Mesh analysis doesn't use current because the sum of the currents around the loop just doesn't make sense. If it's a single isolated loops the current in one side of the loop IS the same current flowing in another side of the loop. Another reason (and another way to think about it) is that currents can branch off into "adjacent" loops at T-intersections and X-intersections at the nodes sitting on top of the sides of the loop. Therefore the current in each of the 4 sides of a rectangle loop (or however many sides) are not necessarily equal unless it is an isolated loops. That is, the current can branch off into adjacent loops one side of the loop, and only the remainder of that current moves on to the next side of the loop.
The sum of voltages around a loop however, equal to zero, since, obviously the voltage at a certain point in the circuit is still the same if you start travelling from that point around a circle back to that same point. So while loop analysis does not use Kirchoff's Current law, it does use Kirchoff's voltage law.
I haven't use nodal analysis since I learned it because, although it made perfect sense while it was being taught, when the homework and exams came around I always ended up with like 50 equations. Heh, I've only recently discovered that you only make equations for the junctions (nodes with >2 connections) and not every single node in the circuit. lol. Needless to say, because of that I've never used nodal analysis for anything useful- always mesh.