It doesn't sound like a hard and fast rule. It seems like it's whatever the manufacturer asks for in terms of "intended working life." This _seems_ to be the original FDA certification for the device:
https://www.accessdata.fda.gov/cdrh_docs/reviews/den130034.p...
The problem here, of course, is it's not the device itself, but the simple remote control peripheral that is designed to switch device operating modes.
The FDA should force manufacturers of "two piece" systems like this to have backup controls on the device itself and to exclude these non-medical components designed for control from any regulation covering "intended design life."
It seems like a daffy middle ground that the FDA lets exist and manufacturers take advantage of when they can.
There is no part two to the certification?
Edit: Seems odd that the loophole is big enough to allow not even uploading the full documentation.
> It seems like it's whatever the manufacturer asks for in terms of "intended working life."
I would guess the FDA makes their certification easier or harder to pass based on how long the device is expected to last. If the 5 year lifespan is what hits the middle ground for that company between cost-of-certification and useful-product-life, then it's up to the FDA to make a longer lifespan feasible to get certified.