What do you mean by threshold?...
formatting link
:-)
At constant Vds, there is a one-to-one function between Vgs and Id. Vgs(th) is _a single point on that curve_, defined by the conditions specified (usually constant Id, and either fixed Vds, or Vds=Vgs i.e. diode strapped).
Whatever happens at other currents, needn't be related to this parameter.
You're asking if, the _spread_ in the parameter corresponds to the same spread, using a similar test but at higher current?
Possibly.
The transfer curve (Vgs to Id) should be more or less consistent between parts, in terms of its shape. Its slope will vary with device geometry, and its offset will vary with process (doping, gate oxide thickness, etc.). It's an offset-and-gain (y=mx+b) sort of situation, except it's not a straight line but a sloppy curve thing, but the change follows that pattern anyway.
Because the gain factor varies too, it's not the case that the Vgs(th) variation is due entirely to offset. Both offset and gain are baked into the figure.
But the same variation is baked into the Rds(on) max figure. For the same Vgs, Id and Tj, Rds(on) will always be less than that max. For more Vgs, it will also be reduced (you can use the Rds(on) vs. Vgs plot to estimate how much, of course, that will no longer be a strictly worst case condition).
So you don't have anything to worry about!
Tim