On a piece like this, with a big feedback loop, no reason tubes shouldn't last forever with little change in performance, and only show sound and gain differences when absolutely dying.
A difference in 1st tube type only lessens the gain in the feedback loop; feedback loops like this are self balancing and correcting systems to a high degree. I don't believe it changes the actual gain of the unit to any real degree. Less feedback gain, softer clipping sound, different harmonic structure.
I can't see why anyone would need to pad the output of an LA-2A, unless interfacing a -10 system.
If you want more output, and usually drive a 10K load, why not use something like an Edcor 15K:10K or 10K:10K for output. You could easily make it switchable between A-24 and Edcor for a roughly 14 dB change in max output level.
First version LA-2A had socket for a plug-in input pad. Put a U pad in front of it designed for 600 ohms, or an H or O if you are driving it with an actual 600 ohm balanced source. An A-10 can take the average modern level fine, but it can be close enough to the border to start sounding pushed. The input pad will improve input transformer headroom. It does nothing to active amp headroom, only changes what the knob position looks like. Or again, change to a 10K:10K or 15K:15K etc input and reduce input gain by nearly 20 dB.