20% mannitol is 100g of six-carbon sugar alcohol, delivered in fresh water.
It is an inert alcohol, closely resembling mannose, but without the aldehyde group which would make it a sugar.
Sugar alcohol, you say.
Yes, it is the reduced daughter product of glucose, reduced commercially in huge vats. Alternatively it can be made when fructose is hydrogenated over a nickel catalyst (which forms a mixture of sorbitol and mannitol).
Humans are probably rather unique in not being able to metabolise it. Indeed, even when sorbitol - another hexose alcohol - is administered to humans and rats, some of it is deposited in the liver as glycogen, giving the impression that it is being metabolised. Not so for the mannitol, which (even when given in an inventive variety of ways) does not undergo any metabolism.
Baxter confess that in their Osmitrol solution the pH is adjusted with sodium hydroxide and hydrochloric acid. The mannitol itself has a pKa is 13.5 so pretty much in all physiologically relevant situations it will not dissociate.
In a way, it is attractively inert, much like a noble gas. The human body simply neglects this substance, and it exerts its effects by virtue of its osmotic effects. There is about 182.2g to every mole, which gives the 100g contents of a half-litre bag an osmolality of about 1100 mOsm/L. Each molecule, having such a small mass and remaining non-polar, is capable of penetrating the junctions between cells rather easily, and so this substance is able to distribute rapidly between extracellular fluid compartments.
Mannitol exerts its most useful effects by remaining in the extracellular fluid exclusively. Its entry into the cell is frustrated basically by an absence of a uptake pump. Human cells are just not that into mannitol. This results in a fairly restricted volume of distribution, roughly parallel to the volume of extracellular fluid (0.47-0.5 L/kg)
The clearance of mannitol is renal, wherein lies its other most significant advantage. The molecule, being freely filtered in the urine, is a powerful antagonist to normal urinary concentration mechanisms, and drags torrential diuresis out with itself, robbing the body of lots of sodium and water. There is no tubular mechanism for mannitol resorption.
It has a rather short serum half-life. First, the dose given into the bloodstream distributes very rapidly into the extracellular fluid, and the intravascular mannitol content fairly rapidly filters though the glomerulus and out into the urine. After 3 hours, 80% of the administered mannitol dose makes its way out of the patient.
What if I drank it?
It would be deliciously sweet. But one would be foolish to drink too much mannitol.
Firstly, anything more than 20g will act as an osmotic laxative. That is about 100ml.
And, one must recall that all those colonic bacteria are perfectly capable of metabolising it into H2, CO2 and water. The gases they produce will work to anger the bowel, causing distension and unpleasantness.