Hi Peter,
What Hickok tester? One that only measures shorts or one that measures leakage and shorts such as the 539 type KS and RD.
If you tubes fade that fast you should have done a "life test". Yes, gas can affect a reading on a Hickok, but it will make a tube read higher not lower since 'gas' takes the grid 'more positive' into the negative voltage range ( makes the tube grid less negative and thus more output as if you turned the pot to 10 or 12 bias from 14). If your tubes fade back 10% from indicated gm in under 5 seconds they are not reliable...period.
The problem with tubes that show such drops is the cathode is worn out or the grid is not able to control the electron flow as it should.
If you Hickok is a model that can show leakage, and ANY tube show leakage under 1 mega ohm, pitch them out. ANY under 3 meg keep an eye on. You get below 1 mega ohm and the chances of a instant full short are greatly increased. Low end Hickok testers will not see anything above 250,000 ohms and that is to near a short for any audiophile amp to have to face. You must remember that your amp using KT88 is well over 500vdc and arcing is a real concern even for NEW tubes.
It is not worth it if you blow a screen resistor or even worse damage the output transistor because some element fuses with another.
If you see a reddish plate or screen wires thru the holes, you best turn it amp off and yank those tubes out for good. Even new tubes should be observed for the first few hours for this sign.
No matter how expensive or sophisticated an amp is, burning second rate, unmatched tubes in it is just plain not worth it. One blowout on one tube and you can be sending that Amp to be repaired, so there is no bargain there. How do I know, after 40 years in this 'business' I have never had an amp or tube tester returned.
warm regards
Bud