Lateral Mosfet Tests - Rauch DVT 250s Restoration

[Xoc1]

I am rebuilding a Rauch DVT 250s Amplifier
550W@4Ohm, 4 pairs of 2sk1058 -2sj162, 80 - 0 - 80 volt rails, & 1KVA transformer per channel.
The driver boards are completely rebuilt with new components, and modified to latest specification.
RF oscillation had blown one PCB to bits. The output fets had been changed, but capacitors to stop oscillation had been left of the output pcb.
I have managed to get the amplifier to behave its self, passing most service tests - DC offset, CMR rejection, Off load voltage swing, & RF oscillation.
Then I noticed audible distortion on one channel, that was not really apparent untill I ran it down to a 4 Ohm load.
I managed to isolate the problem down to one set of output mosfets, and when I stripped this PCB down one mosfet was not screwed down.
I have tested all the mosfets with a resistance meter, and tested them for VGS in situ on the heatsink and maximum variation is about 0.03V.
I used the information here for simple tests
MOSFET testing
and Nelson Pass's excellent info
http://www.passdiy.com/pdf/mos.pdf.
According to these tests this would give the mosfets a clean bill of health.
But I have only really tested them at 12 Volts, not 160V.
I am wary about the effort involved in mounting & rebuilding the output boards, The pcb tracks are beginning to look fragile! So a new set of mosfets have been ordered.
This leaves me with 16 used fets to play with......
Are there any more tests that I could use to try to isolate any rogue devices?
Also does anyone have any experience of what happens when a Lateral fet is not fixed down? Would re-fixing the part in position, and rebuilding the output board stopped the distortion under load. Or is that fet toast?