The purely negative utility monster (whether it is in a ferociously large amount of pain or not), that also has by definition no diminishing returns in its utility function, just hits zero pain at some point. Until it is in pain again, it is simply not part of the equation. The difference is: If your goal is to minimize X, you can't go on forever without diminishing returns (but with diminishing returns, you can) whereas if your goal is to maximize Y, you can go on forever with or without diminishing returns.

edit: It depends on how the function is defined. Above, I used allocated resources vs. utility (utility = relieve from suffering). But a negative utility monster would be possible if its condition got automatically worse and if it had no diminishing returns of (f.e.) suffering per unit pain, but all the other beings had.