Well, with a mobile phase of straight water or organic solvent without any ions in the mobile phase, one can do some strange things with the right stationary phase. I remember publications by Blasius using crown ethers as the packing. You can separated a series of salts with the same anion, but different cations, or a series of salts with the same cation, but different anions. With other words, if you would inject a solution of sodium chloride and potassium bromide, you could get 4 peaks: sodium chloride, sodium bromide, potassium chloride and potassium bromide.
However, this all goes belly up, when you add a salt to the mobile phase.
I will scratch my head a bit more about your observation, and the possible reason for it.