There are a couple of registers that I also use in the
interrupt indeed, ie: a,r0 (@r0).
What I do is:
At the start of the program I "init" the
chip, I configure the interrupt and set the the stack
at 30h, address space below 30h is variable space for me.

At the start of the interrupt I push the Accu on the stack, copy r0 to a and pushes Accu on the stack again.

After the interrupt (just before RETI) I pop back accu
and store that in R0 and then I pop back accu once more.

Other addresses aren't used in the interrupt.