Magnesium

The people tested were mostly older adults with night-time leg cramps, not athletes cramping mid-race, so this settles the mineral rather than the sport. The pregnancy trials in the same review could not be pooled at all because their outcomes were reported inconsistently, and a separate Cochrane review graded those same trials as far less biased than this one did. Nothing here tests magnesium during exercise.

The larger figures come from a slice of the data examined after the fact rather than the review's main comparison, which makes them the weaker number even though they are the more striking one. The ten pooled reviews overlap, so some of the same original trials are counted more than once. And blood pressure is a stand-in here: none of these trials counted heart attacks or strokes.

This is magnesium sulphate at monitored clinical doses through a vein; magnesium sold in capsules has never been shown to do anything for eclampsia risk. The trial ran across 33 countries with very different standards of care, so the background risk it was measured against varied from site to site. And the Cochrane review is built around this same trial, so the two are not independent of each other.

The reviewed trials establish the direction but not the size; how far the risk falls is not settled here. This is a different job from preventing a mother's seizures, working a different way, and the two should not be read as one use. And it applies to births that are very preterm, not to pregnancy generally.

Everyone in these trials already had one of those conditions, so a healthy adult has no reason to expect the same shift. The trials ran between one and six months, which says nothing about a year on it. And 543 people across 11 trials is a modest base for a figure this precise.

The drip is given in a clinic under supervision, and nothing about that result transfers to a capsule taken at home. Some of the pooled studies assigned participants poorly, by the authors' own account. And each half of this rests on fewer than a thousand people.

Seven trials and 325 people is a thin base, and the authors ask for larger, better ones before anybody draws a clinical conclusion. An effect this big out of a pool this small is also the pattern produced when small trials that found nothing never reach publication. Magnesium was compared with placebo, never with a prescription antidepressant.

Four trials and 332 women leave enough room that a small real benefit could still be hiding inside the result. Cramps were defined and counted differently from trial to trial, which blunts any pooled figure. And this is cramps in pregnancy, not the ordinary night-time leg cramps most people mean.

The two trials used different magnesium compounds, so the split may be about which form sat in the capsule rather than about magnesium. Both groups improving by a third in the null trial points at expectation and the natural course of cramps doing much of the work. In the oldest positive trial, blood magnesium never actually rose, which makes it hard to credit magnesium with the improvement. And the trials counted cramps so differently that no single number can be drawn across them.