Another option to enable reuse of your Reverb returns would be to create a global parameter for your Dialogue volume, and use that to automate a Gain effect just before your send:
You could also turn this into a Preset Effect to make it easier to add to each of your Dialogue events:
The reason the VCA does not affect the level of the Reverb send is because the VCA only modifies the final output level of entities assigned to it. The return gets signal from the event before the event’s final output, and because the return has it’s own output you hear this reverberated signal, even though the final output of the event is now muted.