12/06/2026
Magnesium does not cross the intestine by mass action, where more in the lumen means more in the blood. It crosses through TRPM6, a channel fused to a kinase that sits in the lining of your colon and kidney and pulls magnesium across one ion at a time. That channel has a ceiling.
Fine and colleagues measured exactly where the ceiling sits. They fed normal subjects a standard meal spiked with escalating doses of magnesium, from nothing up to 80 mEq, and tracked what actually entered the body. The total amount absorbed kept climbing with every increment, so a bigger dose is never wasted in the sense of giving you zero. But the fraction absorbed collapsed, from 65 percent at the lowest intake to 11 percent at the highest.
The absorption curve was not a straight line. It bent, and it bent in a way their math captured almost perfectly: one saturable mechanism that maxes out, sitting on top of a second mechanism that passively absorbs a fixed 7 percent of whatever you swallow no matter how large the dose. Past the point where TRPM6 saturates, you are running on the 7 percent. Everything else stays in the gut lumen, where unabsorbed magnesium is osmotically active and pulls water in behind it. That is the loose stool. It is not a side effect to push through. It is the visible signal that you have exceeded what the gate can move, and the excess is on its way out.
What you absorb then has to get into cells, and that is a second channel entirely. TRPM7 is the ubiquitous one, expressed across nearly every tissue, and it governs the intracellular magnesium pool that stabilizes ATP, runs DNA repair enzymes, and gates nerve signaling. TRPM6 and TRPM7 are the only two of the eight TRPM channels highly permeable to magnesium, and they are not interchangeable. One is the border crossing at the gut. The other is the door into the cell. Neither responds to brute force.
The Pardo systematic review reached the dosing point from a different direction, pooling fourteen human bioavailability studies and concluding that the percentage absorbed is dose dependent, the same curve Fine described. It also found a real difference between forms, with inorganic magnesium appearing less bioavailable than organic, which matters but is a smaller lever than the marketing on a single bottle suggests. Fine's own data made a related point decades earlier in a way nobody selling capsules likes to repeat: magnesium from almonds was just as bioavailable as from a soluble magnesium salt, and commercial enteric-coated magnesium was absorbed far worse than the uncoated form. The coating that sounds sophisticated lowered the number that matters.
None of this means magnesium supplementation does not work. It means the dosing strategy printed on most bottles is built backward. Smaller amounts, taken more than once across the day, with food, is how you stay on the steep part of the curve where most of what you take actually crosses. One large dose at bedtime is the worst version: it saturates TRPM6, hands the rest to the 7 percent passive route, and sends the remainder through your colon. The biology here is not new and it is not contested. It is settled human pharmacology from 1991 that the supplement aisle has spent thirty years not mentioning, because "take more" sells better than "take less, more often."
Fine et al., J Clin Invest, 1991. doi:10.1172/JCI115317
Pardo et al., Nutrition, 2021. doi:10.1016/j.nut.2021.111294
Sun et al., Channels, 2015. doi:10.1080/19336950.2015.1075675
Madlmayr et al., Life Sci Alliance, 2026.
