An LED filament lamp carries the same cap as the incandescent lamp it replaces, and the same gauge sheets apply. What changes is the mass distribution, the heat path through the cap, and the way the cap is attached, and those change which failures are likely.
Filament lamps are often heavier at the base than an equivalent incandescent because of the driver. The cap cement therefore sees a higher sustained torque when the lamp hangs in a pendant, and creep becomes more likely. A cap that passes dimensional gauging and a room temperature pull test can still rotate in service, so a creep check at elevated temperature is worth adding to the type test.
The driver dissipates heat into the cap rather than into the bulb, so cap temperature can be high even when the luminous output is modest. Cap temperature rise measurement is the relevant test, and it should be run at the worst-case mounting orientation rather than in free air, because a recessed fitting changes the answer completely.
Filament lamps are assembled in high volume with automated cap fitting, and the variation between machines can be larger than the variation within a machine. Gauge by machine, not by lot, or a single out-of-adjustment head will be hidden inside an acceptable lot average.
