Thank you for the quick reply!
That is a good reason and a good guess. But, the block format already uses some "byte compressed" integers, notably the size of all vectors, vtx, vin and vout. That format is as far as I understand called compact size and differs some from the CVarInt class.
Below is a snippet from the protocol specification where they explicitly say that the CVarInt only is used for local storage, but doesn't say why.
https://en.bitcoin.it/wiki/Protocol_specification#Variable_length_integer"Modern BitcoinQT has also CVarInt class which implements even more compact integer for the purpose of local storage (which is incompatible with "CompactSize" described here). CVarInt is not a part of the protocol."
Any ideas regarding the above quote?