There's an article on replenishers published in the BJP Almanac in the 1950's. It's was part of a series of research documents.
Essentially Ilford tested how developing agents were depleted by development and how bromide & iodide build up impacted the replenished developer. This was during the evaluation & design of a PQ replacement for D76 for photofinishing.
There's two types of replenishment, bleed where a fixed amount of replenisher is added per sq meter of film processed, some developer must be bled off first to allow the volume to stay constant, that's how D76 is replenished. This is to keep the level of by products like iodide & bromide leached from the film low.
Topping up replenishment just adds what's used and carried over to keep the volume constant, Ilfoprd's PQ photofinishing dev Autophen could be used that way, Metol is highly susceptible to iodide & bromide build up, phenidone isn't. Colour processors use the top up system.
A PQ version of D76 can be replenished virtually ad-infinitum while D76 needs dumping every 6 months or so, depending on throughput.
So you can calculate the extra developer you need to add, and from testing the ph show what additional buffering is needed, you'll see that replenisher never has any bromide even when it's used in the developer.
That's the simple answer
In practice Ilford did extensive lab testing and with commercial photo labs on a larger scale before finally releasing Autophen.
Ian