Yes, I change electrolytic caps as a general rule in recent devices. Anything from the last 10-15 years, let's say. I have junk from the late '60s with big lytics that still work, so it's not a mystery how to make long-lasting caps; people choose to build crap.
I haven't done any measurements but I suspect as the caps age, they cause high leakage currents to flow in them. Currents which depend on the working voltage, ie you won't measure anything too bad at 5V, but at 400V you will.
Further, I think that the crimping of the pin to the foil inside also gives up, probably because there isn't enough/any corrosion inhibitors in the cap.
Cap goes, transistors fry, diodes short, fuses blow. Just make sure there isn't any charring on the pcb, that stuff is conductive. You can route out a channel, that takes care of that.
And who even uses DIP MOSFETs? 
Tim
Low cost assembly, all single-sided stuff made with through-hole parts and wave soldering. You're not going to go out of your way to add a surface mount part in that case. ("case". ahah)
For example, in my Garrison fan repair story, the DC supply that feeds the controller uses a weird and wonderful part called the LNK362 in DIP.
You may be worried about thermal resistance of the DIP case, they get around that by using 4 pins for the source.