Despite the many other (unsourced) answers stating the opposite, this is actually a very valid concern. Not from a physical weakness angle, but rather, even though a shuttle mission was a couple of weeks long at most, the Space Shuttle Program was concerned enough about the deterioration of hand/eye coordination and the length of time since the commander had practiced the manual landing, that an onboard flight simulator using a laptop PC was created. This simulator was utilized late in the mission to practice a few landings.
The simulator was called PILOT (Portable In-Flight Landing Operations Trainer) and is described in the linked 1993 article from when it started flying.
You can see from this page from the flight plan of the very last shuttle mission that it continued to be used throughout the program. For this mission it was used on Flight Day 12, setting up for a deorbit and landing on Flight Day 13.
Here is the page from the generic Orbit Ops checklist describing how to start the sim.
Flight data file documents from the JSC FDF Page.
Incidentally, while it's true that the shuttle was completely fly-by-wire, the rudder pedals used a rather stiff spring to load the controls, and were by far the most difficult manual controls to manipulate. However, the shuttle rudder was rather ineffective throughout most of the entry due to blanking at the high angles of attack flown. The rudder pedals were mostly utilized for nosewheel steering after touchdown. Due to these reasons, the rudder was in fact deactivated by software until the vehicle speed dropped below Mach 5.
Rudder deactivation reference: page 2.22-18