delayed to this extent, then it may not be an appropriate choice of analgesia for emergency symptom relief.
Our meta-analyses pooled studies with different population characteristics, discrepant primary and secondary outcomes with different treatment comparators. Hence, it undermines our confidence in drawing conclusions on the relative merits of intravenous magnesium in the treatment of acute migraine in adults.
Future studies need to adopt more robust internal validity, with adequately powered sample size, acceptable randomization method, allocation concealment, clear inclusion criteria, double-blinding, use of a placebo as a controlling agent, intention to treat analysis and adequate follow-up.
Consensus on the standard care of acute migraine based on evidence will help to avoid the heterogeneity of research methodology, as observed in our meta-analyses and overall, improve the quality of research in acute migraine.
Conflicts of interest
There are no conflicts of interest.
1. Martin VT, Behbehani MM.Towards a rational understanding of migraine
trigger factors.Med Clin North Am2001;85:911–941.
2. Sun-Edelstein C, Mauskop A.Role of magnesium in the pathogenesis and treatment of migraine
.Expert Rev Neurother2009;9:369–379.
3. Jain AC, Sethi NC, Balbar PK.A clinical electronencephalographic and trace element study with special reference to zinc, copper and magnesium in serum and cerebrospinal fluid in cases of migraine
4. Boska MD, Welch KMA, Barker PB, Nelson JA, Schultz L.Contrasts in cortical magnesium, phospholipid and energy metabolism between migraine
5. Ramadan NM, Halvorson H, Vande-Linde A, Levine SR, Helpern JA, Welch KM.Low brain magnesium in migraine
6. Coan EJ, Collingridge GL.Magnesium ions block an N
-methyl D-aspartate receptor-mediated component of synaptic transmission in rat hippocampus.Neurosci Lett1985;53:21–26.
7. Baudouin-Legros M, Dard B, Guichency P.Hyperreactivity of platelets from spontaneously hypertensive rats. Role of external magnesium.Hypertension1986;8:694–699.
8. Altura BT, Altura BM.Withdrawal of magnesium causes vasospasm while elevated magnesium produced relaxation of tone in cerebral arteries.Neurosci Lett1989;20:323–327.
9. Weglicki WB, Phillips TM.Pathobiology of magnesium deficiency: a cytokine/neurogenic inflammation hypothesis.Am J Physiol1992;263:R734–R737.
10. Ferrari MD.Biochemistry of migraine
11. Mody I, Lambert JD, Heinemann U.Low extracellular magnesium induces epileptiform activity and spreading depression in rat hippocampal slices.J Neurophysiol1987;57:869–888.
12. Goldstein S, Zsoter TT.The effect of magnesium on the response of smooth muscle to 5-hydroxytryptamine.Br J Pharmacol1978;62:507–514.
13. Peikert A, Willimzig C, Kohne-Volland R.Prophylaxis of migraine
with oral magnesium: results from a prospective multi-centre, placebo-controlled and double-blind randomised study.Cephalalgia1996;16:257–263.
14. Koseoglu E, Talashoglu A, Gonul AS, Kula M.The effects of magnesium prophylaxis in migraine
without aura.Magnes Res2008;21:101–108.
15. Frank LR, Olson CM, Shuler KB, Gharib SF.Intravenous magnesium for acute benign headache in the emergency department: a randomised double-blind placebo-controlled trial.CJEM2004;6:327–332.
16. Cete Y, Dora B, Ertan C, Ozdemir C, Oktay C.A randomised prospective placebo-controlled study of intravenous magnesium sulphate vs. metoclopramide in the management of acute migraine
attacks in the emergency department.Cephalalgia2005;25:199–204.
17. Ginder S, Oatman B, Pollack M.A prospective study of IV magnesium and IV prochlorperazine in the treatment of headaches.J Emerg Med2000;18:311–315.
18. Corbo J, Esses D, Bijur PE, Iannaccone R, Gallagher EJ.Randomised clinical trial of intravenous magnesium sulphate as an adjunctive medication for emergency department treatment of migraine
headache.Ann Emerg Med2001;38:621–627.
19. Bigal ME, Bordini CA, Tepper SJ, Speciali JG.Intravenous magnesium sulphate in the acute treatment of migraine
without aura and migraine
with aura. A randomised, double-blind, placebo-controlled study.Cephalalgia2002;22:345–353.
20. Demirkaya S, Vural O, Dora B, Topcuoglu MA.Efficacy of intravenous magnesium sulphate in the treatment of acute migraine
21. Zidverc-Trajkovic J, Pavlovic AM, Jovanovic Z, Sternic N, Kostic VS.Efficacy of intravenous magnesium sulphate in severe migraine
attacks.J Headache Pain2001;2:79–82.
22. Mauskop A, Altura BT, Cracco RQ, Altura BM.Intravenous magnesium sulphate relieves migraine
attacks in patients with low serum ionised magnesium levels: a pilot study.Clin Sci1995;89:633–636.
23. Mauskop A, Altura BT, Cracco RQ, Altura BM.Intravenous magnesium sulphate rapidly alleviates headaches of various types.Headache1996;36:154–160.
24. Coppola M, Yealy DM, Leibold RA.Randomized, placebo-controlled evaluation of prochlorperazine versus metoclopramide for emergency department treatment of migraine
headache.Ann Emerg Med1995;26:541–546.
25. Jones J, Pack S, Chun E.Intramuscular prochlorperazine versus metoclopramide as single agent therapy for the treatment of acute migraine
headache.Am J Emerg Med1996;14:262–264.
26. .Headaches: diagnosis and management of headaches in young people and adults, NICE Clinical Guideline 1502012.London:NICEAvaliable at: http://guidance.nice.org.uk/cg150
[Accessed 1 March 2013].
27. Tfelt-Hansen P, Pascual J, Ramadan N, Dahlof C, D’Amico D, Diener HC, et al..Guidelines for controlled trials of drugs in migraine
: third edition. A guide for investigators.Cephalalgia2012;32:6–38.
28. Thomas SH, Stone K.Emergency department treatment of migraine
, tension, and mixed-type headache.J Emerg Med1994;12:657–664.
29. Marcus DA.Migraine
and tension-type headaches: the questionable validity of current classification systems.Clin J Pain1992;8:28–36.
30. Ellis GL, Delaney J, DeHart DA.The efficacy of metoclopramide in the treatment of migraine
headache.Ann Emerg Med1993;22:191–195.
31. Tek DS, McClellan DS, Olshaker JS, Allen CL, Arthur DC.A prospective, double-blind study of metoclopramide hydrochloride for the control of migraine
in the emergency department.Ann Emerg Med1990;19:1083–1087.
32. Bell R, Montoya D, Shuaib A.A comparative trial of three agents in the treatment of acute migraine
headache.Ann Emerg Med1990;19:1079–1082.