REF-1595 · Garrison et al., 2020 · Cochrane Database of Systematic Reviews 2020, issue · Paper · PubMed · Free full text

In older adults with presumed night-time leg cramps, magnesium did not reduce how often cramps came (5 trials, 307 people, moderate certainty) and showed no significant effect on their intensity or duration; the pregnancy trials conflict and are all at high risk of bias, and the review found no randomised trial of magnesium for cramps during exercise (search to September 2019). From the abstract

Main caveat: Written from the abstract: magnesium forms, doses and trial lengths are not given there, and the trial-level detail was not read.
fairly strong evidenceMeasured on a mixed groupRead from the abstract only

Garrison and colleagues updated the 2012 Cochrane review of magnesium for skeletal muscle cramps. On 9 September 2019 they searched eight databases and two trial registries for randomised trials of magnesium, in any form, against no treatment, placebo or another cramp therapy, in any group of people with cramps.

Eleven trials with 735 participants qualified (118 people in crossover trials also served as their own controls): five in pregnancy-associated leg cramps (408), five in idiopathic cramps (271) and one in liver cirrhosis (29). Every trial gave magnesium by mouth except one, which used slow intravenous infusions.

For idiopathic cramps, studied mostly in older adults thought to have night cramps, magnesium and placebo did not differ significantly on any measure of cramp frequency, intensity or duration at four weeks. The pregnancy trials could not be pooled and pointed in different directions. Oral magnesium came with mainly gastrointestinal side effects.

The authors judge it unlikely that magnesium gives clinically meaningful cramp prevention to older adults, and they found no trial in exercise-associated cramps.

Participants

11 trials, 735 participants (118 crossover participants counted again as controls).
Idiopathic cramps: 5 trials, 271 participants, mean age 61.6 to 69.3 years.
Pregnancy-associated leg cramps: 5 trials, 408 women.
Liver cirrhosis: 1 trial, 29 people, only some of them with cramps.

Protocol

Search (9 September 2019): MEDLINE, Embase, CENTRAL, the Cochrane Neuromuscular Specialised Register, CINAHL Plus, LILACS, AMED and SPORTDiscus; WHO-ICTRP and ClinicalTrials.gov for ongoing or unpublished trials; a citation search in ISI Web of Science.
Selection: randomised trials of magnesium supplementation to prevent skeletal muscle cramps in any clinical group; comparators no treatment, placebo or another therapy.
Comparisons: nine trials against placebo; one against no treatment, calcium carbonate or vitamin B; one against vitamin E or calcium.
Methods: two reviewers selected trials, extracted data and assessed risk of bias; participant-level data were obtained for four trials, one of them unpublished.
Funding and interests: PubMed lists support from the Canadian Institutes of Health Research (CIHR); the first author is an author of one included study.

Key Results

Idiopathic cramps, magnesium vs placebo, at 4 weeksEffect (95% CI)Trials (people)Certainty
% change in cramps per week from baseline (primary)MD −9.59% (−23.14% to 3.97%)3 (177)moderate
Cramps per weekMD −0.18 (−0.84 to 0.49)5 (307)moderate
At least 25% fewer crampsRR 1.04 (0.84 to 1.29)3 (177)high
Cramps rated moderate or severeRR 1.33 (0.81 to 2.21)2 (91)moderate
Most cramps lasting 1 min or moreRR 1.83 (0.74 to 4.53)1 (46)low
Pregnancy: no meta-analysis was possible. Against no treatment, one trial found no significant benefit on a three-point efficacy scale; of three placebo trials, one found no benefit, one found benefit on frequency and intensity, and one gave frequency results that could not be reconciled. All five pregnancy trials were judged at high risk of bias.
Cirrhosis: one small trial with limited reporting found no difference in cramp frequency or intensity.
Harms (all settings pooled): major adverse events in 2 of 72 people on magnesium and 3 of 68 on placebo; withdrawals for adverse events did not differ significantly; minor adverse events were more frequent on magnesium, without a statistically clear difference (RR 1.51, 95% CI 0.98 to 2.33; 4 trials, 254 people; low certainty); gut effects such as diarrhoea affected 11% (10% on control) to 37% (14% on control).
Gaps: no randomised trial of magnesium for exercise-associated cramps, or for cramps of disease such as motor neuron disease, other than the cirrhosis trial.

Practical Application

Magnesium is unlikely to prevent rest cramps in older adults, and this review found no trial of it for cramps during exercise: in older adults with night-time leg cramps it did not significantly reduce cramp frequency (−0.18 cramps a week, 95% CI −0.84 to 0.49; 5 trials, 307 people; moderate certainty). Minor, mostly gut, side effects were reported more often on it (RR 1.51, 95% CI 0.98 to 2.33, low certainty).
For a cramp during a race or a hard session, the base's guidance does not run through supplements: the cause is contested, with neuromuscular fatigue the best-supported thread (PHYS-31, REF-232); stretch gently until it releases and continue easier, and salt or extra fluid is not the fix (SYNTH-68).
In pregnancy the trials conflict and are all at high risk of bias, so this review settles nothing there; not from this paper, the base's own practice: that decision sits with the obstetric team (SYNTH-95).
Not from this paper; the base's own practice: a low magnesium level found by a blood test is a clinical finding for a doctor to treat, not a reason to self-prescribe for cramps.

Limitations

Written from the abstract: magnesium forms, doses and trial lengths are not given there, and the trial-level detail was not read.
No trial in athletes or in exercise-associated cramps; the idiopathic trials were small (271 people across five) and their main outcomes were measured at four weeks.
The pregnancy trials are all at high risk of bias and could not be pooled.
The first author is an author of an included study.
The search ran to September 2019.
How this entry was checked and changed
Why this level

the main claim rests on mostly small trials (five idiopathic-cramp trials, 271 people), which the rubric places at A-; risk of bias read in the abstract: for idiopathic cramps only one of five trials was at high risk, heterogeneity was minimal (I² 0-12%) and the frequency outcomes are moderate to high certainty; the pregnancy trials (all at high risk, conflicting) and the single cirrhosis trial support no conclusion. The absence of any exercise-cramp trial is an applicability point, not a level point (rule 4)