However, I'm not fully convinced there's a real need to "reuse" a variable name. If it's going into the same calculation, then some sort of indexed variable should be used. If it's going into another equation, then a new name should be picked. Even if you can deassign a variable, the usage of the same name implies that the variable is being reused, as is, since the reader might not even notice the deassignment.
Some times you want to 'de-assign' a variable so that the symbolic processor will use the variable name, rather than its defined value.
[actually what is happening is that it becomes a self recursive definition]
e.g. if r:=5 and then we define area:=pi*r^2 we can ask the value of area=xx.x but we can't easily ask for the symbolic differential because mathcad will think that all the values are fixed so the differential is zero.
If we insert an r:=r then the symbolic engine will know to stop at the recursion and give d/dr(area) -> 2*pi*r
Further, you can pre assign the result to a function
circ(r):= d/dr(area) -> 2*pi*r
If later you update the formula for area, the follow on equations will live update as well, and you will be able to use the default numeric value that was already in place for r [i.e. the r=5 above].
It just takes a little care to make it look neat sometimes. e.g. place the r:=r just across the right margin for a pretty print