If the particles interact with something between emission and detection, this can localize the particles to specific positions, and that will destroy the interference pattern. This is what happens in the double slit experiment with C70 molecules that I mentioned earlier. (Higher air pressure means less interference. Air messes up the results).
In a "successful" double slit experiment, one that does produce an interference pattern, the particles have made it through the slits, all the way to the detector (often a photographic plate) without getting significantly disturbed by photons or anything else. Each particle that reaches the detector would be putting it in a superposition of classical states if we could also prevent the detector from interacting with its environment. But we can't. It will scatter photons unless the room is completely dark, and it will emit photons if it isn't at absolute zero temperature. And even if it was, it would still interact gravitationally. These interactions will "mess up" the superposition of detector states. If the detector is a photographic plate, they will cause the discoloration to appear at a specific location rather than at a superposition of locations.
Is it correct to say that this "messes up the results"? The answer depends on what result we were hoping to get in the first place. If we wanted to determine the state that the system was in before the measurement began, then the result has indeed been messed up. But it makes more sense to include the requirement that the measuring device must be in a specific classical state as a part of what we mean by a "result". If we do, these interactions are the reason why experiments have results at all (at least in the MWI).