Tht's right - if you need fast response it is impossible to slew the gate of the pass device up and down with 100nA. This is probably a device that changes it's bias if it is in slew - similar to the "gain boosted" or "Slew-rate enhanced" op amp.
So we cheat, and slew the gate with 2mA, but after everything has settled out - look, my Q current is only 100nA!
Although it sounds nice, it's not too practical - the wideband noise from this LDO will be way too high to use in sensitive applications, and it will be very sensitive to glitches on the power supply due to it's very faint q-current.