I had a look at the March 2014 Guide.
For accessories, the key paragraph is Section XII:
If your submission is for a specific component of the electrosurgical device, you are still expected to evaluate your component while connected to other components of the electrosurgical device and consistent with how you intend or expect your component will be used.
The guide uses "component" to mean both accessories
and the ESU unit, while "device" means the whole system, which is a bit confusing. So the paragraph applies equally to the unit as well as accessories. Also the paragraph appears in a section which covers both electrical safety and EMC. It comes after a paragraph on EMC, so a reader could assume the context is just EMC, but a proper reading covers both electrical safety and EMC.
So in other words, it is a fairly wide requirement, saying that if the submission is for any single component in the system, you should consider system issues as well.
That said, engineering common sense should still apply in selecting which requirements apply to which parts in a system. For electrical safety, for example, there are several tests on accessories which are reasonable, for example HF/mains dielectric strength, HF leakage etc.
For EMC though, it is 100% clear that tests are not required on simple accessories such as cable or electrode.
Yes, these kind of accessories can certainly influence the EMC result, but the source of the emission is the ESU unit, not the accessory.
Unless there is a contractual relationship between the manufacturer of the accessory and the ESU unit to report changes and exchange design information, any tests done by the manufacturer of the accessory are meaningless, because they have no control over design and production of the ESU unit, nor do they have the information about the internal design to know what are the worst case conditions, settings, mode etc, nor are they able to assess immunity without the designer's technical information and risk management file.
The ESU unit manufacturer is free to make changes at any time, and they are responsible to re-evaluate EMC after any changes. As the guide suggests, they should consider tests with a variety of accessories to represent the range expected in normal use.
But the manufacturer of the cable or electrode is unable to know what changes were made. It is a completely unacceptable for the FDA to insist on these tests.
I sure that any FDA reviewer asking for EMC tests on a cable or electrode is just working off a script without thinking, and they are sure to see common sense once explained.