Heck of a good question. On the one hand, you'd expect that anything that can intersect with the black hole would eventually do so until it all ran out and the black hole quieted down, unless low-angular-momentum ellipticals are so chaotic that more just keeps raining down. On the other hand, these black holes only grow and aren't getting any smaller (on the timescales that matter (seriously it would take so long for a supermassive black hole to evaporate that the relevant question is 'is matter infinitely stable or only nearly infinitely stable')).

I do know that a feedback mechanism has been proposed to keep cluster gas clouds a consistent temperature, in which comparitively tiny amounts of cluster gas inflow increase the activity of the black holes proportional to the amount of cool inner-cluster gas and thus a negative feedback loop is closed and the clouds remain fairly stable.