https://www.secumar.com/en/servicing/
For bike helmets, independent safety labs have actually done the science and found no measurable deterioration in impact testing nor material properties with helmets as much as 26 years old. (Link to journal paper at end of summary linked below.)
Bicycle helmets just add a few mm of valuable deceleration distance in the best case, that's all they do apart from protecting from purely cosmetical abrasions. The maximum amount of extra deceleration distance is provided at one very specific kind of impact. When the material properties change, the maximum will be achieved at a different impact. But you cannot predict which one you will have that one time when you actually use the helmet. The old one might be better or worse, nobody knows.
If manufacturers knew exactly what would eventually hit you, lots of engineering could be applied to precisely tune the helmet to that impact, but they don't know so they cannot do that. So all the engineering goes into achieving comfort while hitting regulation targets and into marketing features that only need to appeal to "common sense". There is no engineering at all going into the main purpose. Even aerodynamics, a clearly defined performance metric orthogonal to protection, is just barely starting to move on from hand-wavey "fast looks" to actual engineering.