No... and Yes
This can work, but not in the way you're imagining it.
The victim's body has already absorbed the salt that the predator will use. It's already dissolved into the fluids and tissues of the body. The act of forcing it out of that dissolved state, only to have it re-absorbed into the very fluids it was drawn from won't do anything. You haven't dehydrated the body because the body has the fluids already.
But that doesn't mean you don't have a whomping good premise. Let's just change the effect. From the Australian Government's Better Health Channel we learn:
The body loses salt through urine, perspiration, vomiting and diarrhoea. If too much salt is lost, the level of fluid in the blood will drop. Hyponatremia is a condition that occurs when the sodium in your blood falls below the normal range of 135–145 mEq/L. In severe cases, low sodium levels in the body can lead to muscle cramps, nausea, vomiting and dizziness. Eventually, lack of salt can lead to shock, coma and death. (Source)
In other words, you're paralyzing or killing your victim because you're drawing the salt out of the bodily processes needed to keep the body functioning. The fact that the salt hasn't left the body is irrelevant, it isn't where it's supposed to be. It's like setting an electrical breaker inside a breaker box but not actually plugging it in. Yup, the breaker's inside the box! But good luck turning on your computer.
In a phrase, what your predator is doing is causing shock via hyponatremia. What I like about this solution is that you have different effects depending on the size or type of your victim (I love that list of possible conditions due to salt loss).
You'll be handwaving the speed of the reaction because salt is, well... it's a rock. And moving rocks takes time no matter how small the amount. That's the nature of chemistry. But I'd ignore this little inconvenience. I like the idea.