I got to looking at my "prepper" sorts of supplies recently, and found I had some well-out-of-date iodine products -- out of date enough that back when I bought them the movement was called "survivalist" rather than "prepper". So out of curiosity I decided to test them to see whether they were still any good.
The first item was a little bottle of pills for sterilizing water, Coghlans brand, containing tetraglycine hydroperiodide. The bottle had 50 pills, with "8 mg titratable iodine" each. These bottles are date coded, and this one was from 1999. (For reference, the first three or four digits on the code stamped on the bottle are month and year.) The manufacturer says that an unopened bottle of these pills lasts at least four years, but this was twenty.
Now, I once had a bottle of iodine water disinfection pills that went bad after a few years, despite never having been opened. It was easy to tell that it had gone bad: the metal bottle cap had disintegrated and its surroundings were stained with iodine. Iodine is hard to contain: it diffuses through plastics and oxidizes metals. Glass works well, but it's hard to make an enclosure entirely out of glass, though chemists do in fact do so for exactly this purpose; they're called ampoules, and the YouTube chemist known as Nile Red has a nice video on how to make them. If you want your iodine stuff to last indefinitely, that's how to package it.
On the present bottle, the cap was in perfect condition, inside and out. It was also screwed on only loosely: the main thing holding the iodine in was the seal glued to the top of the glass bottle. That seal seemed to be in perfect condition. Taking it off, I found its inside stained black but intact.
Testing was by titration with sodium thiosulfate (photographers' "hypo"). I dissolved the entire bottle's worth of pills, which produced a brown mixture with some crud at the bottom, filtered off the crud, and added a bit of starch solution, in this case water from boiling spaghetti, to make the color change more emphatic. I also made up a solution of 1 gram of sodium thiosulfate in 100 ml of water. Then I added the latter to the former until the color disappeared; that took 75 ml of solution. Ideally it would have been 78 ml, so these pills still had 96% of the claimed amount of iodine; they had hardly deteriorated at all. Indeed, the 4% difference could easily be just my own experimental error. Whatever that seal was, it sure did its job. (It was mostly plastic, but might have had a metal film incorporated into it; I didn't look closely enough to tell.)
Your mileage may vary, of course. This particular bottle had mostly been kept indoors at comfortable temperatures. But now I know how to judge other bottles of that brand: unscrew the cap and examine its condition and the condition of the inner seal. If there are no signs of leakage (brown stains, iodine smell, and/or corrosion), the pills are still good.
Next up was a bottle of "Rad Block" potassium iodide pills, expiration date "AUG 03", and made three years before that. These are for use in a nuclear emergency (reactor leak or atomic warfare); one takes them to saturate the thyroid gland with iodine and prevent radioactive iodine from being absorbed.
You might think that potassium iodide is an inert salt, like sodium chloride, without any propensity to go bad. But that's only true until oxygen diffuses in and oxidizes the iodide to iodine, which can then leak out; water vapor (which also diffuses in) may also contribute to the reaction. Iodized salt, which is commonly kept in unsealed containers, loses its iodine due to this, and loses enough of it to make public health experts worry a little.
This bottle was sealed, but was plastic, which both oxygen and water can diffuse through. (In this case the plastic was HDPE, but in the long run pretty much any plastic allows diffusion through it.) Opening the bottle, the pills had clearly gone bad: they were reddish-brown, and with little bits broken off in places. These pills are supposed to be "slightly yellow", with a "clear coating to prevent any bitter taste"; the reddish-brown color would be the iodine contaminating that coating.
So they looked bad; but just how bad were they? As witness the use of iodine for water disinfection, a little free iodine isn't going to lead to health problems.
After dissolving the whole bottle's worth in water, I measured the pH, to see how much iodide had been replaced by oxide or hydroxide. Either would end up as hydroxide when dissolved, and would make the solution alkaline. Instead it was nearly neutral, so there was no sign of this having happened. Even if CO2 had also leaked in and converted the hydroxide to carbonate, that would still have been alkaline. (Of course there may have been some other ingredient in the pills to moderate or cancel out the pH change.)
After filtering off the fillers in the pills, I then tried to extract the iodine as a pure element, by oxidizing the iodide with hydrogen peroxide and sulphuric acid, somewhat along the lines of this video. This was not a very successful attempt -- a lot of brown water made its way down the drain -- but out of the 10 grams of iodine that was supposed to be in the bottle, I got 7.65g of elemental iodine. So certainly the vast majority of the iodine was still there, and probably almost all of it, the loss just being during my own processing. Likely the pills would have been just fine to take, though there is a tiny chance of some sort of toxic iodine-containing chemical being formed as a result of decomposition. They'd have had to be real klutzes to formulate iodide pills that would produce such a decomposition product, but real klutzes do exist in the world.
In a nuclear emergency, given the choice between taking old expired potassium iodide pills like this or taking no pills, I'd still take the pills; even if there were some toxic iodine chemical produced, the levels probably wouldn't be significant. But I'd still rather not have to choose. Potassium iodide pills are readily available; the better sort now seem to have nominal shelf lives of six years rather than three, packaged in foil-sealed blister packs.
They're more expensive than my old ones were, though. To replace them I opted to buy the pure chemical (USP grade), which turns out to be reasonably available for about two orders of magnitude less money per amount of iodide. To dose it requires a scale reading in milligrams, but these days tiny scales that do that are cheap. The pure chemical might not be pleasant to eat, but the quantity needed is so small that I'm sure a way could be found. I'm storing it a glass jar with a metal lid, with silica gel and oxygen absorbers along with it, in the hope that it'll last indefinitely.