The "Mark Coles" school of thought....I agree.
However the work being done will need an EIC (not a MWC) by virtue that every circuit is being changed. The person doing the work will have to sign it off for every circuit and the date of installation will be the new cert date. Therefore I am compelled to believe that it should be done in accordance with the latest editions. The original installer could quite legit say that after the new work is done it is no longer her install.
The problem is that it creates for you a whole mess of inconsistencies and illogical positions, and I believe that if you can interpret a set of technical and engineering requirements in two ways, one of which is consistent and logical, and the other is not, then the latter must be the wrong way to interpret it.
My approach deals with the work that you do, as in "I...hereby CERTIFY that the said work for which I have been responsible is to the best of my knowledge and belief..." is what has to comply.
So what is that work?
Joining cables, so you have to comply with the regs for joints - fitness for purpose, properly rated, properly done, properly enclosed and accessible etc.
Installing cables, so you have to comply with the regs on things like current carrying capacity, protection, (inc RCD and locations if you conceal your cabling in a wall or partition etc).
As a result of the above, you should confirm by testing that you have not messed up the fault loop values, that you still have proper continuity of ring final conductors, that you still have adequate installation resistance etc.
You've (possibly) installed a new CU, so you need to certify that the MCB ratings are correct for the circuits, and record their types, you need to check the operating times of residual current detectors, and so on.
You've modified the PEBs, so you have to certify that they are OK, and if they were legacy undersized ones you have to replace them, if they were clamped to pipes in the wrong places or with the wrong clamps you have to sort that out, and so on.
All through though you are dealing with the compliance of the work for which you are responsible, and there are no inconsistencies between what you think you should sign to certify and what you can sign to certify.
The problem with the "you become responsible for the whole circuit" approach is that that is utterly impossible.
You can have no way of knowing whether 522.6.6 has been complied with and you have no realistic way of finding out. Therefore you cannot certify that the existing installation complies with BS7671:2008.
Ditto 526.3.
If a previous occupant had drilled through a ring final cable and bodged a repair, you would have no way of knowing that somewhere directly buried in plaster was a pair of 5A choc-blocks and a few cm of 0.75mm flex, nor would any testing that you carried out tell you that this had been done. So another uncertainty which stops you certifying that the existing installation complies with BS7671:2008.
And it just goes on and on and on, with reg after reg after reg, compliance with which you cannot possibly know or determine, and therefore you cannot certify.
So to pick out one or two, and use the argument that you have to do those because you're becoming responsible for the whole installation is inconsistent, and illogical, and therefore cannot be right.