Fred suggested quick & easy solution which usable in most cases.
If possible (and it usually is) then I prefer to have everything inside style itself. The solution I decide for, depends on song and style complexity.
For example... if I'm only using two variations (i.e. MAIN A and MAIN B), then I simply create a third variation (copied from MAIN A or MAIN B), where I disable voices I don't need (Rhy1 & Rhy2). And during playing I simply switch to desired variation. If all variations are already taken, then I create FILL-IN with particular voices disabled. This can also be done by making an INTRO with disabled Rhy1 & Rhy2 -is probably the best solution, because it's easy to remember during playing.
It's up to you what you do, I'm just saying there are many choices.
Bogdan