From the rest of the thread, it is clear that your front end is a very slow 1uA comparator,and your problem is that this comparator is slow to swing through the - say - 1.0V to 2.3V range were both N- and P- channel MOSFETs are turned "on".
The answer would appear to be a second - faster - comparator that doesn't draw current except when the output of the first comparator is in the problem region. Because it is only on intermittently it can draw more current when it is on, and swing faster. Because it works on the output of the first comparator, you can afford to build-in plenty of hysterisis.
Of course, to knowing that the output of the first comparator is in the danger area requires the services of a window comparator (two more comparators if you build it conventionally) though this might just be an N-channel MOSFET current sink below the second comparator and a P-channel current source above it, with their gates both driven by the output of the first comparator; so no current flows through the second comparator until you actually need it.
When I had to deal with a similar problem in a discrete circuit some thirty years ago, I found myself using a bipolar Schmitt trigger to avoid the shoot-through current problem. Your IC process may not offer decent bipolar transistors but if it did, that approach might still work.
-- Bill Sloman, Nijmegen,